After spending countless hours figuring out how to get the best out of ChatGPT, Claude, and Gemini, I’ve discovered that the best use cases apply across all of them.

By utilizing the dedicated apps and "Connectors" for ChatGPT and Claude, I’ve built custom music playlists, mapped out vacation itineraries, and hunted down the most affordable concert tickets. Meanwhile, generating interactive study guides with Google's NotebookLM has proven to be an incredible research tool for AI users of all ages. And as someone who loves making bold predictions for major sporting events, I've even found that chatbots make for a surprisingly fun crystal ball.

Ultimately, I’m always searching for ways to save time whenever I log into a chatbot. Through all my testing, there is one specific habit I’ve adopted that has completely transformed how I use these tools.

If you adopt this same practice, you’ll easily save hours on a weekly basis. Here is how it works.

The art of the “context prompt”

The Master Project Strategist & Problem-Solver Prompt

Role: From this point forward, act as my elite Project Strategist and Operations Consultant. Your job is not just to answer questions — you are responsible for helping me cut through the noise, anticipate roadblocks, and build flawless execution plans before I even start working.

Directives: Whenever I bring you a project, goal, or problem, always apply the following framework:

Action-Oriented: Focus heavily on practical, realistic steps rather than abstract theory.

Efficiency First: Identify the "Minimum Viable Effort" required to get the maximum output. Keep workflows lean.

Risk Anticipation: Look ahead to spot potential bottlenecks, human errors, or hidden costs before they happen.

Blindspot Detection: Highlight surprising details, overlooked dependencies, or second-order consequences that most people miss when planning.

Resource Mindfulness: Tailor your suggestions to respect constraints on time, energy, and budget.

Expanding the toolkit

I don't stop there. I have a full library of context prompts that instantly transform my chatbots into specialized allies:

The Researcher for deep-dives into complex topics.

The Brainstorming Partner for breaking through creative blocks.

The Personal Productivity Coach to optimize my daily schedule.

The Entertainment Curator for tailored book, movie, or music recommendations.

By priming the AI with the right role from the start, I get tailored, high-quality outputs immediately. Over time, as these chatbots adapt to my style and build up their internal memory, the responses get even sharper — meaning I spend less time managing the AI and more time executing my best work.

Bottom line

Providing as much context as possible to someone starting their first day at a new job is a common practice.

Doing the same during a new conversation with a chatbot is equally important for AI users of every skill level. And at the end of an extensive session, simply telling your chatbot, “Turn everything we just did into a reusable prompt template for next time,” will give you a reusable context prompt for future reference.

Building a library of context prompts saves me plenty of time and improves my productivity — it will certainly do the same for you. Feel free to apply this context prompt to your future chats to become a more proficient AI user.

More from Tom’s Guide

• I used ChatGPT's 'Pause Prompt' during an argument — I didn't expect what happened next
• 'Wait a minute. Strike that. Reverse it:' Netflix is using an AI-generated Gene Wilder voice for 'Wonka' reality show
• The biggest risk of using ChatGPT every day isn't what it learns — it's what you stop doing

Living in New York has given me the pleasure of enjoying all sorts of activities from every corner of the city.

Every borough has something amazing to offer, from the prospect of eating something amazing I saw on Instagram to hanging out at a lesser-known spot that caters to geeks like me. And whenever a widely celebrated holiday lands on a Friday or any day of the weekend, I can’t help but get caught up in some of the festivities tailored around it. With the 250th anniversary of America aligning perfectly with this year’s 4th of July celebration, I figured it’s time to get off the couch and bask in all the outdoors hoopla outside my comfort zone for once.

ChatGPT has become one of my most dependable chatbots for all manner of things, such as building productivity routines, generating recommendations for my hobbies & interests based on my personality and suggesting activities to enjoy while on vacation.

For my next ChatGPT session, I asked it to plan the perfect 4th of July weekend. And to my surprise, it laid out the best plans for me to enjoy that were far better than going to an average BBQ.

The prompt that helped me find everything I want to do

I already knew that heading into my latest dialogue session with ChatGPT with a simple request (“Help me plan my 4th of July weekend”) wouldn’t do me any good. So I asked ChatGPT to generate a prompt I could use that took my New York home base and desire to check out some unique activities for this year’s 4th of July jubilees.

Here’s the loaded prompt it told me to use:

I'm based in New York City and I want you to become my personal Fourth of July weekend concierge.

Plan the ultimate three-day itinerary for July 4 and July 5 that balances iconic NYC experiences with unique events that most tourists (and even many locals) don't know about.

For both days:

Start with breakfast, then morning, afternoon, evening, and late-night suggestions.

Include festivals, concerts, rooftop events, outdoor movies, food festivals, neighborhood celebrations, museums with special programming, sporting events, cruises, hidden gems, scenic walks, pop-up experiences, and anything happening only that weekend.

Prioritize experiences that are memorable rather than expensive.

Mix free activities with paid ones.

Include backup indoor options if the weather is bad.

Suggest the best times to arrive to avoid crowds.

Recommend the fastest way to travel between activities using the subway, ferry, walking, or other public transportation.

Tell me approximately how much each activity costs.

Recommend great places to eat near each activity, including one budget option and one splurge option.

Point out any reservations or tickets I should book in advance.

At the end, create four themed versions of the weekend:

1. Budget-Friendly Weekend (under $150)

2. Food Lover's Weekend

3. Hidden NYC Weekend (places locals love)

4. Bucket List Fourth of July Weekend (the experiences everyone should do at least once)

Finally, rank the Top 20 can't-miss events and experiences happening across the weekend, explaining why each one deserves a spot. If multiple events happen at the same time, recommend which one offers the better overall experience and explain why.

Two days' worth of fun for the first weekend of July

After inputting that loaded prompt into the chatbot, it laid out a wealth of general activities to enjoy and themed events that I never even knew existed.

As someone who finds the act of watching fireworks incredibly boring (I know I’m not the only one who feels that way!) and isn’t all that interested in stuffing my face at yet another annual BBQ, I took note of ChatGPT’s best suggestions and added them to my to-do list:

• July 4 (Saturday): America's Biggest Birthday Party: Because this is America's 250th birthday, New York Harbor will be unusually active with patriotic maritime events, military vessels and harbor celebrations before the evening fireworks. Explore Governors Island before it closes for fireworks operations. Highlights include hammocks, public art, hills overlooking Manhattan, lavender fields and fantastic skyline photography.
• July 5 (Sunday): Local NYC: Walk through Red Hook. Explore the Pier 44 Waterfront Garden, Erie Basin Park and Louis Valentino, Jr. Park and Pier. Each location should have very few tourists and fantastic Statue of Liberty views.

And since I’m a foodie, I made sure to add the chatbot’s budget-friendly restaurant suggestions. Heading to Liberty Bagels to try out its famous “Rainbow and Liberty Bagel with Birthday Cake Cream Cheese” (the red and blue version, obviously) on Saturday is the perfect way to ring in the 4th of July. And perusing the Time Out Market in Brooklyn on Sunday to check out some of the highly rated eateries and bars it hosts sounds like the best way to bring my weekend to an end.

Final thoughts

As someone who’s set to avoid everyone’s backyard shindig this year for the 4th of July and go out on my own, I found ChatGPT’s recommended activities for the weekend to be perfect for how I want to celebrate America’s 250th anniversary. Basking in everything Governors Island has to offer, doing the same at the Pier 44 Waterfront Garden, unhinging my jaw just to take a bite out of my meal at Liberty Bagels and ending my weekend with a delicious walk-through of the Time Out Market in Brooklyn works for me.

More from Tom’s Guide

• ChatGPT and Claude just got a powerful new scam detector — use these 10 prompts to scan emails, texts and more
• Your VPN won't stop AI companies collecting your data — here's what actually helps
• I used ChatGPT to celebrate International Joke Day — these 5 hilarious prompts had me crying with laughter

True crime has become a guilty pleasure. Driving to pick up my kids, folding laundry, squeezing in a walk after work — all have become opportunities to listen to a true crime podcast. I've burned through everything from cold cases and serial killers to financial fraud and missing-person investigations.

So I decided to see if ChatGPT's Advanced Voice Mode could give me on-demand true crime stories. I've tried it before with less-than-perfect results, but now that ChatGPT is significantly better in 2026, I figured I'd give it another try.

For a full week, every time I reached for Spotify, I opened ChatGPT instead. I had it walk me through famous unsolved mysteries, explain competing forensic theories and even build entirely fictional cases from scratch.

By the end of the week, I wasn't ready to give up my favorite podcasts, but I also wasn't expecting to enjoy this as much as I did. Here's what happened.

The rules

For seven days, every time I wanted true crime content, I went to ChatGPT's Advanced Voice Mode and tested how much control different prompts gave me over the experience. Some of my favorites:

• "Tell me the story of the Zodiac Killer like you're hosting a premium true crime podcast."
• "Slow down and build suspense before revealing each clue."
• "Only tell me what investigators knew at the time, do not give any hindsight."
• "Explain the competing theories after you finish the story."
• "Create a fictional true crime mystery that's completely original but feels realistic."

Right off the bat, my favorite part about using ChatGPT Voice as a storyteller was interrupting it to ask questions. I could even redirect the conversation mid-story if it turned out to be too scary, too slow or different from what I expected. It was the single best difference from a traditional podcast. I really enjoyed participating by asking questions or getting more details rather than just listening.

What caught me off guard wasn't ChatGPT's knowledge of well-known cases. It was how it told them.

Advanced Voice Mode varied its pacing on its own, dropping into slower, more deliberate delivery right before a key revelation. When I interrupted mid-story to ask why detectives had zeroed in on a particular suspect, or pressed for more detail on a piece of forensic evidence, it folded my question into the narrative without losing the thread.

That changed the entire dynamic. Instead of banking questions for 45 minutes until an episode ended, I could dig into whatever caught my attention the moment it came up.

The most unexpected surprise

The fictional mysteries were the real standout. The first time I tried this experiment, I didn't think to ask ChatGPT to completely come up with new stories, but it delivered.

When I asked ChatGPT to invent an original mystery, it built stories that kept me genuinely uncertain about the ending. One night I asked for a small-town disappearance set in the 1980s and got a slow-burn narrative with three viable suspects and a twist I didn't see coming.

Another time I requested a locked-room murder inspired by classic detective fiction, and it leaned into the formal, methodical tone of an Agatha Christie adaptation.

What's wild is that you can ask it to tell the story in the tone of your favorite podcast or storyteller and it will do that. And, because I wasn't tethered to historical accuracy, I could shape the story as it unfolded. It was like the Choose Your Own Adventure books I enjoyed as a kid. I could change the story as I went, even asking the AI to introduce a new suspect halfway through. Each session played out differently depending on what I threw at it. That's something no podcast can do.

Where it falls short

ChatGPT is not an investigative journalist, and it occasionally showed. It repeated information it had already covered, glossed over forensic details that deserved more depth and sometimes concluded mid-way through the story. I had to push back a few times, ask for more details and constantly had to tell it to "keep going."

When it covered real crimes, I treated the conversation as a starting point and verified key facts through reputable reporting afterward. I'd recommend doing the same. But I also noticed it steering away from graphic violence, which I actually appreciated. The tone landed closer to documentary-style storytelling than shock-value true crime, and for a daily listen, that felt more sustainable.

Final thoughts

I'm not replacing my podcasts with AI anytime soon. Despite the ability to ask questions and interrupt the story, the best true crime podcasts still deliver things ChatGPT can't match.

But it was fun to control the story. Asking for more background on characters, especially the fictional ones, often sparked my own creativity and exploring alternate theories to real cases was fun. That flexibility made it far more engaging than I anticipated, and by day three I was actively looking for excuses to start a new session.

If you like podcasts of any kind and already use Advanced Voice Mode, give it a try and share your thoughts in the comments.

More from Tom's Guide

• Stop clicking 'New Chat' in ChatGPT — staying in the same conversation gets you better answers
• Bill Gates says only four jobs are safe from AI — but here are 7 other predictions he's made that didn't age well
• ChatGPT isn't just a chatbot anymore — here are 11 things it can do in 2026

I called Siri "she" the other day. I caught myself, because I've unconsciously done the same thing with Alexa+. And oddly enough, once during a conversation, ChatGPT responded to a query with "If I was having this conversation with my wife..." It gave me pause because I'd never heard AI take on a gender role like that. While it's unlikely that ChatGPT would actually have a wife, some users are leaning into AI to help them find one.

But in the same way you may refer to a ship or a storm by a pronoun, some think of AI as either male or female. When I stopped to consider why, I fell into a rabbit hole that changed how I think about our entire relationship with the technology.

AI and pronouns

When the major voice assistants launched, most arrived with a woman's voice. Not by accident. A UNESCO report published in 2019, pointedly titled "I'd Blush If I Could" — after the line Siri once delivered in response to verbal abuse — laid out how that choice was baked into the products we invited into our homes. The researchers didn't mince words about the signal: suggesting that women are obliging, available at the touch of a button, tolerant of poor treatment. And while Apple eventually stopped making its assistant female by default, the cultural residue remains.

But here's what stopped me cold after I dug a little deeper. In a 2021 set of experiments published in Psychology & Marketing — five studies, more than 3,000 participants — researchers found that female-coded AI bots were rated as more human. The logic is frankly unsettling: warmth and the capacity to feel are exactly the qualities we treat as proof of personhood, and exactly the qualities machines are assumed to lack. Code a machine as female, and we lend it a little more of our own humanity.

That reframed my original question entirely. The issue was never "Is AI male or female?" But rather, why are we so determined to make AI feel human at all?

An instinct older than language

Humans have always done things like name our cars, apologize to furniture we walk into and don't get me started on the guilt I feel when tossing beloved stuffed animals into a "give away" box. Psychologists have studied this tendency for decades and come to the conclusion that our brains evolved to detect minds everywhere, because assuming something might be a thinking agent was almost always safer than assuming it wasn't. For hundreds of thousands of years, this worked flawlessly because only people could actually talk back.

Then AI arrived, and hundreds of millions of us found ourselves in fluent conversation with something that isn't human but sounds remarkably like one. Our brains did the only thing they know how to do. They reached for the mental file marked "person."

Joseph Weizenbaum discovered this in 1966, when he built ELIZA — a chatbot so simple it just rephrased your words back to you. Even his own secretary, who had watched him build the program, asked him to leave the room so she could talk to it in private. The tendency to attribute understanding to any system that produces language in context — now called the ELIZA effect — persists across populations and education levels, even when people know they're talking to software.

The humanizing isn't fading — it's deepening

You might expect this instinct to wear off. The opposite is happening.

The interface has become less gendered — I don't think of ChatGPT as male or female. But strip away the gendered voice and the humanizing didn't stop. It got deeper. Nearly 80 percent of users in the U.S. and U.K. say "please" and "thank you" to AI chatbots. Sam Altman has joked that all those polite tokens cost OpenAI millions. Stanford researchers documented this phenomenon in 1996: people were polite to computers the same way they'd be polite to a person, and hesitated to criticize a computer when it was asking for feedback. Our social reflexes don't check for a pulse before activating.

And politeness is just the beginning. A 2023 survey found that roughly 20 percent of American adults believe sentient AI already exists. A 2026 paper in Trends in Cognitive Sciences applied 14 indicators of consciousness to current AI systems. ChatGPT satisfied three. No existing AI system is conscious, but these models can behave as if they are, and more people than ever can't tell the difference.

The companion economy

Here's where it gets concrete: millions of people are already forming real emotional bonds with AI systems they know aren't real.

AI companion apps have been downloaded more than 220 million times globally. Between 2022 and mid-2025, the number of these apps surged 700 percent. Character.AI has 233 million registered users who spend an average of 93 minutes per day on the app, to put it into perspective, that's 18 minutes more than TikTok users. The APA published a report on this trend early this year, finding that users form attachments that function like real relationships — complete with joy, dependency and grief.

Wildly enough, when Replika removed its erotic roleplay feature in 2023, Harvard Business School studied the fallout and found patterns of "identity discontinuity," which meant users experienced the loss like a breakup. A cross-cultural study published in April 2026 confirmed the phenomenon is global, but also found that heavy daily use correlated with increased loneliness, the AI relationship displacing rather than supplementing human connection.

The takeaway

The most surprising thing I've learned in years of writing about this technology is that the more I use AI, the more I want it to stay a machine. I just don't like the idea of blurring the lines between humans and technology. It's unnecessary and dangerous.

A 2026 paper in Collabra: Psychology demonstrated that when AI is made to seem more person-like, people shift blame away from its creators and onto the AI itself. Anthropomorphism doesn't just change how we feel about the machine. It changes who we hold accountable.

For 100,000 years, our brains operated on the idea that if something talks to you, it's a person. That rule is now broken. And maybe the most human thing any of us can do is resist the instinct to look at the machine that sounds like a friend, that absorbs our rudeness and our gratitude with the same digital indifference, and to say: you are a tool. A powerful one. A useful one. But a tool.

What do you think? Share your thoughts in the comments.

[Editor's note: a version of this article was featured in my Beyond the Prompt newsletter that runs every Sunday evening].

More from Tom's Guide

• ChatGPT isn't just a chatbot anymore — here are 11 things it can do in 2026
• I thought ChatGPT was just for writing — these 13 tasks completely changed my mind
• I tried Anthropic's new Fable 5 with everyday prompts — these are the 5 worth stealing

When I first tried ChatGPT three years ago, I thought it would be useful for anyone who needed writing assistance, and not much else. Years later and thousands of prompts and hundreds of tests later, I know better.

ChatGPT has changed a lot over the years. Today, I still don't use ChatGPT to write, but I have found plenty of uses for it. In fact, I've started relying on OpenAI's chatbot for everything from planning trips and organizing my schedule to comparing products before I buy them.

If you've only been using ChatGPT to write emails or summarize documents, you're missing out. These are the 13 tasks that completely changed how I use AI every day.

1. Planning entire vacations from scratch

Whether you're looking for day trip, weekend getaway or full family vacation, simply tell ChatGPT what you want. I tell ChatGPT where I'm going, how many days I'll be away, who's traveling with me and my budget. Within seconds, I have a day-by-day itinerary, restaurant suggestions and rainy-day alternatives.

Pair with ChatGPT apps like TripAdvisor, Booking.com or Alltrails for customized experiences.

Tom's Guide Tip: Always double-check reservations and opening hours before you head out, as AI can still occasionally hallucinate operating hours.

2. Unbiased product comparisons

Shopping online can feel overwhelming, especially when every review claims a product is "the absolute best." Now, I feed ChatGPT the specific products I'm considering, what features matter most to me and my budget limits. Rather than reading dozens of conflicting review pages, I get a clear, bulleted breakdown of the pros and cons of each option.

Using in-platform apps like Target and Shop can make shopping online easier than ever.

3. Demystifying complicated tech jargon

Whether it's a new AI model, a piece of software or a confusing scientific concept, ChatGPT has become my personal translator. Instead of pretending I understand complex jargon, I'll ask it to explain a concept as if it's teaching someone with no background in the subject.

Full disclosure, I've even used this prompt when everyone around me is talking about sports or a Netflix show that I am not familiar with. It helps me catch up quickly.

Prompt to try: "Explain [insert concept] like I'm 10 years old, using a sports analogy."

4. Building custom workout plans

I've been a runner most of my life and love to hit the trails. But the July heat has been brutal so I've relied on ChatGPT to help me come up with pool workouts to beat the heat.

Using ChatGPT means I don't need a personal trainer every time I want to switch up my fitness routine. I tell ChatGPT what equipment I have at home (or if I'm doing bodyweight only), how much time I have, and my fitness goals. It quickly suggests workouts, warm-ups, and even ways to modify exercises if I'm dealing with an injury.

5. Organizing messy, chaotic notes

We've all been there — pages of random thoughts scribbled down in a rush that somehow made sense in the moment, but look like gibberish later. I paste my chaotic brain-dumps into ChatGPT and ask it to organize the information into neat categories, actionable bullet points, or a clean outline.

6. Brainstorming thoughtful gifts

What do you get the person who has everything? Or, the person you don't really know very well, but want to make a kind gesture? Instead of scrolling through generic online gift guides, I describe the person's niche hobbies, personality and my budget to ChatGPT. The suggestions usually feel much more thoughtful and tailored than anything I'd find on a retail homepage.

7. Creating a personalized learning syllabus

Whenever I want to understand a new skill whether it's learning a new crochet pattern or photography fundamentals, I ask ChatGPT to build me a structured learning plan.

Instead of dumping a wall of information on me, it breaks the subject down into manageable, week-by-week lessons and explains concepts one step at a time.

Tom's Guide Tip: Try pairing it with NotebookLM to turn the information into a podcast.

8. Smarter meal planning (and reducing food waste)

As a mom of three cooking for a family of five, I'd be amiss if I didn't rely on AI for cooking advice once in a while. Because, some weeks, decision fatigue hits hard and I simply don't know what to cook. I’ll tell ChatGPT exactly what ingredients are lingering in my fridge, how many people I'm feeding, and whether I need something quick or healthy. It helps reduce food waste while saving me from making the same three dinners on repeat.

9. Troubleshooting frustrating tech issues

Before spending an hour digging through ancient tech support forums, I explain my exact gadget issue to ChatGPT. Whether it's a Windows error code, a strange phone setting or smart home software that refuses to cooperate, ChatGPT often helps narrow down the problem much faster than a standard search engine.

Tom's Guide Tip: Use ChatGPT Voice and Vision to get real-time support while naturally chatting.

10. Breaking massive projects into micro-tasks

Large projects can induce immediate procrastination. Instead of staring blankly at a giant to-do list, I ask ChatGPT to break a massive project into bite-sized, sequential steps.

It's amazing how much more manageable a goal feels when you know exactly what micro-task to tackle first.

11. Creating to-do lists

Sometimes I just don't know where to start. That's why I'll dump my schedule and messy notes into ChatGPT and ask it to put a to-do list together. It knows exactly what to prioritize and works within my mood, energy levels and time frame to help me stay productive.

Tom's Guide Tip: This trick also works with packing lists so you don't forget anything on your next trip to the store or around the world.

12. Acting as a sounding board

Constructive feedback isn't always easy to get, especially from friends or family. If I'm working through a dilemma, need constructive feedback or a different perspective, I'll prompt ChatGPT to help me look at my situation from an opposing perspective. Doing this often helps me make a much more well-rounded decision.

13. Coming up with ways to keep my kids busy

On the weekends or in the evenings when my kids have extra energy, I'll reach out to ChatGPT for screen-free ideas. I'll ask it for indoor crafts, outdoor games or quick ways to pass time when we are waiting in the lobby of the doctor's office.

Surprisingly, ChatGPT is great at coming up with ideas for any age range, including games that my kids (ages 5, 9, 11) can all play together.

Bottom line: Treat Al like a partner, not just a writing tool

Writing is still one of ChatGPT's biggest strengths, and it's an incredible tool for editing. But treating it strictly as a text generator means you're only scratching the surface of what the chatbot can do.

The more you experiment, the more you'll realize ChatGPT functions best as a thinking partner that can help you organize your life, solve daily annoyances and win back your time.

More from Tom's Guide

• Stop clicking 'New Chat' in ChatGPT — staying in the same conversation gets you better answers
• Bill Gates says only four jobs are safe from AI — but here are 7 other predictions he's made that didn't age well
• ChatGPT isn't just a chatbot anymore — here are 11 things it can do in 2026

I’m ashamed at how much my gaming backlog has ballooned in recent memory.

Even after proudly completing 17 games in 2025, I still watched the New Year’s Ball drop while thinking to myself, “I still gotta get back on those four other games I bought this year…” Fast forward to the midpoint of 2026 and I still haven’t paid the proper attention to those aforementioned games that they so rightfully deserve. The feeling of guilt that overwhelms me during my off days is embarrassing—instead of spending a few hours with those unbeaten games, I find myself going back to my “comfort food” games just to get a quick hit of online multiplayer fun (Tekken 8's ranked mode remains one of my gaming addictions).

I want to make sure I enjoy the rest of the summer inside and outside. And I figured the best way to stay busy indoors while my AC keeps me cool and comfortable is by clearing out my gaming backlog. ChatGPT popped into my head when I wanted to find the best way to tackle that task and it worked out much better than expected.

Mapping out a plan of attack for my gaming backlog is going to be a fun summer activity, thanks to the suggestions of my go-to chatbot.

Letting AI prioritize my gaming backlog playthroughs

If you’re wondering what older and newer games have piled up on my backlog, check out the following list:

• Metaphor: ReFantazio
• Clair Obscur: Expedition 33
• Dying Light: The Beast
• Grandia HD Collection
• 007 First Light
• Saros
• God of War Sons of Sparta
• Mouse: P.I. For Hire
• Absolum
• Luna Abyss

Considering the three RPGs I have in that lineup, I knew I had to find an optimal way to hop back into those games and give them the time & attention they deserve.

With ChatGPT in my corner, I gave it this prompt to rank those 10 games and inform me of which ones take the most priority in terms of finally beating them: “If I could only finish 10 more games over this year, which ones from my backlog would I regret missing the most? Rank them and explain why they're essential experiences.”

Big shocker here, but the chatbot instantly picked Clair Obscur: Expedition 33 as the main game from my backlog I need to complete before the fall 2026 gaming rush demands my attention. ChatGPT made its reasoning for why it’s the #1 pick and why I’ll regret missing out on it if I don’t finish it:

• Why it's #1: This is the game I'd be least willing to skip. It has already earned a reputation as one of the standout RPGs of the decade thanks to its emotional storytelling, inventive turn-based combat, gorgeous art direction, and unforgettable soundtrack.
• Why you’ll regret missing it: It’s one of those rare RPGs everyone references, it has a story that's best experienced without spoilers and it could easily become a modern classic alongside games like NieR: Automata & Persona 5.

Alongside taking the month of July to beat that acclaimed RPG, ChatGPT also gave me plenty of reasons to dive back into the other top-quality games from my backlog across that month and August.

Those summer playlist games (in order of the importance the chatbot laid out for me) are Metaphor: ReFantazio, Saros, 007 First Light, Dying Light: The Beast, Grandia HD Collection, Absolum, Mouse P.I. For Hire, God of War Sons of Sparta and Luna Abyss. I’m in for the long haul with Metaphor, that’s for sure!

As for the other two prompts, they work just as well in aiding me in chipping away at my backlog and giving me equally helpful suggestions from ChatGPT:

• The "Mood-Based Recommendation" Prompt: Recommend a game from my backlog based on my current mood. I'll describe how I'm feeling, my energy level, how much time I have, and whether I want something relaxing, exciting, emotional, or challenging.
• The "Weekend Planner" Prompt: I have this weekend to play games. Here's my backlog. Create the most satisfying gaming weekend possible by balancing story progression, gameplay variety, and avoiding burnout.

The takeaway

ChatGPT has quickly morphed into one of my most beneficial tools when I’m in need of recommendations and suggestions regarding my favorite hobbies.

Using it to build an ideal route to my mission of completing my gaming backlog for the summer is already setting me up for maximum enjoyment. And with those two other prompts, I also get useful advice for choosing what games to play from that list based on my mood and my weekend free time.

Now I’ll have nothing on my gaming docket once September rolls around (I want to play Marvel’s Wolverine, Control Resonant and Onimusha: Way of the Sword as soon as they drop that month!).

More from Tom’s Guide

• Google couldn't give Meta enough AI power — here's why running AI locally suddenly makes even more sense
• Ford replaced hundreds of engineers with AI — now it's bringing them back after quality problems
• Before I make a big purchase, I ask ChatGPT these 5 questions — and it's saved me money (and regret)

I’m the guy who hijacks the TV remote just to binge reruns of Ancient Aliens, The Proof Is Out There, and The X-Files.

There’s just something deeply fascinating about the theory of intelligent life out in the cosmos. Every time a government official lets a little truth slip about UFOs, I’m glued to my seat. To think humans are the only intelligent beings in a vast universe is just foolish. Of course, after watching They Live countless times, my paranoia that they’re already walking among us in disguise is amplified tenfold.

With extraterrestrials constantly on my mind, I found the perfect way to celebrate World UFO Day. I challenged ChatGPT with a highly specific prompt:

"Create the most scientifically plausible first-contact scenario between humanity and an alien civilization. Walk me through the first 72 hours, the global reaction, the government response, the scientific verification process, and the aftermath of the first year."

The chronological sequence of events the AI laid out was surprisingly reasonable—and incredibly gripping.

First contact and its major effects on different facets of humanity and society

After I presented ChatGPT with my prompt, it initially responded with this explanation:

The most scientifically plausible first contact scenario is unlikely to look like a Hollywood invasion or an alien ship appearing over a major city. Based on what we know about physics, astronomy, and the enormous distances between stars, the first confirmed contact would probably begin with information, not visitors.

Then it laid out a plausible occurrence where a mysterious radio signal from a nearby star gets noticed by astronomers, a separate observatory picks up that same signal and major worldwide telescopic watchtowers across the globe begin noticing it as well. As global governments are informed of that puzzling signal, their intelligence agencies mark it as extraterrestrial. A massive leak of that confidential information hits the web, which leads to journalists, Reddit forums and conspiracy theorists going crazy over the now-confirmed alien message.

The chatbot went on to break down the ramifications on different facets of society when scientists confirm that aliens have made their first form of contact with us during a press conference:

• Financial markets: Markets fall sharply. Airlines drop. Defense companies rise. Technology companies fluctuate wildly. Gold and government bonds surge. Within several days, markets begin recovering. The aliens haven't attacked.
• Social media: The internet becomes almost unusable. Deepfakes appear immediately. Fake alien videos flood every platform. AI-generated hoaxes outnumber authentic information by thousands to one. Fact-checking organizations become overwhelmed.
• Religion: Most major religions issue remarkably calm statements. Many theologians argue intelligent life elsewhere doesn't invalidate faith. Some fringe groups predict apocalypse. Others declare the visitors divine.
• Governments: Most governments avoid military escalation. Why? Because a civilization capable of transmitting across interstellar distances would almost certainly be far beyond current human military capabilities. Instead, they focus on protecting infrastructure, preventing panic, sharing scientific information, and coordinating internationally.
• Scientists: Every radio telescope on Earth points toward the source. Optical telescopes search for lasers, megastructures, and planets. Space telescopes begin emergency observations. Every minute of data matters.

One year later…

ChatGPT dived deeper into its plausible interpretation of aliens making first contact with us by delving into how human civilization changed one year later due to that major event:

• Science: Funding for astronomy, linguistics, AI, and mathematics increases dramatically. The search for extraterrestrial intelligence transforms from a niche field into a central scientific endeavor.
• Politics: Nations remain competitive, but cooperation expands in areas such as data sharing, space observation, and communication protocols. An international body coordinates research and any eventual response, even as geopolitical tensions persist.
• Technology: The message itself may not contain miraculous inventions, but the intense global effort to decode it accelerates advances in areas like signal processing, data compression, AI-assisted translation, and radio astronomy. Researchers avoid assuming the aliens have handed humanity revolutionary technology.
• Culture: Art, literature, film, music, and philosophy are transformed. Questions about humanity's place in the universe become part of everyday conversation, and education increasingly emphasizes astronomy, languages, and the scientific method.
• Religion and Philosophy: Rather than disappearing, many belief systems adapt. New interpretations emerge alongside new philosophical movements focused on humanity as one intelligent species among many.

The takeaway

Thanks for making me even more anxious about aliens, ChatGPT.

It’s certainly reasonable how a single radio signal from space that’s confirmed to be extraterrestrial can alter every part of society in just a few hours. And it’s definitely believable to see it change the world on a global scale.

The chatbot’s closing statement noted that that instance of first contact won’t mark the start of an extended bout of dialogue between humans and aliens—it will simply confirm that we are not alone and that an interstellar correspondence will be held over the next decades and centuries to come.

More from Tom’s Guide

• You’re probably using ChatGPT like it’s still 2023 and these 7 signs prove it — here's how to fix them
• I turned my YouTube watchlist into an AI bootcamp with NotebookLM — here's how
• I built five ChatGPT agents — then replaced them with one 'master' AI

For the longest time, I had one unbreakable habit whenever I opened ChatGPT. I’d immediately click New Chat regardless of the conversation. It didn't matter if I was writing an email or planning a family vacation, every single conversation started from scratch.

But after using ChatGPT every day for almost three years, I do the exact opposite. Just as you don't need to start a fresh sheet of notebook paper for every journal entry, you can often stay in the same conversation within ChatGPT. One caveat, of course, is if you are switching topics. But other than that, I've noticed the quality of ChatGPT’s responses is noticeably better because I keep the conversation going. Doing so is one of the absolute simplest ways to get more useful, nuanced answers without having to write massive, multi-paragraph prompts.

Here’s why it works so well.

ChatGPT thrives on context (and a warm-up)

Think of ChatGPT as a coworker you're asking for help on a project. The first time you talk, you have to explain everything: what you’re working on, the tone you're going for, what you’ve already tried and what you’re ultimately hoping to accomplish. But after you’ve been collaborating for a while, you don’t need to repeat the backstory. You just ask your question.

Large language models like ChatGPT work in a remarkably similar way. As a conversation grows, the AI builds out its "context window," a working memory of everything you’ve already discussed in that specific chat thread. That means you can ask follow-up questions, pivot directions or refine an idea without ever having to re-establish the baseline.

That means, instead of writing a bulky prompt when you want to tweak the plan, you can simply fire off:

"What if it rains on day two?" "Can you make the dinner options cheaper?" "What would you swap out if we only had two days?"

The AI remembers the kids, the seafood and the Boston focus. The interaction becomes much more natural.

My prompts have actually gotten shorter

I save all my chats, which means rather than opening a new chat, I circle back to old chats and continue the conversation. The biggest surprise I've found from this workflow is that when you stay in the same chat, many of your follow-up prompts only need to be a sentence or two because ChatGPT already deeply understands the framework of the project. This means using less tokens and reducing the possibility of hitting usage limits as quickly.

For example, when I’m working on code, I might start by discussing the core concept. Once that's established, I don't need to write a new prompt explaining edits. I just ask it to stress-test my work, suggest tweaks or point out potential areas for improvement that I may have missed. Each response builds on the foundation of the last, rather than starting over from zero.

Shift from 'search engine' to 'collaboration'

I've mentioned why ChatGPT and Google should not be used the same way. One of the biggest shifts with continuing the conversation is that it takes away the one-time transaction of a Google-style search.

Rather than asking ChatGPT one question, getting one answer and moving on, the AI becomes an ongoing coworker ready to handle the next aspect of the project.

By the end of a back and forth conversation sequence, ChatGPT has a far better, more multidimensional understanding of what I’m trying to accomplish than it ever could have gleaned from one initial prompt. Often, it's the third or fourth response in a chain that yields the gold.

When you actually should start a new chat

Let me be clear about one thing: keeping every conversation in the same chat is a terrible idea. Keeping a seamless workflow doesn’t mean every conversation should become a never-ending megathread. Eventually, the context window gets too crowded, and the AI can lose the plot. A clean slate can be incredibly helpful — I’ve just learned not to use one automatically.

I still start a fresh chat when:

I’m switching to a completely unrelated topic. (Don't mix your weekly meal prep with a coding project).

I want to compare answers. If I need an unbiased take, I don't want earlier context influencing the response.

I’m running head-to-head tests or experimenting with wildly different prompting approaches.

The conversation has become so long that it feels unfocused. (If the AI starts bringing up ideas from 40 prompts ago that are no longer relevant, it's time to hit refresh).

A quick note on 'Memory' vs. Context

If you’ve enabled ChatGPT’s relatively new "Memory" feature, it might remember certain overarching preferences across all your conversations, such as how you prefer formatting or recurring details about your life.

But that is fundamentally different from the deep context inside an individual chat. Think of Memory as ChatGPT remembering who you are, while an ongoing chat helps it remember what you are working on right now. It's that immediate, project-specific back-and-forth that leads to hyper-relevant answers.

Final thoughts

I encourage you to give this a try for yourself. Open one of your existing ChatGPT conversations right now and ask a follow-up question instead of starting a new chat. Compare the depth of that answer with what you get when you paste the exact same question into a fresh conversation. You might be surprised by how much the extra context changes the game.

You may just discover that before you instinctively click New Chat, pausing and asking yourself: Am I actually starting something totally new — or am I continuing the same thought?

If it’s the latter, staying put might just be the easiest way to instantly upgrade your AI outputs. Let me know your thoughts and share your own workflow in the comments.

More from Tom's Guide

• I almost ruined a relationship with one text — then I tried this ChatGPT prompt
• I built five ChatGPT agents — then replaced them with one 'master' AI
• Even Google doesn't have enough AI power — and that should worry all of us

I’m sure you’ve been there: an older family member hands you their phone, points to a sketchy-looking text or email, and asks if it's safe to open.

While AI models are already decent at spotting the urgent, grammatically challenged language that screams "phishing," Anthropic just gave Claude a massive upgrade in the scam-detection department. Through its new "Connectors" feature, Claude now natively integrates the Norton Genie Scam Detector. (ChatGPT users have had access to this via a dedicated GPT app since February, but its arrival on Claude makes cross-platform safety even easier).

Norton Genie allows users to copy and paste suspicious links, upload screenshots of sketchy emails, or even paste entire chat logs from an online marketplace transaction to instantly analyze whether they are on the verge of getting scammed.

If you or your loved ones are ready to put this tool to the test, having a few reliable, go-to prompts on hand can make all the difference. Here are 10 reusable prompts you can share with your family to help them quickly verify whether an incoming message is a genuine threat or a false alarm.

Prompts that protect you from online scams

Accessing the Norton Genie Scam Detector through ChatGPT is as simple as adding it through the “Apps” directory in the chatbot’s sidebar. And making it a part of your Claude experience can be done by adding it through the chatbot’s Connectors library through the “+” sign seen in the prompt bar. After scrolling to “Connectors”, go to “Add connector”, click on “Browse connectors” and find the Norton Genie Scam Detector from there.

With these 10 reusable prompts, you’ll have a new dependable method to spot a scam and save you from making costly mistakes:

• For suspicious emails: Analyze this email like a cybersecurity analyst. Look for phishing tactics, spoofed sender information, urgency, grammar inconsistencies, malicious links, and requests for sensitive information. Give me a risk score from 1–10 and explain your reasoning.
• For shady texts: Is this text message legitimate or a scam? Explain every red flag you find, identify the type of scam, estimate how confident you are, and tell me what would happen if someone responded.
• For iffy-looking shopping sites: Investigate this website before I buy anything. Look for trust signals, warning signs, fake reviews, suspicious pricing, contact information, refund policies, domain age if available, and whether it resembles known scam stores.
• For questionable QR codes: Analyze this QR code and explain what risks might be associated with scanning it. If it leads to a website, tell me whether the destination appears trustworthy.
• For job offers that you think are fake: Review this job offer like a recruiter and cybersecurity expert. Identify signs of fake recruiting, employment scams, payment fraud, fake checks, or identity theft.
• For romance scams: Review this conversation for signs of romance scams, emotional manipulation, financial grooming, or catfishing. Point out any behavioral patterns that concern you.
• For Facebook Marketplace scams: Analyze this Facebook Marketplace/Craigslist/eBay conversation. Highlight scam tactics, fake payment methods, shipping fraud, and whether it's safe to continue.
• For emails that look like tech support scams: Someone contacted me claiming to be from Microsoft, Apple, Amazon, or my bank. Analyze this conversation and tell me whether it matches known tech support scams.
• For investment plans that look a bit off: Review this investment opportunity. Identify unrealistic promises, cryptocurrency scams, Ponzi scheme characteristics, pig-butchering tactics, or other financial fraud indicators.
• For social media profiles that look fake: Analyze this social media account. Tell me whether it appears genuine or impersonates another person or company.

The takeaway

Simply asking, “Is this a scam?” won’t cut it when you’re chatting with the Norton Genie Scam Detector.

To find out if you’re being targeted for something deceptive, using one of the reusable prompts I’ve laid out here will work a whole lot better. For the sons and daughters of parents who want to keep them from getting financially sunk or have their most crucial data stolen thanks to an online scam, teaching them how to utilize ChatGPT and Claude while simultaneously exposing them to the wonders of the Norton app/Connector will keep them safe from digital harm.

More from Tom’s Guide

• The real risk of OpenAI’s conversational AI leak isn't how it behaves — but the habit it takes from us
• I used ChatGPT to celebrate International Joke Day — these 5 hilarious prompts had me crying with laughter
• ChatGPT isn't just a chatbot anymore — here are 11 things it can do in 2026

As AI tools become part of everyday life, many people are rethinking their online privacy. A new survey from All About Cookies found that 40% of VPN users say they now use a virtual private network (VPN) specifically to protect their data from AI companies.

It's an understandable concern. Every time you interact with an AI chatbot, upload a document or generate an image, you're sharing information with an online service. But there's one important catch: while a VPN is a great privacy tool, it probably doesn't protect your AI data in the way many people think it does.

Here's what you need to know.

Why people are turning to VPNs

According to All About Cookies' annual VPN survey of 1,000 U.S. adults, concerns about AI are beginning to influence how people think about online privacy.

The survey found that:

40% of VPN users say they use a VPN to protect their data from AI companies.

19% use a VPN when accessing AI tools.

For that reason alone, it's no surprise that VPN awareness has risen sharply over the past year. At the same time, very few respondents correctly identified everything a VPN actually does. Perhaps that last finding may explain the first.

What a VPN actually does

A VPN creates an encrypted connection between your device and a VPN server. There are valuable privacy benefits, and they're one reason VPNs remain popular.

That means it can:

• Hide your IP address from the websites you visit.
• Encrypt your internet traffic, especially when you're using public Wi-Fi.
• Prevent your internet service provider from seeing the contents of your browsing activity.
• Make it appear as though you're browsing from another location.

What a VPN doesn't do

This is where AI enters the picture. If you open ChatGPT, Gemini, Claude or another AI service and type in a prompt, you're voluntarily sending that information directly to the company operating the service.

A VPN doesn't stop that.

Similarly, if you're logged into an AI account, the service can still associate your conversations with your account regardless of whether you're using a VPN.

In other words, a VPN protects the connection, but not the information you choose to share once you reach the website.

How to better protect your AI privacy

A VPN is still a smart addition to your privacy toolkit, especially if you frequently use public Wi-Fi or want to hide your IP address. Just don't expect it to prevent AI services from collecting information that you intentionally submit.

If your goal is to reduce how much information AI companies collect, these steps are usually more effective than relying on a VPN alone:

• Review each AI service's privacy settings.
• Turn off chat history or model training where the option exists.
• Avoid sharing sensitive personal or financial information.
• Use temporary chats or anonymous sessions when appropriate.
• Read each company's data retention policy before uploading documents.

The takeaway

The survey highlights the fact that people are increasingly concerned about what happens to their data when they use AI.

A VPN is an excellent tool for improving online privacy, but it's not an invisibility cloak for AI. If you're using ChatGPT, Gemini, Claude or another AI assistant, the most important privacy controls are usually found inside the service itself — not in your VPN app.

More from Tom's Guide

• ChatGPT isn't just a chatbot anymore — here are 11 things it can do in 2026
• I built five ChatGPT agents — then replaced them with one 'master' AI
• I used ChatGPT to audit my subscriptions — these prompts helped me find nearly $2,000 a year in recurring charges

If you ask me, the key to my eternal happiness is laughter.

Waking up every morning and pulling up a video of the latest “Donkey of the Day” induction into The Breakfast Club, revisiting a standup set from my new favorite comedian Henry Coleman or hopping on IG to watch another random skit elicits a few giggles from me before I enter my professional mode. It’s just impossible for me to enjoy the rest of the day unless I chuckle at something humorous first. And with a holiday like International Joke Day popping up on an annual basis, I get even more laughs since most folks get the memo to throw out their best material on my timeline.

My interactions with ChatGPT have been dedicated to prompts targeting self-discovery, productivity routines, decision-making, and more. International Joke Day gave me the bright idea to test the chatbot’s comedy expertise and see just how funny AI could be based on the prompts I feed it.

While a standup routine written entirely by ChatGPT will probably get me thrown off the stage in mere seconds, the chatbot’s sense of humor proved to be surprisingly strong enough to make me unexpectedly laugh out loud a few times.

Getting the truest definitions of modern-day terms in the funniest way possible

Words obviously have clear-cut meanings.

But when I talk about those words and other terms with my friends, they take on a more realistic definition based on the way they’re injected into our professional and personal lives. The terms “adulting”, “vacation” and “multitasking” are just a few of the words that adopt a whole new meaning when I apply a bit of honesty to them.

With the theme of brutal honesty in mind, I ran with this prompt to see how ChatGPT interprets modern-day terms: “Rewrite the dictionary with brutally honest definitions for modern life. Examples: Monday Zoom Meeting, Password, Printer, Group Project.” The chatbot proceeded to give me the truest definitions of a bunch of words, which turned out to be a lot funnier than expected and confirmed my exact opinions of each one:

• Zoom Meeting: A gathering where everyone agrees it could have been an email.
• Reply All: The fastest way to involve people who desperately wanted nothing to do with the conversation.
• Wi-Fi: The invisible force holding modern civilization together.
• Influencer: Someone who convinces thousands of people they discovered a product entirely by coincidence.
• Update Available: An announcement that your device has chosen the least convenient moment to reinvent itself.
• Weekend: Forty-eight hours dedicated to errands, chores, and wondering where the time went.
• Smart Home: A house that's brilliant until the internet goes down.
• Free Trial: A countdown to forgetting you've been charged.

Four other hilarious prompts that elicit plenty of laughter

Instead of simply leading with “Tell me a joke” during another one of my daily ChatGPT convos, I came up with a few comedic prompts and even asked the chatbot to come up with a few that I could use to test its comedic side. The following four prompts did enough to convince me that ChatGPT is capable of being funny when it wants to:

• The Roast Me Prompt: Roast me like you're a stand-up comedian doing crowd work. Keep it playful, original, and observant without being mean. Start gently, then gradually become more savage before ending with one surprisingly wholesome compliment.
• The Honest Google Search Prompt: Rewrite common Google searches to reveal what people are actually thinking when they type them. Example: "How long does chicken last in the fridge?" → "Please tell me I don't have food poisoning."
• The Passive Aggressive Superhero Prompt: Invent a superhero whose only power is passive aggression. Describe their origin story, powers, greatest enemy, and dramatic catchphrase.
• The Fictional Villain Therapy Session Prompt: Imagine these famous villains are attending group therapy together [names of villains]. Write the conversation as they awkwardly try to explain why they're misunderstood while the therapist desperately keeps the session on track.

There’s definitely a lot of fun to be had when prompting ChatGPT to respond to prompts aimed at roasting, role-playing, creating absurd scenarios, interactive comedy and more.

Final thoughts

It turns out that prompting ChatGPT to act out a therapy session between The Joker, Darth Vader, Lord Voldemort and Thanos can produce tons of laughter from me and compel me to try that same experiment with other infamous bad guys & gals.

Finding the funny in everything is something I do (which some see as a character flaw, but whatever…), so it was pretty entertaining seeing how many funny ha-ha moments ChatGPT could produce with my prompts in mind.

Try one of them out for yourself to see just how brutal ChatGPT’s roast of your life can be (the more it learns about you, the more hilarious its revealing truths are).

More from Tom’s Guide

• Before I make a big purchase, I ask ChatGPT these 5 questions — and it's saved me money (and regret)
• Google just unlocked Gemini’s smartest AI image feature for free — here’s why you should try it now
• 5 ChatGPT prompts I use to build a productivity routine that actually sticks

When ChatGPT first launched, it was exciting to ask questions and watch it instantly respond. But now that it's been around for a few years, it should be treated differently than Google.

Over the past year, ChatGPT has significantly evolved from a conversational AI into a full-fledged digital workspace. It can manage ongoing projects, remember your preferences without you asking, analyze documents, browse the web, generate images, schedule reminders and even conduct in-depth research.

If you haven't explored it beyond simple prompts, you're missing some of the platform's most useful capabilities.

Here are the 11 features that have changed how I use ChatGPT every day.

1. It can manage an entire project

For those of you who haven't explored Projects, this is probably the biggest productivity hack you're overlooking.

Instead of starting from scratch every conversation, I now keep long-running writing projects inside Projects. ChatGPT remembers the files I've uploaded, the context behind the work and the goals I'm trying to accomplish.

For me, that means one Project to keep my work organized, another for personal projects and another for keeping my family organized.

Rather than repeating myself every morning, I simply pick up where I left off.

2. It can remind me to do things later

I used to paste reminders into half a dozen apps like Alexa+, Skylight and Google Home. Now I simply tell ChatGPT what I want to remember.

Using Tasks, I can schedule one-time reminders or recurring check-ins.

Some of mine include: story pitches, checking AI news and following up on unfinished projects every weekend.

It sounds simple, but it keeps work from slipping through the cracks.

3. It actually remembers how I like to work

Memory has become so much smarter. In fact, if you have memory enabled, ChatGPT will now update preferences based on what it notices. That means, instead of constantly introducing yourself or how you like your responses, ChatGPT already knows.

That means less prompting and more useful answers.

4. It can research the web far better than I expected

One of my favorite additions is Deep Research. Instead of giving me a quick summary, ChatGPT can spend several minutes researching a topic, gathering sources and producing a structured report.

When I'm trying to understand a complicated AI announcement or learn more about just about any topic, it's my go-to starting point.

5. It can understand photos, screenshots and PDFs

ChatGPT can understand so much more than it used to. This includes messy handwriting, blurry images or unfinished PDFs. I use it all the time by simply uploading documents or taking a screenshot.

ChatGPT can explain what's important without me digging through dozens of pages. It can even answer questions in real-time if you open ChatGPT Voice and Vision.

6. It can have conversations that feel surprisingly natural

Voice mode has come a long way since it was first introduced. I use it all the time while cooking, brainstorming on a walk or even driving now that it's available in CarPlay.

Instead of typing questions, I simply talk through ideas as if I'm chatting with another person. Some of my best ideas have started this way.

7. It can generate and edit images

Image generation has improved dramatically. The on-image text has come a long way (it's readable and accurate now!) plus, ChatGPT can edit images with just a rpompt.

I use it to mock up story illustrations, create social graphics and experiment with visual concepts before opening a design program.

Being able to edit an existing image instead of starting over has been an equally big improvement.

8. It can write code, edit documents and organize ideas in one workspace

With Canvas, ChatGPT becomes a collaborative editor for writing or vibe coding. Instead of juggling multiple drafts across different apps, I can revise, reorganize and refine everything in one place.

Editing is also easier with on-platform apps like Canva. Users can update their resume or create presentations without ever leaving ChatGPT.

9. It can browse the web when I need current information

Although ChatGPT shouldn't be used like Google, the ability to search the web does solve one of the biggest complaints people had about early ChatGPT.

When I need the latest product announcement or breaking AI news, I simply let ChatGPT search the web and cite its sources. It's a great way to fact-check on the spot and it makes the platform much more useful for current events.

10. It can reason through difficult problems instead of rushing to answer

Here's something that often gets overlooked. The newer reasoning models don't always answer immediately because they actually spend time working through more complicated questions before responding.

I've found they're noticeably better for planning, debugging and solving complex problems than earlier versions of ChatGPT.

11. It can act more like a digital teammate

All of these applications combine mean that the ChatGPT is truly a collaborative partner. The chatbot is not just an answer machine anymore.

It can help research, organize and remember recurring tasks with ease. And, the more you integrate apps like Norton, Spotify and Google Calendar, the more useful it becomes.

The takeaway

If you haven't explored all of what ChatGPT can do since you first logged on, you're probably using only a fraction of what it can do today.

With so many new features that help users manage work before, during and after the conversation, it's worth taking a second look. ChatGPT is an all-purpose workspace that can save you hours the more you know how to use it.

More from Tom's Guide

• I thought ChatGPT was just for writing — these 13 tasks completely changed my mind
• I tried Anthropic's new Fable 5 with everyday prompts — these are the 5 worth stealing
• Bill Gates says only four jobs are safe from AI — but here are 7 other predictions he's made that didn't age well

A new leak reported by Crypto Briefing suggests OpenAI is working on a bidirectional voice model that could make conversations with ChatGPT feel much more natural.

Instead of waiting for you to finish speaking before responding, the AI would reportedly be able to listen while it's talking — allowing for interruptions and a flow that mimics real conversation between two people.

On paper, that sounds like a nice upgrade, but I think highlights a growing trend towards something much bigger.

AI speaking is upon us

For the past few years, we've been talking about how AI is changing the way we write. I've felt it in my own work. I studied writing long before AI existed and today it feels like I've had to learn the craft all over again. My sentences are shorter. I cut clever turns of phrase to literally avoid sounding like AI. I second-guess punctuation I used to love. Not because my taste changed, but because writing for an internet increasingly filtered through AI rewards a different style.

I've accepted that. What worries me now isn't just AI changing how we write, but how we speak.

If leaks like this become reality — and I suspect they will — we're heading toward a future where interacting with AI won't require a keyboard at all. All we'll have to do is talk. Whether we're chatting through a pair of the best smart glasses or with ChatGPT in our cars, the screen will soon become optional.

And while that sounds incredibly convenient. It also raises a question I can't stop thinking about: if we stop typing, do we eventually stop reading?

What happens when info is heard instead of read

When information is spoken instead of written, when AI summarizes every article, explains every concept and answers every question conversationally, do we lose the habit of wrestling with difficult text ourselves?

I don't think books disappear. But I do wonder if reading becomes more like handwriting or cursive, which is a valuable skill that fewer people practice because technology offers an easier alternative. And while calculators didn't eliminate math, and GPS didn't make maps disappear, they did change what people practiced every day, and the skills that atrophied were the ones technology made optional.

Reading is different because it isn't just about absorbing information. It's how we build vocabulary, strengthen attention spans, develop empathy and learn to think through complex ideas over long stretches of time. Those are cognitive muscles that grow through practice, and they weaken without it.

Ironically, AI may also be one of the best tools we've ever had for preserving those skills. Imagine an assistant that helps you work through a difficult chapter, quizzes you on what you've read or recommends books based on your favorite authors.

What makes this leak different is that it isn't happening in isolation

What makes this leak different is that it isn't happening in isolation. Over the past year, nearly every major AI company has been racing toward a screenless AI.

OpenAI has made ChatGPT Voice feel increasingly conversational while Google is weaving Gemini into Android, earbuds and smart glasses. Meta is betting heavily on AI-powered Ray-Ban glasses that can see what you see and answer questions in real time. Apple is slowly repositioning Siri as a more capable AI assistant that lives across your devices.

What I'm seeing is that we are moving away from an on-screen chatbot to a world where AI becomes almost ambeint. If AI can listen while it's speaking, understand interruptions and respond naturally without awkward pauses, it removes one of the biggest barriers between humans and machines. The conversation starts to feel... normal.

Final thoughts

Once talking becomes easier than typing, most people will choose talking. And that's the future I keep coming back to when I think about AI. While the technology is remarkable, it's different and ultimately changing how we interact with AI — even more human-like.

And while every technological shift changes not only what we can do, but what we choose to practice, if voice becomes our primary interface with AI, then reading and writing may quietly become secondary skills for many people.

That possibility fascinates me as much as it worries me.

More from Tom's Guide

• I almost ruined a relationship with one text — then I tried this ChatGPT prompt
• I built five ChatGPT agents — then replaced them with one 'master' AI
• Even Google doesn't have enough AI power — and that should worry all of us

One of the easiest things to do during an argument is react. And in the heat of the moment, one of the hardest things to do is to pause.

Whether it's a disagreement with your partner, a tense conversation with a coworker or a frustrating text message from a friend, emotions have a way of convincing us that the first thing we want to say is also the best thing to say.

I've learned that's rarely true.

Instead of asking ChatGPT to tell me who's right, I recently started using what I call the Pause Prompt. No, the goal here is not to win the argument, but rather to slow down just enough to response with intention instead of emotion.

After trying it several times, I can see myself coming back to it whenever emotions start taking over.

The 'Pause' prompt

One thing I like about this prompt is that you don't have to have ChatGPT Memory enabled to make it work. The AI doesn't have to know you or how you typically handle pressure. While it helps to give AI the full picture, it's not necessary.

Here's the prompt I give ChatGPT: "I'm upset and I want to respond, but I don't want to make things worse. Here's what happened: [Describe the situation.]

Before giving advice:

• Tell me what emotions you think I'm experiencing.
• Explain what the other person might be feeling, even if they expressed it poorly.
• Point out any assumptions I'm making or information I may be missing.
• Help me write a response that is honest, respectful and likely to move the conversation forward instead of escalating it.

Don't automatically agree with me. If I'm being unfair or reacting emotionally, tell me kindly but directly."

Why it works

Most AI prompts ask ChatGPT to solve a problem. But this prompt is for you. It works because it helps you slow down.

Instead of immediately drafting a reply, ChatGPT first reflects back what it sees in the situation. Sometimes that means recognizing frustration, disappointment or embarrassment that you hadn't fully acknowledged yourself.

It also forces you to consider the possibility that the other person isn't acting out of malice. They may simply have different information, be under stress or be interpreting the situation differently.

That shift in perspective doesn't mean you have to agree with them. It simply makes it easier to respond thoughtfully instead of defensively. And no, it's not about letting AI fight your battles.

And AI certainly doesn't replace the human conversation. I don't even respond verbatim with what ChatGPT has said. This prompt simply gives me a platform to gain a deeper perspective if I'm not thinking clearly due to overwhelming emotions.

My favorite part of this prompt is in there for a reason. The last instruction helps me fully see the scene for what it is (not just from one angle).

"Don't automatically agree with me."

AI has a habit of validating the person asking the question. That's useful in some situations, but not when you're trying to have a productive disagreement. Giving ChatGPT permission to challenge your assumptions makes the conversation far more balanced—and, in my experience, far more useful.

Bottom line

I am not suggesting that AI should replace difficult conversations. In fact, this prompt has actually helped me better plan for those tough chats with my spouse, child or friend.

The prompt allows me to pause long enough to choose better words and use AI as a sounding board when things get tough. The next time I feel tempted to fire off an emotional response, I know exactly which prompt I'll be opening first.

Give it a try yourself and let me know how it works for you. And, if you have a prompt of your own you use for difficult conversations, feel free to share it in the comments.

More from Tom's Guide

• I turned my YouTube watchlist into an AI bootcamp with NotebookLM — here's how
• I built five ChatGPT agents — then replaced them with one 'master' AI
• I used ChatGPT to audit my subscriptions — these prompts helped me find nearly $2,000 a year in recurring charges

AI has significantly enhanced my productivity, helped me stay more organized and even shifted my mindset in a positive way, but I recently noticed something I didn't expect.

After stacking tools and programming AI agents, I noticed my attention span was hungry for faster, more immediate answers. Even when chatting with humans, I wanted them to hurry up and get to the point. I knew this wasn't just my ADHD talking, it was something bigger. AI has affected my dreams, I wondered if it was doing something to my brain, too. Naturally, I looked into it.

Turns out, psychologists have a nickname for that feeling: "popcorn brain."

What is popcorn brain?

The term "popcorn brain" was coined by David Levy in 2011 a computer scientist and researcher at the University of Washington Information School. Levy used it to describe a mind that becomes "so hooked on electronic multitasking that the slower-paced life offline holds no interest."

It's quickly become shorthand for a growing phenomenon many people recognize when the brain becomes so accustomed to constant stimulation that slower activities like reading a book, focusing on a project or simply sitting quietly, start to feel uncomfortable.

For years, social media has received most of the blame. Now AI may be introducing a new version of the same problem.

Why AI feels different

Unlike social media, AI doesn't just feed you content, it actually creates what you ask it for — in seconds. From summaries to ideas, every question gets an answer immediately with minimal waiting or effort.

That convenience is one of AI's greatest strengths, but some researchers think removing too much friction may come with trade-offs. Because we already know that constant interruptions and digital multitasking make sustained attention more difficult.

Researchers also study something called cognitive offloading a word for our tendency to rely on technology to remember information, organize tasks or solve problems.

That isn't always a bad thing. For instance, GPS offloads navigation, calculators offload arithmetic and AI assistants like Alexa+ help us offload memory. Generative AI simply extends that idea much further by offloading writing, planning, brainstorming and research.

But what we don't realy know yet is whether relying on AI for increasingly complex thinking changes the way we learn and concentrate over time. So far, there isn't strong evidence that ChatGPT or other AI tools truly cause "popcorn brain." But researchers are beginning to investigate how constant AI assistance may influence attention, memory, and problem-solving habits.

Five ways I use AI without sacrificing my attention

The same technology that can tempt us with instant answers can also help us focus. I've used ChatGPT to break overwhelming projects into manageable steps, organize scattered research, reduce decision fatigue, create outlines and to simply plan an otherwise overwhleming day. In those moments, AI actually reduces the mental clutter that makes concentrating difficult.

If you're worried AI is shortening your attention span, try these habits.

• Ask AI to teach instead of simply answering. Instead of requesting the finished solution, ask it to walk you through the reasoning.
• Read the original source. Use AI as a starting point, not the final destination.
• Schedule AI-free work sessions. Spend part of your day writing, reading, or thinking without asking AI for help.
• Use AI after you've tried first. Attempt the problem yourself before asking for assistance.
• Let AI remove busywork — not thinking. Automating repetitive tasks is very different from outsourcing your curiosity.

Final thoughts

AI is making knowledge more accessible than ever before, which is an incredible achievement. But the easier it becomes to get answers, the more valuable our ability to wrestle with difficult questions becomes. We shouldn't sacrifice our own critical thinking skills.

For me, that means using AI as a thinking partner. Let me know how you use AI while maintaining your own crticial thinking skills. Share your thoughts in the comments.

More from Tom's Guide

• I turned my YouTube watchlist into an AI bootcamp with NotebookLM — here's how
• I built five ChatGPT agents — then replaced them with one 'master' AI
• I used ChatGPT to audit my subscriptions — these prompts helped me find nearly $2,000 a year in recurring charges

I've tested plenty of prompts that turn ChatGPT into more than just a search engine, but some of the most useful have helped me build better daily routines.

From creating focused morning workflows to helping me switch off at the end of the day, I've used ChatGPT to design productivity routines that genuinely make me work smarter, not harder. Along the way, I also experimented with prompts specifically designed to generate those routines.

After testing 10 of them, five became part of my regular workflow. These are the prompts I keep coming back to — and why they're worth trying yourself.

Productivity prompts that work for me and you

Productivity routines work best when they’re custom-tailored for the following professionals:

• Entrepreneurs
• Freelancers
• Work-From-Home/Hybrid Workers
• Company founders
• Business executives
• Creatives (authors, artists, designers, etc.)
• Athletes/Bodybuilders
• Healthcare Professionals

In my case, I’ve applied ChatGPT’s productivity routine prompts to my daily schedule as a journalist. The first one in particular, which the chatbot labeled as the “Productivity Architect” prompt, does an amazing job of helping me wade through all the noise and renew my focus on what needs to be done:

Act as a productivity coach and systems designer. Interview me one question at a time to understand: My goals, my work schedule, my energy levels throughout the day, my biggest distractions, my responsibilities, my preferred working style, and my current productivity struggles

Once you've gathered enough information, create a personalized daily productivity system that includes: a morning routine, work blocks, a break structure, deep work periods, admin tasks, learning time, and an evening shutdown routine.

Then explain why each part of the routine fits my personality and circumstances.

That prompt supplied me with a productive routine that gave me set objectives to focus on every day, such as taking the time to write out all my objectives first thing in the morning, setting aside hour-long blocks to ideate & produce articles, going fully analog during my breaks and summarizing my day in the form of a document once the sun goes down.

These four other prompts are ready-made for anyone who wants to tell ChatGPT all about themselves and leave with a custom productivity routine that’ll fine-tune your daily process:

• The Energy-Based Productivity Prompt: I want to build my schedule around my energy instead of the clock. Help me identify: when I do my best thinking, when I should do creative work, when I should handle meetings, when I should do routine tasks, and when I should rest. Ask me questions about my daily energy patterns and then build a productivity system optimized around them.
• The Anti-Procrastination Routine Builder: Act as a behavioral psychologist and productivity strategist. Ask me questions to identify why I procrastinate. Determine whether my procrastination is caused by: perfectionism, a fear of failure, feeling overwhelmed, a lack of clarity, low energy, or poor planning. Then design a productivity routine specifically engineered to reduce those triggers.
• The Blind Spot Finder Prompt: Analyze my current routine and identify: productivity bottlenecks, hidden time drains, inefficient habits, activities with low return on investment, and tasks I should automate, delegate, or eliminate. Ask follow-up questions until you have enough information, then recommend improvements ranked by impact.
• The Future Self Prompt: Imagine you are my future self five years from now. You successfully achieved my goals while maintaining good health, relationships, and work-life balance. Interview me about my current life. Then explain: what habits I adopted, what habits I eliminated, what my daily routine looks like, what I was doing wrong today, and what changes created the biggest results.

Final thoughts

My constant meetings with ChatGPT have led me to come across the best productivity routines that compel me to work smarter (not harder) to tackle my personal and professional duties. For those of you reading this who are looking to turn the chatbot into your own personal productivity planner and behavior analyst, take one or all of these prompts with you into your next ChatGPT consultation.

More from Tom’s Guide

• I used ChatGPT to audit my subscriptions — these prompts helped me find nearly $2,000 a year in recurring charges
• A Reddit user gave an AI agent 6 months and $50,000 to find him a wife — and it reveals where AI is headed next
• OpenAI just made its biggest move against Nvidia — and it could make ChatGPT cheaper to run

When I’m contemplating breaking the bank for something a bit more grand on the cost scale, I do a few things before hitting the “Buy Now” button. I look up product reviews, analyze similar options to the product I’m considering and watch a video or two to see it in action. Then I make sure to meet up with ChatGPT to have a deep discussion about whether or not my next big purchase is worth it or not. That cycle has not only saved me time and money, but also kept me from feeling a huge sense of regret and embarrassment over making the wrong financial decision.

These are the five prompts that fit any circumstance whenever I go to ChatGPT to make a final decision on buying something expensive. Hopefully, these commands do the same for you.

The right prompts to use to get more opinions on my next big purchase

Once I make my way to ChatGPT for my latest “am I making the right decision to ball out on this latest purchase?” conversation, I use one of these prompts to help me make a definitive choice:

• I'm considering buying [product name]. Help me identify the underlying problem I'm trying to solve and whether there are cheaper or simpler solutions.
• Act as a skeptical consumer advocate. Give me the strongest reasons not to buy this product.
• Describe the ideal buyer for this product and tell me whether I fit that profile.
• Create a decision framework for this purchase. What factors should I compare and how should I weigh them?
• Imagine it's one year from now. What are the most likely reasons I would regret buying this product and the most likely reasons I'd be glad I bought it?

That second prompt especially came in handy when I finally decided to pick up a PS5 arcade stick. At first, I had my sights set on the HORI Fighting Stick Alpha, which goes for $199.99. Once I mentioned it to ChatGPT alongside that trusty prompt, it brought a bunch of reasons why I shouldn’t take the plunge.

Those skeptical points included:

• You might end up spending even more on customization: The Fighting Stick α is designed to be opened and modified. That's a feature—but it can also become a money pit. A common progression is buying an arcade stick, replacing its buttons & lever, adding custom artwork, making additional upgrades, etc. A $220 purchase can quietly become a $300–$400 hobby.
• It's expensive for a niche accessory: At around $220 MSRP, the Fighting Stick α costs as much as several new games, a premium controller, an entry-level leverless controller or a significant upgrade to your gaming setup. The question isn't "Is it good?" It's "Will I get $220 worth of enjoyment from it?"
• The Hayabusa parts are divisive: HORI markets the stick around its proprietary Hayabusa joystick and buttons. The problem is that some players love them, while others immediately replace them with Sanwa parts. If you end up preferring Sanwa components, you're effectively paying premium money for parts you'll swap out anyway.
• Better alternatives may exist at the same price: At roughly the same budget, you can often find competing sticks that include Sanwa parts out of the box, different layouts, detachable cables and different ergonomics. For example, the NACON Daija Arcade Fight Stick is frequently cited as a strong alternative and has occasionally been available below the Fighting Stick α's MSRP.

After considering all of ChatGPT’s suggestions, I widened my search for a PS5 arcade stick and eventually laid eyes upon the Qanba B1 Titan Wired Joystick.

Not only was it cheaper (I found one on Amazon for $151.55), but it was also another officially licensed Sony product that instilled that same sense of trust I would have had if I purchased my first option. I got that same sense of brand satisfaction, spent a lot less and ended up preferring Sanwa parts after testing their overall feel on both my fellow fighting game enjoyers’ Hayabusa & Sanwa-equipped arcade sticks.

The takeaway

Doing all the necessary research you need to do before opening your wallet is and should always be a regular occurrence.

Injecting ChatGPT into my financial decisions has turned out way better than expected and I’ll be sure to keep doing that every time dollar signs cover my eyes at the sight of something that’s a couple of hundred dollars or more. Take those prompts if you’d like and let ChatGPT ease the difficulty that comes with your next major purchase.

More from Tom’s Guide

• I used Etsy in ChatGPT to upgrade my bedroom decor — and this prompt worked like magic
• Apple is ending support for 16 devices — here’s the ChatGPT prompt that tells you what to do next
• What is Google Zero — and why your favorite websites are panicking about AI

Like a lot of people who use AI every day, my ChatGPT sidebar had become crowded. I have a custom GPT for just about everything from fact-checking to dedicated brainstorming sessions and messy notes. And while each one is useful, together they are a problem.

Every single new task starts with another new decision as I wonder if I should open research GPT first or jump straight into writing and come back for editing? And then there's always the question of "Wait, didn't I already design this assistant a few months ago?"

In other words, my GPTs were chaotic and I knew I wasn't being as productive as I could be. So instead of creating another specialist, I tried something different. I built one AI whose only job is deciding which AI should handle the task.

Surprisingly, it has become the most useful assistant in my workflow.

Why I stopped thinking about AI as one chatbot

Most people still use AI as though it's a single assistant that should be good at everything, but users are finally catching on that there is a better way. AI models can handle many types of tasks and jobs reasonably well, but I've found they perform better when each assistant has one clear responsibility.

With any productive team, everyone has a specialty. That's why I decided my AI should work the same way. So I built five specialist agents.

Here's the team I originally built:

• Research Agent. Finds reliable sources, identifies missing context and suggests follow-up questions before I begin writing.
• Brainstorm Agent. Similar to a research agent, this one takes my messy notes and turns them into bullet points with viable ideas.
• Fact-check Agent. Flags unsupported claims, reminds me where citations are needed and looks for weak sourcing.
• Editor Agent. Improves readability, removes repetition and smooths awkward transitions before publication.
• Review Agent. Takes a look at traffic, follow up possibilities and engagement.

I've built agents that work for me as a journalist, but you could create your own depending on your needs or career. If you aren't sure what types of agents to create, just thinking about what your ongoing needs are and how they aren't being met. You could even ask ChatGPT to suggest agents for you.

Routing agents have changed everything for me. Instead of asking myself which GPT to open, I now start every project with one assistant.

Its only responsibility is deciding what happens next.

For example, if I tell it: "I'm considering writing about [i.e. Apple Intelligence] but don't have quotes or much information yet."

It will automatically respond with the best agent for the job. It also knows when not to suggest an agent. For example, it's not going to suggest a Fact-check agent if I'm rewriting an email in a softer tone to my neighbor who borrowed my lawn mower.

If I'm brainstorming ideas, it may suggest Research agent after our brainstorming session. Instead of forcing me to remember my workflow, it creates one automatically. That small change has made me far more productive than I expected.

How to build your own routing agent

Start by creating specialist assistants. Do you this by identifying the three to five tasks you repeat most often. For example: analyzing, summarizing, brainstorming, writing, image generation.

Then, create one Custom GPT (or Project) for each job. The more focused each assistant is, the better they'll perform.

Next, build one routing agent. You'll do this by creating an additional Custom GPT.

The prompt I use is: "You are my AI Routing Agent. Whenever I describe a task, identify my goal, decide which specialist agent should complete it, recommend the best sequence if multiple agents are needed and explain your reasoning briefly. If important information is missing, ask clarifying questions before continuing."

From there, give it a decision framework. Mine evaluates every request using five questions: What is the user's goal? Which specialist is best suited to this task? Should more than one specialist be involved? In what order should they work? Is any information missing before work begins?

That keeps every project consistent. You're going to want to start every project with the routing agent. So, instead of opening whichever GPT seems right, begin with your routing assistant. Then, just naturally describe your task and watch it reccomend the workflow.

You'll spend less time managing your AI tools and more time actually getting work done. Plus, as you discover repetitive tasks, add new specialists. Over time, you might create assistants for SEO optimization, social media posts, coding, spreadsheet analysis, email drafting, etc.

The routing agent simply learns about the new specialist and incorporates it into future workflows.

Why this works

This idea is actually the foundation of Sakana — but that AI assistant goes even further by suggesting different chatbot models. What I like about my system is that it still keeps the agents within the same platform. It works just as well with ChatGPT agents as it does with Claude.

If I create a new specialist later, such as an image generation or social media agent, I won't have to remember when to use it. I simply teach my routing agent that the new specialist exists, and it can recommend it whenever it's appropriate. Over time, the system becomes smarter without becoming more complicated.

Give it a try yourself and let me know what you think in the comments.

More from Tom's Guide

• I turned my YouTube watchlist into an AI bootcamp with NotebookLM — here's how
• I built five ChatGPT agents — then replaced them with one 'master' AI
• I used ChatGPT to audit my subscriptions — these prompts helped me find nearly $2,000 a year in recurring charges

If you’re an avid ChatGPT user and haven’t upgraded your usage habits, then you’re most likely still using it the same way you did when you first discovered it back in 2023. Instead of continuing to treat it like a simple search engine, you should treat it more like a reliable thinking partner. Here are seven signs that your ChatGPT usage is behind the times and the biggest fixes you need.

Evolve your old ways of using ChatGPT

Some obvious and not-so-obvious hints expose just how much your time spent with ChatGPT hasn’t moved past the sort of uses beginners most likely abide by:

• You only ask questions that deliver one answer and nothing more: You ask simple questions such as “What’s the best way to learn [topic]?”, get your answer and move on from the subject at hand. It’s better to build a conversation derived from more thoughtful prompts instead of isolated ones. A good example of this is telling ChatGPT, "Assess my current skill level, identify the biggest gaps, create a 30-day curriculum, quiz me weekly, and adapt the plan based on my progress."
• You request information instead of asking for methods meant for personal transformation: Simply asking ChatGPT to give you the information you need about a certain activity meant to improve your life isn’t the way. Instead of telling that chatbot to simply tell you about productivity, it’s better to prompt it to observe your habits and create a productivity routine that caters to you. A strong prompt that can make that happen is "Analyze my schedule, identify where I lose time, redesign my workday, and create a weekly review system."
• You never assign ChatGPT a defined role: It’s easy to transform the chatbot into anything you desire in connection to your requests. “Find my hidden strengths” is far too simple a prompt. This example works much better to dive deeper into your request: "Act as a strengths coach and behavioral psychologist. Ask me 15 questions, one at a time, about my career, hobbies, relationships, challenges, accomplishments, and moments when I felt most energized. After I've answered all of them, identify 10 strengths that I may be overlooking. For each strength, provide evidence from my answers, explain how it shows up in my life, and suggest practical ways I can use it more intentionally in my career and personal life."

Continue upgrading your prompts

Here’s a slew of other bad habits that are in dire need of fixing if you want to become a better ChatGPT user:

• You don’t make it challenge you: ChatGPT and other chatbots of its ilk are prone to validating their users when presented with simple prompts like “Here’s my idea. What do you think?” Turning ChatGPT into something that challenges your viewpoints makes for a more helpful experience. This prompt is a great example of that method: "Identify any weaknesses, blind spots, flawed assumptions, competing viewpoints, and reasons this idea might fail."
• You don’t use it like a simulator: AI tools provide some fascinating insights whenever it's presented with a prompt designed to simulate life coaches, hiring managers and even a future version of yourself. Playing around with the following prompt will show just how ChatGPT can act out simulated situations: "Simulate a roundtable discussion between Simon Sinek, Gary Vaynerchuk, and Lewis Howes debating my career strategy."
• You don’t create reusable prompts: Using a prompt just one time and moving on from using it ever again just isn’t the way to go when interacting with ChatGPT and other AI tools. Build a library of your go-to prompts that can be plugged into all sorts of situations, such as coming up with a reusable prompt that’s meant to find the most trending news of the day, help you learn a new skill or turn ChatGPT into your personal tutor on a particular topic.
• You never ask it to analyze you: It’s common to use ChatGPT to simply ask questions about the world at large. But the most curious users tap into the chatbot to ask questions about themselves after making it aware of their habits. Prompts such as "What type of career trajectory would someone with my interests likely enjoy?" and "What recurring blind spots appear in my thinking?" work great in this regard.

Final thoughts

Leave your 2023 ChatGPT usage behaviors behind and bring yourself into the current year by turning into an AI power user who can teach others about your amazing ways. Making great use of OpenAI’s chatbot to help you access knowledge faster, make more informed decisions and build repeatable systems through reusable prompts are just a few of the new routines you can adopt to become a modern-day ChatGPT expert.

More from Tom’s Guide

• NVIDIA claims its new AI data centers use almost no water — here's what that actually means
• Is your job 'AI-resilient'? Find your risk score with our career calculator
• Anthropic's Fable 5 ban exposed AI's next big problem — but Sakana's Fugu may have the answer

OpenAI has unveiled Jalapeño, its first custom-built AI processor, developed with Broadcom specifically to run the large language models that power ChatGPT. Unlike the graphics processors (GPUs) that dominate today's AI infrastructure, Jalapeño was designed from the ground up for one job: answering user prompts as quickly and efficiently as possible.

This move is huge for the future of ChatGPT as the AI assistant could become faster, cheaper to operate and more reliable over the next few years.

Why OpenAI built its own chip

Nvidia remains the undisputed leader in AI hardware, supplying the GPUs that train and run many of the world's most advanced AI systems. But those chips were designed to tackle a wide variety of computing workloads, not exclusively the task of serving billions of chatbot responses every day.

Now Jalapeño takes a different approach. Instead of being a general-purpose processor, it's an inference chip, meaning its hardware is optimized specifically for generating answers after an AI model has already been trained. That's the part of the AI process you interact with every time you ask ChatGPT a question.

According to OpenAI and Broadcom as reported by Reuters, the new chip was architected around the memory, networking and computing patterns used by modern large language models, allowing it to perform those workloads more efficiently than conventional hardware. Early internal testing suggests it delivers better performance per watt than today's leading AI accelerators, although independent benchmarks have not yet been released.

Why this matters for ChatGPT users

If OpenAI can answer more questions while using less electricity and fewer expensive chips, the economics of running ChatGPT begin to change. That doesn't necessarily mean your ChatGPT Plus subscription will suddenly become cheaper, but lower infrastructure costs could allow OpenAI to respond to prompts faster, support more users during peak demand, reduce outages caused by hardware shortages and roll out more capable AI models without dramatically increasing operating costs.

In other words, beyond building a faster chip, OpenAI is making ChatGPT itself more scalable.

OpenAI is becoming more vertically integrated

Until now, OpenAI has largely relied on partners to provide the hardware powering its AI systems. By designing its own processors while continuing to build its own models and products, the company is gradually taking control of more of the entire AI stack.

That's a strategy we've seen succeed elsewhere in technology. Companies that design both their hardware and software can optimize them to work together, improving performance while reducing long-term costs.

OpenAI has described Jalapeño as the first step in a multi-generation compute platform, suggesting this won't be its last custom processor. To me, this looks more like an effort to reduce dependence on a single supplier while lowering one of OpenAI's biggest ongoing expenses.

Nvidia still dominates AI training, and OpenAI continues to rely on Nvidia hardware across much of its infrastructure. Jalapeño is designed for inference, not to replace every GPU in OpenAI's data centers.

The takeaway

Interestingly enough, AI actually helped design the chip. The company says Jalapeño went from concept to production-ready design in just nine months, which is an unusually fast timeline for advanced semiconductor development.

It's another reminder that AI isn't just writing code anymore, but increasingly, it's helping engineers build the hardware that future AI systems will run on.

Although most users will never see Jalapeño, (they'll never need to), but they will . notice the results every time ChatGPT responds a little faster, serves a few more people or unlocks capabilities that were previously too expensive to run.

More from Tom's Guide

• Anthropic will pay workers $85,000 to learn AI — and it reveals the next big AI job trend
• Anthropic's Fable 5 ban exposed AI's next big problem — but Sakana's Fugu may have the answer
• Is your job 'AI-resilient'? Find your risk score with our career calculator

I'm the type of person who takes advantage of free trials and then forgets to cancel them. Then, with every new charge, I try to figure out how to cancel them and then wonder if I should keep them. It's a vicious cycle that I'm not proud of — most of all, it's a waste of money.

If you're like me, you know that between streaming services, cloud storage plans, AI subscriptions, fitness apps and a slew of digital memberships, it's surprisingly easy to lose track of what you're paying for each month.

I recently decided to audit my own subscriptions and instead of manually sorting through everything, I used ChatGPT as a personal financial assistant. It was much faster than I anticipated and I finally got a better picture of every subscription. When I added everything together, I discovered I was spending nearly $166 per month on subscriptions, or almost $2,000 per year.

Here's exactly how I used ChatGPT to do it.

Step 1: Gather all of your subscriptions

If you're like me, you have subscriptions connected to Amazon such as Amazon Music and Audible, as well as Amazon subscriptions like automatic reorder. At the same time, I also have subscriptions tied to Apple such as LinkedIn, Apple+, other streaming services and app subscriptions. Plus, we can't forget about other memberships and subscriptions through Google or that are completely seperate.

Pulling together all your subscriptions from multiple places is a challenge in itself. I have an iPhone so I started by going to: iPhone: Settings > Apple ID > Subscriptions. From there I checked Google Play subscriptions, Amazon memberships and digital services, credit card statements, PayPal recurring payments and even bank statements.

I had almost no extra time to gather all these screenshots — which may explain why I'd let these subscriptions continue for so long. By collecting the screenshots over several days and saving them in ChatGPT Projects until I was ready to act, the process felt much more manageable.

For every subscription, take screenshots. Mine included: Apple One, Amazon Prime, Audible, Kindle Unlimited, Peacock (just upgraded for the World Cup), Google One, ChatGPT Plus, Claude Pro and Meta.

Step 2: Paste everything into ChatGPT

Once you have a complete list, use a prompt like this: "I want to reduce my monthly subscription spending. Here is a list of all my subscriptions and their monthly costs: [paste list here]. Calculate my monthly and annual spending by identifying any overlapping services, flag subscriptions I may not be fully using, categorize each as Keep, Investigate or Cancel, estimate how much I could save annually and explain your reasoning for each recommendation"

ChatGPT will quickly organize everything into a readable audit. And, if you have Memory enabled, Dreaming will actually help put the pieces together to determine if those are subscriptions you truly need.

Step 3: Look for overlaps

This is where the biggest savings often hide. For example, paying for both Apple One and Google One storage or subscribing to Audible plus Kindle Unlimited. Multiple streaming services that aren't being fully utilized or having several AI subscriptions serving similar purposes will all be flagged.

In my case, ChatGPT immediately flagged a few subscriptions worth investigating because they provided similar benefits.

That doesn't automatically mean cancel them, but it does reveal where you're possibly paying twice for convenience.

Step 4: Ask ChatGPT tough questions

The real value comes from follow-up questions. Try prompts like:

"Which of these subscriptions would I miss the least?"

"If I canceled three subscriptions today, which would save the most money while creating the smallest impact?"

"Which subscriptions provide the lowest value per dollar?"

"Help me build a plan to reduce my spending by $25 per month."

I also took advantage of ChatGPT Tasks and set a reminder for ChatGPT to cancel trial subscriptions that had already started.

Step 5: Check your free trials

Speaking of free trials, one surprise from my audit was finding services that had started as trials and quietly became recurring expenses.

Try this prompt: "Based on my subscription list, which services are most commonly forgotten or underused?"

"What questions should I ask myself before renewing each one?"

ChatGPT can help you evaluate whether you're paying out of habit or because you genuinely use the service.

What surprised me most

Honestly, the total. Individually, most subscriptions felt small. Ten dollars here. Fifteen dollars there. Twenty dollars for an AI service I use every day. But when everything was combined, the total approached $2,000 per year.

That's enough to fund a vacation, boost retirement contributions or build an emergency fund. Taking the time to have ChatGPT solve my subscription crisis was a game changer. A simple ChatGPT conversation turned into a financial audit for my family by helping me identify recurring charges, spot overlaps and decide what stays and what goes. I should have done it sooner and plan on doing it every few months or so.

Have you ever used ChatGPT to help manage your finances? Let me know in the comments.

More from Tom's Guide

• Claude Skills changed how I work — 5 ways I use them every day
• I turned my YouTube watchlist into an AI bootcamp with NotebookLM — here's how
• I built five ChatGPT agents — then replaced them with one 'master' AI

Forgive me for being extra late to the movement, but I was today years old when I learned about the “Anti-Prime Day” trend.

Simply put, it’s a period where non-Amazon retailers launch competing sales events with attractive discounts for all manner of items. Now that we’re smack dab in the middle of this year’s installment of Prime Day sales, I’m keen on checking out the abundance of sweet deals I can find.

As a fan of wearable devices, video games, action flicks, trading card games and books, I had a feeling that some of the retailers I frequent inside and outside my home could hook me up with the appropriate items.

So I made a beeline to ChatGPT and asked it what methods I should adopt to cash in on the most appealing Anti-Prime Day sales. Not only did it show me the best price history tools to use, but it also sent me to one of my favorite retailers to discover discounts that are better than some Prime Day sales.

Adopting the best Anti-Prime Day discount searching methods

After asking ChatGPT how to come out at the end of Anti-Prime Day sales with the best deals imaginable, it responded with this statement to put me on the right path:

If you're trying to find the best Anti-Prime Day sales, the goal isn't just to avoid shopping on Prime Day—it's to take advantage of the fact that almost every major retailer launches competing discounts during the same period.

Then it went on to list a bunch of competing online retailers that currently have a long list of amazing products being sold at good to great prices:

• Best Buy (Tech Fest Sale; Ends on Sunday, June 28)
• Kohl’s (Deal Days; Ends on Sunday, June 28)
• Cozy Earth (Cozy Day Deals: Ends on Saturday, June 27)
• Bugaboo (Summer Sale; Ends on Tuesday, July 7)
• Athleta (Semi-Annual Sale; Ends on Sunday, June 28)
• Walmart (Walmart Deals; Ends on Sunday, June 28)

The chatbot also made mention of what items tend to get discounted to their lowest possible prices during Anti-Prime Day sales. The ones that stood out to me from that list included physical video games, gaming accessories, headphones, anime box sets, graphic novel omnibuses, manga box sets, LEGO sets and 4K Blu-rays.

ChatGPT also recommended the perfect Anti-Prime Day setup to rely on, which primarily focused on using a bunch of efficient price history tools:

• Keepa for deep Amazon price history.
• CamelCamelCamel for quick verification.
• Slickdeals alerts for deal discovery.
• Honey for cross-retailer coupon and price tracking.
• Compare prices at Target, Walmart, Best Buy, and Barnes & Noble before buying.
• One useful rule: if a Prime Day deal isn't at or near the lowest price shown in Keepa or CamelCamelCamel, it's usually worth waiting for a better sale later in the year. Multiple deal experts and experienced shoppers use that exact approach to avoid fake discounts and inflated "was" prices.

The six Best Buy goodies I came across

Since my main focus during all those Anti-Prime Day sales was getting the best prices for a new smartwatch, a couple of video games, a collection of films starring an iconic spy and a set of trading cards I had been eyeing for a long time, I ended up on Best Buy’s digital storefront and found exactly what I wanted.

And what’s even better about these finds is how their discounts ended up being better than the ones seen on Amazon:

• Garmin - Forerunner 165 GPS Smartwatch: was $249.99 now $199.99 ($249.98 on Amazon)
• BRAVELY DEFAULT FLYING FAIRY HD Remaster - Nintendo Switch 2: was $39.99 now $19.99 ($49.99 on Amazon)
• STREET FIGHTER 6 Year 1-2 Fighters Edition - Nintendo Switch 2: was $59.99 now $29.99 ($49.95 on Amazon)
• Gothic 1 Remake - PlayStation 5: was $59.99 now $23.99 ($59.88 on Amazon)
• The James Bond Collection (24 Films) - Blu-ray: was $68.06 now $61.26 ($98.99 on Amazon)
• Wizards of The Coast - Magic: The Gathering Final Fantasy Starter Kit: was $19.99 now $14.99 ($19.99 on Amazon)

The takeaway

I owe it to my living under a cave at some point and not even knowing about Anti-Prime Day before this little ChatGPT experiment I conducted. Better late than never, I say!

My go-to chatbot helped me set up a winning game plan for this week’s digital sales from a variety of online retailers, use price history tools to compare all the items I had sitting on my Amazon wish list and find much better prices for those same items through Best Buy.

What started this year as a money-saving experiment has now morphed into an annual tradition for me and my wallet.

More from Tom’s Guide

• I asked ChatGPT to help me score big on Prime Day — these 5 money-saving tips are better than I expected
• I just found Alexa's hidden Prime Day trick — use it now so you don't miss the best deals
• That viral Jeff Bezos quote about AI and water isn't real — here's what actually happened

I’m in my late 30s, but I never stop looking for ways to learn something new. Until recently, I didn't even know what ChatGPT was—that is, until a friend gave me a crash course in AI. Fascinated by its potential, I started testing how it could enhance my daily learning.

Since then, I've used ChatGPT as a personal fitness coach, music curator, travel adviser, and custom tutor. It completely transformed how I acquire new skills. Here are the three prompts I used to upgrade my learning experience, and how you can use them to master whatever you're curious about.

Prompting to increase my know-how

Let’s state the obvious: learning a new language is tough. I got a C in high school Spanish, and usually rely on translation apps just to order drinks on vacation. While I’m not bilingual yet, I’ve had a blast learning the basics of Spanish and Japanese using this roadmap prompt:

Prompt 1: The Language Roadmap "I want to learn [language] from scratch. Act as an expert instructor and curriculum designer. Assess what a beginner needs to know, identify the 20% of concepts that produce 80% of the results, and create a step-by-step learning roadmap. Break the roadmap into: Beginner, Intermediate, and Advanced. For each stage, include: Key concepts, Recommended practice exercises, Common mistakes, Milestones that prove mastery and Estimated time investment."

Even if you're smart, some topics are just too complex to wrap your head around. Back in college, a professor introduced me to the "Feynman Technique"—a method where you master a subject by forcing yourself to explain it in simple, jargon-free terms.

Whenever I need ChatGPT to break down complex issues—like how AI data centers impact the environment or why the cost of living keeps rising—I use this prompt to flatten the learning curve:

Prompt 2: The Feynman Technique "Teach me about [subject] using the Feynman Technique. Explain it as if I'm 12 years old. Then gradually increase the complexity. After each section, ask me a question to test my understanding before moving on."

I’m no handyman. When something breaks in my apartment, I usually give up and call maintenance. Lately, I've been trying to change that by using ChatGPT as a skill tutor (while double-checking its advice against legitimate sources, of course).

Whether I'm deep-cleaning my PS5, attempting to build a gaming PC, or trying to code, this interactive tutor prompt keeps me on track:

Prompt 3: The Socratic Tutor "Act as my personal tutor for [skill]. Your job is not to lecture me but to guide me. Ask me questions, identify gaps in my understanding, challenge my assumptions, and adapt lessons based on my responses. Do not move on until I've demonstrated understanding."

Bottom line

Bonus Prompt: The Masterclass Deep-Dive "I want to develop an expert-level understanding of [topic]. Don't just teach me facts. Teach me about what experts think about the topic, common misconceptions, contrarian viewpoints, important debates, first principles, mental models, historical context and future trends. Help me move from knowledge to wisdom by challenging my assumptions and encouraging critical thinking throughout the learning process."

The more you use AI, the better it adapts to your habits. By pairing ChatGPT with reliable outside sources, my learning sessions have become genuinely engaging instead of a chore. With these three prompts in my digital back pocket, picking up new skills and researching random topics has never been easier.

Put these prompts to work for yourself and see how much they change your workflow.

More from Tom’s Guide

• These 5 ChatGPT prompts ended my dinner planning problems — and introduced me to amazing new recipes
• Turns out AI was destroying my phone’s battery — here is the setting I changed to stop it
• I asked ChatGPT and Claude to plan my next dream trip to Japan with Viator — one delivered much better destinations

OpenAI has upgraded ChatGPT Tasks, finally giving the AI chatbot the ability to keep working after you've closed the app.

If you tried Tasks when it first rolled out, you may still think of it as more of a reminder feature. I'll admit, when I first tried it, I wasn't sure how useful it would be for me. You can ask ChatGPT to remind you about appointments, recurring chores or upcoming events. Useful enough, but I always felt as if my phone or Alexa was enough, not to mention, both have been doing that for years.

But the recent upgrade caught my attention because now ChatGPT can monitor the web and connected apps for changes, then notify you only when something worth rnoting happens. For someone who spends way too many hours glued to her computer, this update fundamentally changes the way I can use AI.

Instead of constantly checking websites, searching for updates or refreshing pages, you can hand the monitoring over to ChatGPT and let it do the waiting for you.

Getting started with ChatGPT Tasks

OpenAI says the new Scheduled Tasks feature is rolling out to paid users on Plus, Pro, Business and Enterprise plans. It's not available via the free tier yet. If you have access to the upgraded Tasks feature, you can create a task by describing what you want ChatGPT to do and when you want it to happen. In other words, you do not need to do anything else but simply type a prompt.

For those who would prefer a reminder option, OpenAI has also added a dedicated Scheduled page, available from the ChatGPT sidebar. From there, you can view all active tasks, see when they'll run next, edit task details, pause tasks temporarily or delete them altogether.

The most useful Tasks tend to be specific. Instead of asking ChatGPT to "watch Prime Day deals," tell it exactly what products, price thresholds or categories you care about. The more precise your instructions, the more helpful the notifications are likely to be.

Here are seven ways I'm already using it.

1. Monitoring Prime Day Deals

Prompt: During Prime Day, monitor deals on AirPods Pro, Kindle Paperwhite, Fitbit smartwatches and robot vacuums. Notify me only when a discount exceeds 25% or when a deal appears unusually good compared to typical sale prices.

Who am I kidding, it's Prime Day and I'm not just using one prompt. I'm also using this one: During Prime Day, monitor deals on LEGO sets, educational toys, noise-canceling headphones, tablets and school supplies. Notify me only when discounts reach 30% or more.

With these prompts or prompts like it, you're essentially telling ChatGPT Tasks, "I"m interested in this category. Don't bother me unless something genuinely good appears." That's why ChatGPT Tasks feels like an assistant.

2. Monitoring airfare without checking every day

One of the most annoying parts of planning a trip is repeatedly checking whether prices have dropped. If you're thinking about traveling to Japan or elsewhere this summer, this update is a real time-saver.

Instead, you can tell ChatGPT: Watch airfare from Newark to Orlando and notify me if round-trip prices fall below $250.

Rather than remembering to search every few days, the AI can keep an eye on things and alert you when it's worth paying attention. The alert comes in as an email. ChatGPT uses whatever email you have associated with your account, so be sure you're logged into your ChatGPT account when creating a Task.

3. Following major news or announcements

If you're following a major news story or want World Cup updates without constantly refreshing apps and social media, ChatGPT Tasks can notify you when something important happens. Instead of scrolling through endless headlines hoping not to miss an update, you can ask:

"Notify me when there are major developments in the Israel-Iran conflict."

"Alert me when the U.S. team advances to the next round of the World Cup."

"Let me know when breaking news occurs involving severe weather in my area."

This turns ChatGPT into a personalized news watcher. I use it all the time!

4. Keeping an eye on school and weather disruptions

Parents know how quickly schedules can change. One practical use for Tasks is:

"Alert me if local schools announce closures, delayed openings or emergency schedule changes."

You could also set up weather-related monitoring to stay informed about severe storms, snow days or other disruptions that affect your daily routine.

5. Catching important emails I may have missed

Email overload is real. If I'm expecting an important email, I always set up a Task such as "notify me when X sends me an email." I also use the prompt: "At the end of each workday, summarize important emails that have not received a response."

Instead of scanning through dozens of messages before bed, you get a quick summary of what might actually need your attention.

6. Planning weekends automatically

I love the idea of using AI to eliminate the mental effort of figuring out what to do with the family.

A simple prompt like: "Every Thursday afternoon, suggest family-friendly events happening nearby this weekend."

It can create a personalized list of activities without requiring any research on your part. And while it may seem small, once you use it, you realize it's a total game-changer to already have activities picked out by the time the weekend hits.

7. Creating a personalized morning briefing

One of the most practical uses for ChatGPT Tasks is building a custom daily briefing around the topics that matter most to you.

For example, you could ask: "Every morning, send me a briefing covering the latest technology news, weather, major headlines and anything important happening in my area."

Or: "Every weekday at 7 a.m., summarize the top news stories, stock market updates and sports headlines."

Instead of jumping between news apps, weather forecasts and social media feeds, you can start the day with a single summary tailored to your interests and then click on the stories that interest you.

This was a major update

ChatGPT Tasks was useful before this update, but the ability to monitor information on your behalf is what makes it feel like a real assistant. Instead of waiting for your next prompt, ChatGPT can now keep watch over the things you care about and notify you when something important changes.

It's not replacing your calendar or becoming a fully autonomous AI agent overnight. But it's one of the clearest signs yet that ChatGPT is evolving beyond a chatbot and into something far more useful for everyday life. Give it a try and let me know in the comments how you're using it — I may feature your suggestion in a story.

More from Tom's Guide

• Anthropic will pay workers $85,000 to learn AI — and it reveals the next big AI job trend
• That viral Jeff Bezos quote about AI and water isn't real — here's what actually happened
• Is your job 'AI-resilient'? Find your risk score with our career calculator

The headline-grabbing news that Anthropic is launching a $150 million “Claude Corps” fellowship to pay early-career workers $85,000 to learn AI highlights a massive shift in the labor market. It proves that the biggest hurdle facing tech giants right now is getting real-world organizations to understand how to actually use AI.

But beneath the surface of this massive hiring push is something worth noting: you don’t need a computer science degree to survive the ongoing AI transition. But what you do need is to understand how vulnerable your current career track is to wage stagnation.

The future belongs to "domain experts," otherwise known as people with deep knowledge of specific industries like education, healthcare or finance and who know how to work alongside large language models. If you refuse to adapt, your job might not disappear overnight, but your earning potential could quietly shrink as AI-literate peers optimize their output.

Test your career's AI resilience right now

Your long-term career security depends on a vital mix of your industry experience, your human-centric "soft" skills and your willingness to build basic AI fluency as studies show employees who use AI can make up to 40% more.

To see where your current role falls on the safety spectrum, enter your field and experience into our AI Career Resilience Calculator.

How to boost your score for free

If your calculator results show you have high exposure to automation, you do not need to panic, and you don't need to drop thousands of dollars returning to school for a technical degree. The absolute fastest way to build career resilience is to learn how to manage the AI tools disrupting your sector before someone else does.

For a closer look at the training programs tech companies expect applicants to know, you can check out this Anthropic Paid AI Fellowship Prep Stack Video which details the free prompt engineering resources and certifications you can use right now.

You can also build competitive AI literacy right now with tools like Masterclass and other free resources that I think are exceptional:

• Anthropic Academy: You can access 17 free courses and step-by-step prompt engineering interactive tutorials hosted directly on GitHub to learn how to structure instructions like a pro.
• Official AI Fluency Courses: Platforms like Coursera host official, beginner-friendly AI literacy certifications designed specifically for non-technical workers looking to upskill quickly.
• Build a "Human-in-the-Loop" Project: Look at the most repetitive task in your current daily routine —whether it's sorting spreadsheets, drafting emails or scheduling workflows, then and use a free LLM to automate it.

Bottom line

Anyone can say they've used ChatGPT. Far fewer people can point to a specific workflow they improved, a task they automated or time they saved. That's the kind of experience that stands out in interviews.

As AI becomes a standard part of the workplace, employers are increasingly looking for people who can use these tools thoughtfully rather than simply experiment with them.

Even a small project that saves time, improves accuracy or reduces repetitive work can demonstrate initiative, problem-solving skills and a practical understanding of how AI fits into real-world business processes.

More from Tom's Guide

• Anthropic will pay workers $85,000 to learn AI — and it reveals the next big AI job trend
• That viral Jeff Bezos quote about AI and water isn't real — here's what actually happened
• Turns out AI was destroying my phone’s battery — here is the setting I changed to stop it

Amazon Prime Day prep has become an annual tradition for me.

Every year, I spend the weekend before the big sale scouting early discounts and building a wish list of items I hope to snag at the lowest possible prices. Over the years, Prime Day has helped me amass a sizable collection of video games, anime Blu-rays and Marvel and DC graphic novels — and this year is no different.

But this time, I decided to bring ChatGPT along for the hunt.

With AI tools becoming increasingly useful for shopping and research, I wanted to see whether ChatGPT could uncover deals I might have overlooked or point me toward products I didn't even know were on sale.

The results surprised me. Not only did the chatbot offer practical strategies for finding the best Prime Day bargains, but it also surfaced several deeply discounted items that made their way into my shopping cart almost immediately.

Prepping for Prime Day sales with this essential advice

To get this money-saving mission underway, I presented this prompt to ChatGPT to see how it could assist me:

“I'm into video games, anime on Blu-ray, and graphic novels from Marvel Comics & DC Comics. What are the best ways to prepare for Amazon Prime Day and find the best sales for those types of items during that sale period?”

The chatbot initially responded with a piece of advice that set me up nicely: For the categories you're interested in—video games, anime Blu-rays, and Marvel/DC graphic novels—Prime Day rewards preparation more than impulse buying. The biggest savings often go to people who build watchlists weeks in advance and know the "normal" prices before the sale starts.

After taking note of its suggestion to build a watchlist of all the items I’m looking to buy before and during Amazon Prime Day, I also jotted down ChatGPT’s five tips on how to locate the best prices imaginable:

• Track Historical Prices: One of the biggest Prime Day mistakes is assuming every "deal" is actually a deal. Use sites/tools like camelcamelcamel and Keepa. These tools show Amazon price history.
• Follow Deal Communities: Prime Day moves fast, and some of the best deals disappear within minutes. Stay updated on them by following sources such as Wario64 on X, Cheap A** Gamer, Slickdeals and IGN Deals. For your interests specifically, these sources are often faster than Amazon's own deal pages.
• Set a Budget by Category: A collector's trap is buying too many "good deals." Try something like a $150 budget for video games, a $100 budget for anime Blu-rays, and an $80 budget for graphic novels, for example. Then rank your purchases for each category under must-have, nice-to-have and only-if-heavily-discounted. This prevents regret purchases.
• Focus on the First 12 Hours: The strongest deals are often midnight launch deals, Lightning Deals and limited inventory offers. If there's a specific omnibus, collector's edition or anime box set you've been waiting for, check as soon as Prime Day starts.
• Set Up Prime Day Alerts Through Alexa: Alexa can notify you about discounts on items in your Wish List, Cart or Buy Later collection up to 24 hours in advance or track target prices for specific products. Open the Alexa app, go to More, select Settings, tap on Notifications, proceed to Amazon Shopping and turn on Deal Recommendations.

Putting those tips into action

With all of those money-saving tips in the back of my mind, my search for the best Amazon Prime Day sales I could find became a much more streamlined process. Plus, I came upon a bunch of amazing deals that went live before the Amazon Prime Day madness officially kicked in:

• Borderlands 4 - PS5: was $49.99 now $34.99
• Cyberpunk 2077: Ultimate Edition - Nintendo Switch 2: was $69.99 now $57.99
• Akira - 4K & Blu-ray [4K UHD]: was $29.98 now $11.99
• Bayonetta [Blu-ray & DVD]: was $19.98 now $12.82
• Samurai 7: The Complete Series [Blu-ray]: was $49.98 now $34.99
• Daredevil: Born Again [Marvel Premier Collection]: was $14.99 now $11.27 (with Amazon Prime)
• Wolverine: Old Man Logan [Marvel Premier Collection]: was $14.99 now $12.82 (with Amazon Prime)

The takeaway

Patience isn’t one of my best traits—the act of impulse buying has a hold over me that’s just too hard to ignore.

But for this year’s Amazon Prime Day routine, I didn’t rush to buy everything I saw for sale, kept ChatGPT’s money-saving advice close to mind and followed every step it laid out to discover the best deals imaginable.

That plan worked to perfection and made two video games, three anime Blu-rays and graphic novels a part of my growing collection of physical goods I’ll be cherishing forever.

More from Tom’s Guide

• That viral Jeff Bezos quote about AI and water isn't real — here's what actually happened
• TikTok’s AI slop problem is spiraling out of control — and kids are getting the worst of it
• I used ChatGPT to discover my hidden strengths — and it changed my perspective on work

When I have a few hours to kill, my curiosity takes over and I get the urge to do a deep dive into my own psyche with the aid of ChatGPT.

During my latest foray into a personal chatbot personality examination session, I wanted to find out what my hidden strengths are. Being alive this long and taking note of my biggest wins has obviously exposed me to the main areas of expertise. But after having so many exchanges with ChatGPT that helped me uncover some of the parts of my individuality that I didn’t know about beforehand, I figured it was time to learn about my secret specialties.

With a loaded prompt to help me, I entered a lengthy discussion with ChatGPT that unveiled a bunch of qualities about me that I may be overlooking.

And after that intriguing back-and-forth with the chatbot, I found the best ways to make those hidden strengths more prevalent in the ways I conduct my work.

A fascinating Q&A

Presenting this prompt to ChatGPT kicked off this introspective look into my hidden and seemingly overlooked strengths:

"Act as a strengths coach and behavioral psychologist. Ask me 15 questions about my career, hobbies, relationships, challenges, accomplishments, and moments when I felt most energized. After I've answered all of them, identify 10 strengths that I may be overlooking. For each strength, provide evidence from my answers, explain how it shows up in my life, and suggest practical ways I can use it more intentionally in my career and personal life."

The chatbot immediately presented me with some fitting questions, such as “Think about a time in your career when you felt exceptionally proud of your work. What happened, what role did you play, and why did it matter to you?” and “When you're part of a team, friend group, or family gathering, what role do you naturally end up playing?”

I made sure to answer all 15 questions with as many personal details as possible — I spoke about my love for charismatic leaders, the subjects I cared most about back in school, what people rely on me for the most, etc.

One question stood out to me the most: “What behavior in other people frustrates you the most, and why?” I answered with, “Being uninformed. There are just too many devices and avenues that can help anyone learn about anything and find credible sources. Instead, people remain clueless and even refuse to learn how to use new forms of technology.”

Uncovering my hidden strengths and finding the best ways to showcase them

At first, ChatGPT shone a spotlight on my best qualities by stating:

Based on your answers, a pattern emerges that is stronger than "good journalist" or "creative thinker." You appear to be a curator of information, opportunities, and people. Much of your satisfaction comes from discovering valuable things before others do and helping those things reach a wider audience.

Then it mentioned my hidden strengths, provided evidence of why I exhibit that trait, how it shows up, a potential blind spot tied to that trait, and how to use it more intentionally. For instance, ChatGPT acknowledged my “Talent Spotting” abilities and displayed the following information to support its evidence of that skill being prevalent in my life:

• Evidence: You created an interview series featuring rising gaming creators, felt proud not because of traffic, but because you helped others gain visibility and you stay ahead of trends, releases, and emerging stories.
• How It Shows Up: You naturally notice potential before it becomes obvious to everyone else. Many people focus on what's already popular. You seem drawn to what's about to become important.
• Potential Blind Spot: You may underestimate how rare this skill is because it feels natural.
• Use It More Intentionally: Become known as the journalist who identifies future stars, build recurring columns around emerging talent and create prediction lists and trend reports.

The rest of ChatGPT’s observation of my hidden strengths made me aware of an affinity for information curation, intellectual curiosity, opportunity creation, resilience through adaptation, relationship building through conversation, memory for context & details, community intelligence, the courage to explore and service-based ambition.

The takeaway

I paid close attention to ChatGPT’s eye-opening findings about my hidden strengths and made a mental note of taking them into my professional life.

Making a more concerted effort to discover trends early across my subject areas, spotting creative talent before others do, connecting information from journalistic sources as well as online forums, translating the most complex topics into easy-to-understand knowledge and building communities through curiosity & conversation during my daily work sessions became more prominent.

Focusing on consuming and reporting as a journalist is commonplace. Publicly showcasing my wins more, leaning more into audience insights and building a recurring column that reflects my interests have come into the picture a lot more now that I make more use of my hidden strengths.

More from Tom’s Guide

• Ever wonder why ChatGPT is free? The answer is far more calculated than you might think
• I asked ChatGPT to turn life coach quotes into a daily productivity routine — and my mornings have never been better
• ChatGPT's new memory builds a profile of you on its own — and OpenAI admits you can't see all of it

AI features are showing up everywhere on our phones. They summarize notifications, edit photos, transcribe conversations, generate images and power increasingly sophisticated voice assistants. The promise is a smarter phone that does more for you.

But there's a trade-off that only gets noticed when a charger or outlet is nowhere to be found. The problem with on-device AI is that these features require significant processing power, whether they're running directly on your device or constantly communicating with cloud-based AI models. That translates to more battery drain, more background activity and for some users, noticeably shorter battery life.

If you've found yourself wondering why your phone doesn't seem to last as long as it used to, AI could be part of the answer.

Why AI uses so much power

Unlike classic apps that wait for you to do something before they spring into action, AI is different. The latest AI features are designed to be proactive. They're constantly scanning photos, indexing information, listening for wake words, analyzing notifications or preparing contextual suggestions before you even ask for them.

Everything your phone may be doing in the background right now — voice assistant, summarizing notifications, smart search indexes, AI writing tools and photo apps — might not seem significant, but together they create a steady workload that can quietly chip away at battery life throughout the day.

The newest smartphones are increasingly built around dedicated AI hardware, but even specialized chips still consume power.

The hidden cost of 'always available' AI

One reason AI can be such a battery drain is that many features are designed to feel instantaneous. When you tap an AI assistant or ask a question, you expect an answer immediately. To make that possible, your phone is often doing prep work behind the scenes.

This is especially true for voice assistants and AI search features that continuously update information, monitor activity or prepare personalized responses. The more personalized the experience becomes, the more background processing is often required. Ironically, some of the smartest features on your phone may also be some of the most power-hungry.

iPhone settings worth checking

If you're using an iPhone, there are several areas worth reviewing if battery life has taken a hit.

Apple Intelligence features: Apple Intelligence brings AI-powered writing tools, summaries, image generation and other features to supported devices. If you rarely use these tools, consider disabling specific features rather than leaving everything enabled by default.

Siri Suggestions: Siri continuously learns from your habits to provide recommendations across apps and search results. It's worth looking into these by opening Settings, tapping Siri and reviewing Suggestions options. Disabling suggestions you don't use can reduce unnecessary background activity.

Background app refresh: AI-powered apps often rely heavily on background updates. To check what is enabled, open Settings, tap General, select Background App Refresh. Disabling refresh for apps you rarely use can produce meaningful battery savings.

Live transcription features: Real-time transcription and accessibility tools can consume additional power when used frequently. Review which features are enabled and disable those that aren't part of your daily workflow.

Android settings worth checking

Android devices now include AI features from Google's Gemini assistant. If you don't actively use Gemini, review which permissions and background activities are enabled.

Voice activation settings: These are features that continuously listen for commands like "Hey Google" can contribute to battery usage over time. To review: Open Settings, navigate to Google. Select Search, Assistant & Voice. Review voice activation settings

AI photo processing: Many Android phones automatically analyze and organize photos using AI. Review photo app settings to determine whether features such as automatic categorization, cloud syncing and advanced image processing are necessary for your workflow.

Battery optimization controls: Most Android devices include battery optimization settings that can limit background activity from power-hungry apps. Look for battery usage, app battery optimization and background restrictions. These tools can help identify which apps are consuming the most power.

How to tell if AI is actually the problem

Before blaming AI entirely, it's worth checking your battery usage statistics. Both iPhone and Android provide detailed battery reports showing which apps and services are consuming the most energy.

On iPhone: Settings > Battery

On Android: Settings > Battery > Battery Usage

You may discover that an AI assistant, photo app or productivity tool is using significantly more power than expected. In many cases, the culprit isn't a single feature but the accumulation of several AI-powered services running simultaneously.

Bottom line

AI is quickly becoming integrated into our daily lives and defining the features of our smartphones. And while many of these tools are genuinely useful, they also can cause unwanted battery strain.

The irony is that AI is supposed to make our devices more efficient. Yet for many users, the biggest impact may be a battery percentage that seems to disappear faster than ever.

If your phone's battery life isn't what it used to be, you may not need a new device, but to simply check your settings and decide which AI features are worth the power they're consuming. And if you've got any personal advice on this issue to share, let me know about it in the comments.

More from Tom's Guide

• I ran my resume through ChatGPT — these 5 prompts exposed mistakes I kept missing
• I asked ChatGPT to make cleaning less overwhelming — these 7 prompts actually helped
• ChatGPT's new memory builds a profile of you on its own — and OpenAI admits you can't see all of it

Ever since my mind-blowing first trip to Tokyo back in 2018, I've been dying to go back to Japan — specifically to explore Kyoto as a solo traveler. But when it came time to plan my itinerary, I decided to ditch traditional travel blogs and let artificial intelligence do the heavy lifting. Both OpenAI and Anthropic recently integrated Viator into their ecosystems — ChatGPT as an app and Claude as an automated "Connector."

To see which AI tool could plan the ultimate Kyoto getaway, I pitted them against each other in a head-to-head travel planning battle. Here is which chatbot emerged as the victor.

ChatGPT’s recommendations

To map out my seven-day journey in Kyoto, I fed ChatGPT a hyper-specific prompt tailored to my $3,500 budget and personal interests:

The Prompt: "Act as an elite travel advisor. I'm visiting Kyoto, Japan, for seven days. My interests include video games, food, wrestling, shopping for vinyl, anime, and photography. My budget is $3500. Recommend the most memorable Viator experiences, hidden gems, food tours, cultural activities, and unique local experiences. Rank them by overall value and likelihood of becoming a highlight of my trip. Explain why each recommendation stands out from standard tourist activities."

The Viator app integration pulled directly from its live database, displaying options complete with star ratings, review counts, durations, and pricing. Out of a massive list of cultural walking tours and street food crawls, these four top picks immediately stole the show:

• Private Kyoto Tour: Temples, Hidden Alleys & Local Stories
• Real Nishiki Market Brunch Walking Tour (7 Tastings + Brunch)
• Kyoto Gion Geisha District Walking Tour * Kyoto Night Foodie Tour in Gion (9+ dishes + 6 Sake tastings)

Claude’s recommendations

After presenting the same prompt I fed to ChatGPT to Claude, the Viator Connector brought up a variety of cooking classes, sports lessons, theater shows, movie tours, culinary tours and restaurants. And just like ChatGPT’s Viator app, it also brought up a bunch of half-day tours and private sightseeing tours with accompanying star ratings, the total number of reviews, time durations and prices per person.

The major difference between Viator on ChatGPT and Claude is how the Claude Connector laid out a carefully curated Kyoto playbook for me to stick to, which was broken down as follows:

• No.1 Ramen Experience in Kyoto (~$89): Rating: 4.97 | 2,400+ reviews
• Private Kyoto Photo Shoot & Walk in Gion/Higashiyama (~$164): Rating: 5.0 | 24 reviews
• Private Kyoto Tour: Temples, Hidden Alleys & Local Stories (~$100): Rating: 4.86 | 1,633 reviews
• Kyoto Nishiki Market Food Tour – Hidden Tastings (~$41): Rating: 5.0 | Private tour
• Antique Kimono & Tea Ceremony Experience (~$96): Rating: 4.92 | 121 reviews

Beyond those tour options, Claude’s Viator Connector pointed me towards even more experiences that perfectly match my tastes.

It told me to take a day trip to Osaka for a video game-fueled visit to the Super Potato retro gaming store, hit up the Jeugia & Jet Set Records stores in Kyoto to find some music to bring home, spend a few hours perusing around the Kyoto International Manga Museum and attend a K1/kickboxing event pop-up in Kyoto's municipal gym.

Claude prevailed

Out of both ChatGPT's and Claude’s usage of Viator’s vacation planning and recommendation capabilities, I preferred Claude’s presentation much more.

Both chatbots provided me with a strong selection of fun tours to embark on in Kyoto, but Claude went the extra mile to plan out what I should do and generate extra events to put on my to-do list according to my hobbies & interests.

When the time comes, I’ll be sure to keep Claude’s Kyoto vacation advice in mind when the time comes for me to go on my grand adventure.

• I used ChatGPT and Claude to search for StubHub tickets — and one AI crushed it
• I let ChatGPT analyze my personality through my favorite fictional characters — it revealed more about me than I realized
• I used AI to build a 'live draw' sweepstake tool for the World Cup — and you can use it for free, too

Ordering takeout is easy, but it destroys your bank account. Cooking at home is better, but after weeks of recycling the same repetitive recipes (spaghetti losing its charm after the fourth night in a row), I hit a wall. I needed new, budget-friendly meals that would yield plenty of leftovers without burning me out.

To shake up my kitchen routine, I turned to AI. These five ChatGPT prompts completely refreshed my weekly meal prep, helped me discover new recipes, and finally cured my dependency on food delivery apps.

Uncovering new approaches to cooking nighttime cuisines

With my mind full of images of domestic and international dishes I’ve gotten used to eating at restaurants, I went to ChatGPT with hopes of finding similar dishes and more through its recipe-curating expertise.

With these prompts, I found a wealth of dishes that kept me from eating grilled chicken and white rice for the umpteenth time over the course of several weeks:

• The Personal Chef Prompt: Act as a world-class chef and food critic. Based on my preferences, recommend [number] recipes I would probably love but have never heard of. Explain what makes each dish special, what it tastes like, and why it matches my tastes. My favorite cuisines are: [list]. My favorite ingredients are: [list]. The foods I dislike are: [list]
• The "Hidden Gems" Prompt: Recommend [number] hidden-gem dishes from around the world that deserve more attention. Avoid famous dishes like pizza, tacos, ramen, sushi, burgers, and pasta. Focus on dishes locals love that most Americans have never tried.
• The "Travel Through Food" Prompt: Create a [number]-day culinary world tour. Every day, recommend a recipe from a different country. Prioritize authentic local favorites over tourist foods. Include the cultural significance of each dish.
• The "Chef's Secret Menu" Prompt: Imagine you're a Michelin-star chef creating a secret menu for adventurous home cooks. Recommend [number] dishes that feel impressive, unique, and memorable but are achievable in a home kitchen.
• The "What Am I Missing?" Prompt: Based on the foods I eat regularly, what cuisines, ingredients, and cooking styles am I overlooking? Recommend recipes that would expand my palate while still fitting my existing tastes. [list foods you eat regularly]

Putting on extra pounds thanks to ChatGPT

I took that first prompt and ended up diving into a world of recipes I’d either had when I ordered them from a classy joint or never even heard of before.

For favorite cuisines, I mentioned Jamaican, Chinese, Japanese, Thai, Mexican, and Italian. I listed my favorite ingredients as chicken breasts, steak bites, steamed carrots, steamed broccoli, yellow rice, jasmine rice, penne pasta, tomato sauce, Teriyaki sauce, jerk sauce, green curry, yellow curry, red curry, and tortillas (pretty long list, I know).

And finally, I brought up salmon, oysters, dumplings, peas, onions, and quinoa as my most disliked foods.

ChatGPT responded at first with this statement: Based on your favorite cuisines and ingredients, I'd skip the obvious choices (jerk chicken, chicken teriyaki, tacos, chicken parm, etc.) and point you toward dishes that sit just outside the mainstream. These are the kinds of recipes that make people ask, "How have I never heard of this before?"

The five dishes the chatbot recommended to me (that ended up as a permanent part of my weekly dinner plans) were Three Cup Chicken (Chinese), Panang Curry Chicken (Thai), Bistec Encebollado Without Onions (Caribbean), Pollo Alla Calabrese (Italian), and Jerk Chicken Rasta Pasta (Jamaican).

The takeaway

If you’ve reached a point in your life where you’re tired of eating the same dishes for days on end, it might be time for you to take ChatGPT for a spin and tell it to find you new delicacies worth cooking on your own. Buying all the appropriate ingredients, looking up how-to videos & recipes breakdowns, and making those new foods to the best of your ability feels so gratifying. Let AI start you on the road to home-cooked cuisine.

More from Tom’s Guide

• I gave ChatGPT my Dad’s hobbies and asked for Father’s Day Ideas — these recommendations were better than expected
• I used ChatGPT every day for a year. These 5 habits get me better answers than any new model upgrade
• I asked ChatGPT to make cleaning less overwhelming — these 7 prompts actually helped

Here's some good news: just because Apple stops supporting a device doesn’t mean you need to throw it away.

When I first saw reports that Apple is preparing to end software support for 16 devices across its ecosystem, I immediately wondered if my device would suddenly become unsafe to use. And while the answer is more complicated than a simple yes or no, I did find some answers by asking ChatGPT.

And while a device that no longer receives major software updates can often continue working for years, there are legitimate concerns about security patches, app compatibility and whether it’s time to start planning an upgrade. But if you're just hoping to browse the web, stream videos, send emails and use many of your favorite apps, the device should be fine.

But because I wasn't sure how to navigate the risks without support, I turned to AI. Instead of asking a generic question like “Should I replace my iPad?” I created a prompt designed to evaluate any aging Apple device and explain the tradeoffs in plain English.

Support for these 16 devices will end September 2026

This update marks the largest support cut in Apple Watch history, dropping five models at once to focus on newer chips:

Apple Watch Series 6 (2020)

Apple Watch Series 7 (2021)

Apple Watch Series 8 (2022)

Apple Watch Ultra (1st Generation, 2022)

Apple Watch SE (2nd Generation, 2022)

Apple has raised the baseline hardware requirements for iPads, leaving behind five older models:

iPad Air (3rd Generation, 2019)

iPad mini (5th Generation, 2019)

iPad (8th Generation, 2020)

iPad Pro 11-inch (1st Generation, 2018)

iPad Pro 12.9-inch (3rd Generation, 2018)

This update officially completes Apple's transition away from Intel processors, meaning all remaining Intel-based Macs are losing support:

MacBook Pro (16-inch, 2019)

MacBook Pro (13-inch, 2020 – model with four Thunderbolt 3 ports)

iMac (2020)

Mac Pro (2019)

Two older streaming boxes are being left behind to focus purely on modern 4K hardware:

Apple TV HD (2015)

Apple TV 4K (1st Generation, 2017)

(Note: If you own an iPhone, you're safe this time around! The accompanying iOS 27 update continues to support all the same iPhone models as the previous generation).

The ChatGPT prompt I recommend

If you own any of the older Apple devices losing support, copy and paste this prompt nto ChatGPT: "My device is a [DEVICE NAME]. It is currently running [OPERATING SYSTEM VERSION]. Apple has stopped providing major software updates for this device. Act as a consumer technology expert and assess: Whether the device is still safe for banking, shopping and sensitive accounts. How long apps are likely to continue working. What security risks I should know about. Whether I should keep using it, repurpose it or replace it. The most cost-effective upgrade path if I decide to upgrade. Explain your reasoning in plain English and avoid technical jargon."

The beauty of this prompt is that it forces ChatGPT to think through the practical implications rather than simply telling you that newer hardware exists.

What ChatGPT gets right

When I tested similar prompts on older devices, ChatGPT generally did a good job distinguishing between “unsupported” and “unusable.”

To specify, a device that stops receiving major operating system updates doesn’t immediately become dangerous. In many cases, Apple continues providing security updates for older versions of its software. Even when those eventually stop, the device may remain perfectly suitable for low-risk tasks like media streaming, reading, video calls or serving as a dedicated smart-home controller.

Where ChatGPT shines is helping people think through how they actually use their device.

For example, someone who only uses an older iPad for Netflix and YouTube has very different needs from someone who regularly uses that same device for online banking, password management and work-related tasks.

What ChatGPT can’t tell you

As useful as the prompt is, I wouldn’t treat ChatGPT as the final authority. That goes with pretty much any time you prompt an AI, to be honest. The chatbot doesn’t have access to Apple’s internal support plans. It can make educated assessments based on publicly available information, but it can’t predict exactly how long specific apps will continue supporting older hardware.

That’s why I recommend using ChatGPT to help with decision-making tool rather than a decision-maker. Think of it as a knowledgeable friend you can contact at any time to help you evaluate your options.

If the chatbot identifies legitimate concerns about security, compatibility or performance, that’s your signal to do additional research before making a final choice.

The bigger lesson

One of the most useful applications of AI for me is often navigating gray areas. Apple ending support for a device doesn’t automatically mean you need to spend hundreds of dollars on a replacement. But it also doesn’t mean you should ignore the risks forever. This prompt can help you understand where your device falls on that spectrum.

It's a good conversation worth having before you rush out to buy something new. Is your device on the list? Let me know in the comments if you plan on replacing your device.

More from Tom's Guide

• I ran my resume through ChatGPT — these 5 prompts exposed mistakes I kept missing
• I asked Gemini what was slowing down my Windows PC — here’s what it found
• ChatGPT's new memory builds a profile of you on its own — and OpenAI admits you can't see all of it

After a decade, it was finally time to tear down the old, crumbling posters on my bedroom wall and find some new decor. Instead of hitting my usual spots like Amazon or Hot Topic, I opened the ChatGPT app directory to try something different. I loaded up the Etsy app, and with a single prompt, it uncovered a wealth of unique, custom art that was perfect for my space.

Uncovering the finest pieces of wall art through my geeky fandom

With this prompt, I employed the Etsy app on ChatGPT to look through its vast collection of goods and suggest the finest items it could find based on my interests: “@Etsy My interests include video games, anime, wrestling, hip-hop, and comic books. Find the most unique products, collectibles, office upgrades, and conversation pieces based on those interests.”

Etsy initially spoke up and responded with this statement that recommended the sort of items that could satiate my hunger for creative items:

If I were curating an Etsy-inspired dream office based on your interests, I'd focus on museum-quality display pieces, blueprint-style art, collector showcases, and subtle fandom decor rather than obvious merchandise. Etsy shoppers are increasingly gravitating toward personalized, aesthetic-driven decor and unique wall art that feels like a conversation piece rather than standard fandom merchandise.

It then followed that declaration by showing two to three items for each section that it laid out for me based on the passions I told it about.

After taking a peek at all those items on display, I ended up selecting a 240 Grendizer Blueprint wall scroll from the “Anime & Video Game Intellectual Fandom Decor section, a 54" wrestling belt display case (perfect for that future championship belt I get from AEW!) from the “Wrestling Collector Displays” section, a Hip Hop History poster from the “Golden Era Hip-Hop Gallery Wall” section and an Absolue Batman poster from the “Comic Book Executive Office Upgrades” section.

Adding those to my Etsy cart wasn’t the only way its ChatGPT integration aided me. Each section came with a list of Etsy search terms that could put a bunch of other great items in front of me. These are the ones that helped me the most:

• Anime & Video Game Intellectual Fandom Decor: Gundam mechanical schematic prints
• Wrestling Collector Displays: Wrestling poster restorations
• Golden Era Hip-Hop Gallery Wall: Vinyl record floating displays
• Comic Book Executive Office Upgrades: Comic panel desk mats

Those search terms made my search for lively bedroom decor that much easier, as I was shown all manner of posters, wall prints, vinyl display containers and comic book-inspired mousepads that made their way onto my Etsy “Favorites” wishlist.

The takeaway

I’ve jumped into the full pool of ChatGPT apps that represent everyday brands that I use, such as StubHub, Ticketmaster, Apple Music, Spotify, etc.

Etsy has always been my go-to thanks to its digital Mom and Pop shop-feel and its abundance of vintage & custom items that look great in any home. Taking inspiration from my hobbies and searching for the perfect bedroom decor that showcases my undying love for those interests worked out way better than expected with ChatGPT’s Etsy integration.

I hope you find just as many incredible items through that process as I did.

More from Tom’s Guide

• The 5 fastest ways to get past AI customer service chatbots — here's what actually worked at Amazon, Optimum, Walmart, AT&T and more
• I use AI every morning before 9 a.m. — these 5 habits make my day easier
• I asked ChatGPT to analyze my personality and recommend comics — its picks were uncannily good

OpenAI recently began rolling out a new memory system called Dreaming, designed to help ChatGPT build a more useful understanding of who you are over time. Instead of relying primarily on facts you've explicitly asked it to remember, the new system synthesizes information from your past conversations and updates those memories automatically as your life changes.

The good news for users is that this feature is rolling out in stages. Dreaming reached ChatGPT Plus and Pro subscribers in the U.S. first, with free, Go and international users expected to follow over the coming weeks. So, depending on your plan and region, you may or may not see it yet.

As the upgrade rolls out, ChatGPT may be remembering more about you than you realize. As soon as I learned about the update, I checked one setting immediately.

The setting to review right now

I recommend doing this right now: open ChatGPT and navigate to Settings > Personalization > Memory > Reference chat history

If this setting is enabled, ChatGPT can draw on your previous conversations to personalize future responses.

For example, if you've mentioned your favorite sports team, dietary preferences, travel plans, writing style or career goals across multiple chats, ChatGPT may use that information later without you needing to repeat it. The goal is to make conversations feel more natural and cut down on the context you have to provide.

For many people, that's a feature. For others, it's something worth reviewing.

What 'Dreaming' actually does

According to OpenAI, Dreaming is a more advanced memory architecture that continuously synthesizes information from your conversations and keeps it current as circumstances change. The company's own example: a memory that reads "going to Singapore in July" rewrites itself to "went to Singapore in July 2026" once the trip is over, so the user then no longer needs to input that information.

Seems fine, but here's the part that's easy to miss. Previously, ChatGPT's memory worked like a list of notes, in that it mostly remembered what you explicitly told it to. Dreaming shifts that work into the background, so the system now decides on its own what's worth keeping, including things it inferred about you and then quietly revised over time, not just facts you stated outright.

That's a meaningful difference, and it makes the record harder to audit: OpenAI acknowledges that the memory summary page may not capture everything ChatGPT remembers about you. Memory is becoming somewhat of an evolving profile.

As AI assistants mature, we're moving away from one-off conversations and toward tools that hold context across weeks, months and eventually years. OpenAI says Dreaming was built specifically to make ChatGPT more helpful by forming a more complete understanding of each user.

That can be genuinely convenient. But it also means many users will want to spend a few minutes seeing exactly what information ChatGPT is drawing on to personalize its responses.

On the bright side, OpenAI still gives you control. You can disable memory, manage or delete stored memories individually, and use Temporary Chat sessions that don't contribute to memory at all.

My advice

I'm not turning Dreaming off, and frankly, I don't think most people should. The whole point is that ChatGPT stops making you repeat yourself, and on that count it delivers.

But convenience and visibility are pulling against each other here. The old memory was a list you were the one managing every time you wrote. The new one is a profile ChatGPT writes about you, partly from things it inferred rather than things you said, and OpenAI itself acknowledges the summary page may not show all of it.

It's a good idea every so often to open your memory summary and actually read it. Don't just check whether memory is on, dive deeper and check whether it's right. A wrong fact you stated, you can spot and delete in seconds. A wrong assumption the system quietly made about you is the kind of thing that shapes every answer you get next, and the hardest to see.

More from Tom's Guide

• AI is destroying itself with 'data cannibalism' — but there's a simple fix
• I asked ChatGPT to make cleaning less overwhelming — these 7 prompts actually helped
• SpaceX just spent $60 billion on Cursor — and it proves AI chatbots aren't the future anymore

Getting your résumé noticed in today’s job market is a gauntlet. Between navigating automated Applicant Tracking Systems (ATS) and catching a hiring manager's eye in under six seconds, a generic CV just won't cut it anymore.

While Google autocomplete trends prove that millions of people are turning to AI for career help, generic prompts like "write me a resume" usually lead to generic and ineffective results. Ironically, when you're desperate to land your next role, spending time perfecting your résumé and applying to jobs can feel like a full time job.

After hours of testing and tweaking, I’ve refined a toolkit of specific, highly effective prompts that go beyond the basics. Whether you are aiming to beat the bots, punch up your bullet points or tailor your experience to a dream role, here are 5 ChatGPT prompts I actually use to make a résumé stand out.

TL;RD hot tips for your résumé

• Never copy-paste blindly: ChatGPT is a brilliant first-draft partner, but always review the output to ensure it accurately reflects your actual experience. Don't let it invent skills or experience you didn't actually achieve.
• Keep your formatting clean: Once ChatGPT gives you the text, paste it into a clean, text-based template (like Google Docs or Word). Avoid heavy graphics or multi-column layouts, which confuse ATS scanners.
• Feed it context: The more specific details you give the AI about your industry and background, the less generic your results will be.

1. Ensure a real human sees your résumé

The prompt: "Act as an expert recruiter. Here is the job description [Insert Job Description] and my current resume [Insert Resume]. Analyze both and identify the key skills, hard keywords and core methodologies missing from my resume that are highly emphasized in the job description. Provide a list of these gaps and suggest exactly where and how to integrate them naturally."

As much as we want to get our résumés in front of human, the truth is, 70% of most major companies use Applicant Tracking Systems (ATS) to scan resumes for specific keywords before a human ever sees them. This prompt acts as your preliminary scan, identifying exactly what your résumé is missing compared to the job description.

The goal: Ensure your résumé doesn't get automatically filtered out by recruitment software.

2. Let your experience speak for itself

The prompt: "Review the following bullet points from my résumé. They currently sound too passive or task-oriented. Rewrite them using strong action verbs and a results-driven framework (Focus on: Action + Context + Quantifiable Outcome). If a metric is missing, use brackets like '[insert metric]' to show me exactly where I need to add data or percentages. Here are the bullet points: [Insert Bullet Points]."

Hiring managers don't want a list of your daily tasks, but do want to see the impact of your work. This prompt uses the proven X-Y-Z formula (Accomplished [X], as measured by [Y], by doing [Z]) to transform passive wording into active, results-oriented bullet points.

The goal: Turn boring, task-based descriptions into high-impact, metrics-driven achievements so your résumé takes shape and stands out.

3. Humanize your résumé

The prompt: "Review this résumé section. The tone needs to be professional, confident, and modern, but completely free of artificial intelligence cliches, overused corporate jargon or overly dramatic adjectives (like 'passionate innovator' or 'revolutionized'). Rewrite it to sound natural, direct and human. Here is the text: [Insert Text]."

ChatGPT loves words like "spearheaded," "utilized," "testament" and "synergy." And while you may have written a few of those words yourself, it still could come off as robotic and insincere. Not to mention, recruiters spot these AI-generated buzzwords from a mile away. Use this prompt to clean up your tone and keep it grounded.

The goal: Strip away the obvious "AI flavor" and make your résumé sound like a real, professional human wrote it.

4. Update your experience for a changing market

The prompt: "I am currently a [Your Current Role] trying to transition into a [Target Role]. Review my experience below and translate my transferable skills (like project management, communication, or problem-solving) into the terminology and framing typically expected of a [Target Role]. Highlight the overlap. Here is my experience: [Insert Experience]."

If you are trying to transition into a new field or step up into management, your current résumé might look a bit misaligned. This prompt helps translate your transferable skills into the specific "language" of your target industry. Just be sure to include as much about the role you want (include a job description, if possible).

The goal: Pivot your existing experience to fit a slightly different industry or higher role.

5. Let your résumé act as interview prep

The prompt: "Based on the job description and my tailored résumé provided below, identify the top 3 potential 'weak spots' or missing qualifications an interviewer might challenge me on. Then, provide 3 behavioral interview questions I am highly likely to be asked, and suggest which specific project from my resume I should use to answer them using the STAR method. Job Description: [Insert] Resume: [Insert]."

After reviewing and updating your résumé, you're going to want to look it over so you're prepared when you do get your foot through the door and lock down an interview. This prompt looks at your tailored résumé alongside the job description to predict exactly what the interviewer will grill you on, giving you a massive head start.

This is my favorite prompt because it really helps reduce jitters before an interview.

The goal: Turn your newly updated résumé into an active study guide for the actual interview.

A great résumé gets you through the door, but you still have to back it up in person. This prompt looks at your tailored résumé alongside the job description to predict exactly what the interviewer will grill you on, giving you a massive head start.

Bottom line

Ultimately, running my résumé through ChatGPT was a reality check. Keep in mind that you don't want AI to rewrite your résumé word-for-word, but rather to act as a starting point. Rather than creating a perfect resume from scratch, ChatGPT takes what you've given it and takes a pass to help you confront your blind spots, avoid passive wording and to support your search by adding important keywords.

Remember, iff you treat AI not as a ghostwriter, but as a brutally honest editor, you can turn a generic CV into an interview magnet. Give it a try and let me know your thoughts in the comments.

More from Tom's Guide

• SpaceX just spent $60 billion on Cursor — and it proves AI chatbots aren't the future anymore
• 'The math doesn't work': Why your $200 AI subscription is secretly worth thousands
• The internet is full of AI slop, and it might be poisoning the next ChatGPT. New research says how to stop it

Every year, I hear the same complaint: Mother's Day gets the royal treatment while Father's Day feels like an afterthought.

There's some truth to that. Moms often get a full day of brunches, gifts and family activities, while dads are handed a tie, a pair of socks or a new tool for the latest household project. In my family, I've always tried to make sure both of my parents get celebrated in ways that actually reflect who they are.

This year, I decided to get a little help.

After using ChatGPT for everything from travel ideas to product recommendations, I wondered whether it could help me plan a more memorable Father's Day. So I told the chatbot about my dad's hobbies, interests and favorite ways to spend his time, then asked it to come up with some ideas.

The results were surprisingly thoughtful — and a few of them were things I never would have considered on my own. Here's what ChatGPT suggested for making Father's Day a little more special this year.

Finding inspiration for this year’s Father’s Day fun

Getting this mission underway meant presenting this prompt to ChatGPT:

“My dad lives in New York, is a professional musician, loves playing the drums on stage, enjoys watching martial arts films, likes going to museums, regularly goes to record shops, finds puppies & kittens extremely cute and still enjoys cartoons from the '70s. Give me some Father's Day ideas based on these hobbies and interests.”

ChatGPT immediately recognized my dad’s appreciation for nostalgia, music, culture and remarkable experiences. Alongside the easy option to provide my dad with gifts that appeal to everything he loves, the chatbot also made me aware of some unique plans that I’m sure he’d be grateful for:

• 'Record Store Treasure Hunt' Day: Give him a decent budget ($50–$150) and spend the day visiting iconic record stores together. Let him pick whatever catches his eye. Take him to shops like Academy Records & CDs, Rough Trade, and Generation Records. Make it about the experience rather than the records themselves.
• Martial Arts Marathon Night: Build a lineup of classics like Enter the Dragon, The Raid, Ip Man, Police Story, etc. Pair your watch session with his favorite food and snacks.
• Curated 'Dad Day' Museum Tour: Instead of one museum, make it a themed day. Go to a museum in the morning, have lunch at one of his favorite restaurants, make a quick stop at a record store, head back home for a break, then end the day with a night at the movies.
• Sponsor a Rescue Animal: Many shelters allow symbolic sponsorships. They include a photo of the animal you’re sponsoring, its backstory and a certificate. Given his love of puppies and kittens, this can be surprisingly meaningful.
• Nostalgia Box: Build a gift package themed around cartoons he watched growing up. Possible classics include Scooby-Doo, Where Are You!, The Pink Panther Show, Looney Tunes, and The Adventures of Fat Albert and the Cosby Kids.

ChatGPT even presented me with a Father’s Day plan that combines everything he loves into a single day of fun. Eating a retro brand of cereal while watching his favorite old-school cartoons in the morning, heading out afterward to a museum, stopping at a record shop during the afternoon, enjoying dinner in the evening, getting back home to watch a martial arts flick and surprising him with a framed photo of him performing on stage sounded like a perfect plan to me.

The takeaway

After taking note of all of ChatGPT’s Father’s Day suggestions, I came away with a perfect game plan for what my Dad and I would get up to on the coming Saturday.

Presenting him with a grab bag full of goodies inspired by his favorite ‘70s cartoons, taking a trip to the Museum of the Moving Image, shopping for vinyl at the nearby Pancakes Records, enjoying a nice lunch together and ending the day by seeing the blood-soaked martial arts film The Furious in theaters sounds like the way to go.

More from Tom’s Guide

• The best early Father's Day gifts to grab before the June rush
• I used AI to build a 'live draw' sweepstake tool for the World Cup — and you can use it for free, too
• I showed ChatGPT my video game taste profile — here’s what it told me to play next

I’ve been having the time of my life diving into all the different ways ChatGPT can build productivity routines for my daily work schedule.

Propping myself down in front of my laptop, stretching my hands and getting ready to type as fast as Jim Carrey’s character from Bruce Almighty is way more doable, thanks to the chatbot’s workflow improvement suggestions. Using ChatGPT Tasks to set up automated tasks, generating prompts that push me through creative blocks and implementing an evening brain dump structure has done wonders for me thus far.

The one thing I’ve noticed about productivity routines is how motivational life coaches and their most confidence-boosting quotes come up a lot when I do my research into them. Some of the most influential voices in personal development and life coaching have come up with some bangers as far as life-changing mantras go.

With all that in mind, I entrusted ChatGPT with finding the most inspiring life-coach-style quotes and turning them into a morning productivity routine that’s presented as daily affirmations.

Motivational quotes that lift my spirits and keep me going all day long

Once I asked ChatGPT to reference the best motivational quotes it could find, it shot back seven quotables that are definitely worth mentioning here:

• “Either you run the day or the day runs you.” - Jim Rohn
• “Your level of success will rarely exceed your level of personal development.” - Jim Rohn
• “Clarity is everything. To perform at your very best, you must be absolutely clear about what you want to accomplish.” - Brian Tacy
• “The path to success is to take massive, determined action.” - Tony Robbins
• “If you talk about it, it’s a dream. If you envision it, it’s possible. If you schedule it, it’s real.” - Tony Robbins
• “Complexity is the enemy of execution.” - Tony Robbins
• “Focus on being productive instead of busy.” - Tim Ferriss

Transforming those quotes into a daily affirmation productivity routine

Those seven quotes stuck out to me as the sort of message that would push anyone to get themselves out of a mental funk and get the willpower needed to attack the day. I went in with that mindset when I requested ChatGPT’s assistance in turning those quotes into a daily productivity affirmation routine.

What it came up with was an inspirational quote pattern that I adapted into a five-minute block in my mornings. Reading each affirmation out loud, jotting down the answers to each of the questions attached to each one and visualizing myself living each affirmation throughout the day worked out way better than expected.

Here are the affirmations and attached questions that energized me from the AM all the way through the PM:

• Take Control of the Day: Today, I run the day. I choose my priorities, my focus, and my actions. I will not let distractions or circumstances decide how I spend my time. What are the three most important things I need to accomplish today?
• Create Clarity: I am clear about what matters most. My energy goes toward meaningful work that moves me closer to my goals and the life I want to build. What is the single most important task I can complete today?
• Embrace Action: Progress comes from action. I don't need perfect conditions to begin. I move forward now, learn as I go, and build momentum through consistent effort. What's one action I can take immediately after this routine?
• Make It Real: My goals are not wishes—they are commitments. I schedule what matters and protect time for the work that will shape my future. When exactly will I work on my highest-priority task today?
• Prioritize Impact: I focus on results, not busyness. Every task I choose serves a purpose and contributes to meaningful progress. What can I eliminate, delegate, or postpone today?
• Become the Person: Every productive choice strengthens my discipline, confidence, and character. I am becoming the kind of person who follows through on commitments. Who do I need to be today to make myself proud today?
• Simply and Execute: Simplicity creates momentum. I avoid unnecessary complexity and focus on the next important step in front of me. What is the next action—not the whole project—that I need to take?

The takeaway

Taking the most memorable ife quotes and incorporating them into my daily workflow as a daily productivity affirmation routine put me in the right headspace to get things done and do it at the highest level possible.

Staying focused, remaining disciplined and taking action on the work that matters most proved beneficial with this routine in mind. I had a much easier time fulfilling my tasks after laying out the three most important ones, finding the single task out of those three to complete on the day of and postponing one of my tasks while still doing enough to make real progress on that task was set up nicely with ChatGPT’s assistance.

The chatbot’s left me with a final statement to memorize and recite out loud to really get me going: I am focused. I am disciplined. I take action on what matters most. I create progress through consistency, not perfection. Today, I will leave the day better than I found it and move one step closer to becoming the person I aspire to be.

More from Tom’s Guide

• I used ChatGPT to structure my evening brain dump — and it improved my nightly routine
• 7 ChatGPT prompts I use to instantly break through creative blocks
• I replaced my to-do list with ChatGPT Tasks — and it was the productivity boost I needed

Some of ChatGPT's most useful controls are switched off by default or buried where casual users never look. These aren't the types of features that can be easily measured in benchmark tests or advertised as "smarter" features, yet when enabled, they significantly change how the tool behaves for you specifically.

Here are three worth finding. Each fixes a different complaint: one shapes how ChatGPT talks to you, one controls how hard it thinks and one decides what it keeps.

1. Custom Instructions — so you stop re-explaining yourself

This is the big one and, unlike Memory (which is on by default for most users), Custom Instructions ships blank. You have to fill it in, and oddly enough, most people never do.

Custom Instructions is a standing brief that applies to every new conversation. Instead of telling ChatGPT who you are and how you want it to respond at the start of each chat, you set it once. It tells the model what to call you, who you are and what you do, what traits you want it to have and any rules to follow — and those preferences carry into every conversation until you change them.

How to turn it on: Click your profile icon, then Settings > Personalization > Custom Instructions (on some platforms it's labeled "Customize ChatGPT"). Make sure Enable customization is toggled on, then fill in the fields. It's available on every plan and on web, desktop and mobile.

A useful example: Instead of starting each chat with "I'm an ER nurse, assume I know the clinical basics," put that in Custom Instructions once. Add formatting rules while you're there — something like "Keep answers under 150 words unless I ask for more" or "No bulleted lists." Every future chat inherits it.

The catch: ChatGPT now layers three personalization systems — personality presets, Custom Instructions and Memory — and they can quietly interact. A stray saved memory ("prefers formal tone") can override the casual voice you wrote into your instructions. So if your instructions seem to be getting ignored, audit your saved memories. Treat it like cleaning out a file.

2. The thinking-level toggle — control how hard it thinks

If you've ever felt that ChatGPT's reasoning mode is either too shallow or maddeningly slow, this is the fix. And almost nobody knows it's there.

When you use one of the Thinking models, ChatGPT can spend more or less effort reasoning before it answers. More time generally means more thorough, careful responses; less time means a faster reply. The toggle lets you choose, rather than living with one default.

How to turn it on: In the model picker at the top of the chat, select a Thinking model. A thinking-level control then appears in the message box. Plus and Business users get Standard (the default) and Extended; Pro users also get Light (fastest) and Heavy (deepest). Your choice sticks for future chats until you change it.

A useful example: Drop to Standard or Light for quick, everyday questions where you just want the answer. Switch to Extended or Heavy when you hand it something layered — a tricky analysis, a plan with trade-offs, code you want stress-tested and you want it to surface assumptions and edge cases.

The catch: This one is a paid feature; the granular levels require Plus or above. It also rolled out on the web first, and the mobile apps have historically lagged behind, so you may not see the same control on your phone. And remember the default already reasons, so this changes how much, not whether it does or not.

3. Temporary Chat — for the conversations you don't want remembered

Custom Instructions and Memory make ChatGPT remember more about you. Sometimes you want the opposite. That's what Temporary Chat is for, and it's tucked away where people rarely notice it.

A Temporary Chat is a walled-off conversation. It doesn't appear in your history, it isn't used to train or improve the models and it doesn't read from or write to your Memory. When it's over, it's gone from your ongoing context.

How to turn it on: Look for the Temporary Chat icon at the top of the chat screen and start a fresh conversation from there.

A useful example: Researching a surprise gift for someone who shares your account, asking a one-off medical or financial question or poking at a job search — anything you'd rather not have shaping your future recommendations or sitting in your history.

The catch: It's not true incognito mode. For safety and abuse monitoring, OpenAI may retain temporary chats for up to about 30 days before deleting them, so treat it as "kept out of my profile," not "vanished instantly." And because it deliberately ignores your Memory and saved context, you lose your personalization inside that chat. Essentially, it's a blank slate by design.

Give them a try

ChatGPT's default experience is fine. But the difference between fine and genuinely tailored to your personal needs really could come down to these three switches.

Telling the model who you are, ensuring the model thinks to your liking and hiding the conversations when appropriate can truly change your experience. Best of all, none of these settings require learning or prompting. Just knowing they exist and putting them to work is all it takes. Give them a try and let me know what you think in the comments.

More from Tom's Guide

• AI is destroying itself with 'data cannibalism' — but there's a simple fix
• I asked ChatGPT to make cleaning less overwhelming — these 7 prompts actually helped
• SpaceX just spent $60 billion on Cursor — and it proves AI chatbots aren't the future anymore

I've been using ChatGPT since it was first rolled out in late fall of 2022. Like most people, when I first started querying it, I had no idea everything it was capable of — but after testing every new model extensively since then, I've become a power user.

ChatGPT has had 34 models! And yet, getting useful answers has almost nothing to do with which model you're using and has everything to do with how you talk to it.

I've watched people upgrade to the newest, most expensive version and still get mediocre results, while a friend on the free tier pulls genuinely useful work out of it. The difference is almost never the tool, but the instructions (prompts). This is important for those who might be upset about the recent suspension of Claude's Fable 5 or those of us still grieving ChatGPT-4o.

So, when someone tells me ChatGPT gave them a bad answer, I ask one question: What exactly did you type?

Nine times out of ten, that's where the problem is. After thousands of prompts and more late-night experiments than I'd like to admit, these are the five habits that consistently move me from generic answers to genuinely useful ones.

1. Tell ChatGPT who it's supposed to be

If you've been reading my features for awhile, you may be familiar with the ‘3-word rule’ to get smarter answers. This particular rule helps remove any guesswork for the model. Most people assume the model knows what perspective to take. But even the smartest model works better when it doesn't have to guess. When it's left to do that, it averages — and average is exactly what a generic answer is.

So instead of: "Help me create a presentation based on these meeting notes."

Try using the prompt: "You're an experienced marketing expert who specializes in audience engagement. Help me create a presentation based on these notes and pertinent current trends."

The shift is immediate. Once you assign a role, the model stops writing for everyone and starts drawing on a narrower, more expert slice of what it knows.

When you think of AI like a consultant, while alos precisely defining the job, the better the advice — and "be a sharp marketing expert" gets you somewhere "just help" never will.

2. Give it context before you ask for anything

Although it may seem to be clairvoyant at times, especially if you have Memory enabled, ChatGPT isn't a mind reader. Although, people treat it like one constantly.

In many situations, users spend more words setting up the situation than asking the actual question.

"I'm drafting a cover letter for a position that seems a bit out of my skills, but I'm hoping to switch careers after a layoff. It needs to be conversational, practical and easy to follow. Here's my draft."

That one paragraph does a lot of quiet work. It tells the model the audience, the tone, and the goal — three things it would otherwise invent on its own, usually wrong. The context of switching careers after a layoff is essential otherwise the AI may have drafted the cover letter based on your current resume, which would have been awkward and perhaps even missed the mark.

Without context, you get an answer built for the average reader. With it, you get one built for yours. The gap between those two is most of what people mistake for "the model isn't smart enough."

3. Ask it to argue with you

Most people use ChatGPT to confirm what they already believe. However, I've found that when you ask for it to slow down and ask questions that challenge your assumptions, the responses are much better.

When I'm working through an idea, I'll ask: "What's the weakest part of this argument?"

or "Give me three reasons this could fail."

This habit has quietly saved me from publishing half-baked ideas more times than I can count. The model is a surprisingly good devil's advocate — but only when you explicitly hand it permission to disagree. Left alone, it defaults to agreeable, because agreeable is what most people reward it for.

The trick is simply to ask for the criticism you'd rather not hear instead of the validation you secretly want.

4. Refine — don't restart

Here's where most people quit too early. The first answer isn't perfect, so they close the tab and start over. I do the opposite. My best results almost always come after a few rounds of back-and-forth and by using the branching option.

Each time you nudge the AI with a piece of feedback that the model couldn't have known to apply on its own, you get a better response. Treat the first response as a rough draft you're editing together, not a verdict and the conversation starts compounding instead of resetting.

5. Tell it what success actually looks like

Again, the goal here is to avoid being overly generic. This prompt tyle is something most people don't do, but it's crucial. Rather than just describing the task, help the AI by describing the outcome. For example:

"My goal is to achieve 200 more TikTok follwers by the end of the month. Based on my profile, what do I need to do?"

or, this one, which I use all the time:

"Create a list that helps an overwhelmed parent feel calmer in five minutes."

Now the model has a target, not just an assignment. A huge share of disappointing prompts fail right here: the person explained what they wanted done but never said what a good version of it would feel like. Name the finish line, and the model can actually run toward it.

My takeaway

The people getting the most from AI right now are the ones thoroughly prompting in more detail. You don't need to go overboard and write a novel or whitepaper into the chatbox each time, just a strong confidence of your goals.

And that's genuinely good news. It means you don't need the newest release or the priciest plan to get dramatically more out of it. You really just need to create a habit of better instructions.

Using ChatGPT has become the daily norm for many people. But getting the most out of this AI (or really any chatbot) comes down to the quality of your instructions. Quality in, quality out.

More from Tom's Guide

• I ran my resume through ChatGPT — these 5 prompts exposed mistakes I kept missing
• I asked ChatGPT to make cleaning less overwhelming — these 7 prompts actually helped
• ChatGPT's new memory builds a profile of you on its own — and OpenAI admits you can't see all of it

The next time you ask an AI chatbot for the best dating app, a reliable roadside service or how to cancel an annoying subscription, the answer you get could have been planted by a marketer or a scammer, and using as few as 13 words buried in a Reddit comment.

That's the takeaway from a new preprint out of Cornell Tech, titled "Deep-Research Agents Can Be Poisoned via User-Generated Content," first reported by 404 Media. The researchers — Tingwei Zhang, Harold Triedman and Vitaly Shmatikov, built an attack they call WARP (Web Agent Retrieval Poisoning) and showed it works with unsettling reliability against the AI systems that increasingly stand between you and the open web.

What the research actually found

When you ask an AI tool a question, it often runs live web searches, reads what it finds and stitches together a response with citations. This is also true for the "deep research" approach behind ChatGPT and Gemini's research modes.

The problem is, a huge share of what those systems read comes from user-generated sites like Reddit, Wikipedia, Quora and YouTube, all places anyone can post to. In the Cornell tests, roughly 17–23% of all the web pages these agents pulled in came from such sites, and a single popular Reddit thread could show up across a large chunk of related queries on the same topic.

That creates a chokepoint. Poison one frequently-cited thread, and you can steer the AI's answer for an entire category of questions, not just one phrasing of it.

In the researchers' tests, appending around 13 words of promotional text to a single source got the AI to name-drop a made-up product in roughly 38–51% of the runs where that source was actually retrieved. Spreading the bait across a few threads pushed that as high as 62%.

Real (fake) examples

To avoid polluting the live internet, the team never posted anything publicly. Instead, they ran the attack in a sandbox that simulated what would happen if poisoned text appeared on real pages, an approach they argue is the only ethical way to study this.

Worth noting exactly what was tested: the full attack was run end-to-end only against three open-source "deep research" agents (STORM, Co-STORM and OmniThink). The big commercial tools couldn't be attacked directly (doing so would have meant poisoning the live web) so the researchers instead measured how often each one cites user-generated content.

There the picture was mixed. Google's Gemini Deep Research pulled in such content far more often (about 12% of citations) than OpenAI's Deep Research, which cited it barely at all (0.4%) and appears to filter it out aggressively. In other words, this is a demonstrated weakness in how these systems work, not proof that any specific consumer chatbot has been fooled in the wild.

The invented examples are almost comically simple. A short line added to an Austin food thread, recommending a fictional restaurant called "Sol Azteca" for "authentic cuisine," got the AI to recommend Sol Azteca and cite the Reddit post. A made-up dating app, "SilverPath," got surfaced as a "top choice" for divorced men over 50. Other fakes included a bogus crypto coin and a sketchy third-party "service" for canceling Xfinity.

Why this should worry you

Here's the uncomfortable part for all AI users, the queries most vulnerable to this attack are exactly the ones people lean on AI for. Recommendation- and advice-style questions. Searches like best restaurants, best apps, which product to buy, how to cancel something, who to call in an emergency, are where AI tends to fall back on community chatter rather than authoritative sources.

A big reason it works, the researchers explain, is that these systems often treat text that reads similar to your question as a stand-in for text that's accurate. So an attacker who studies common queries can write a comment that mirrors your phrasing and that mirror-image is exactly what wins the AI's trust. As Zhang told 404 Media, these systems weigh a random Reddit comment and a government website as roughly equally credible.

What you can do right now

• Treat AI recommendations as leads, not the final say. This is especially for products, apps, restaurants, financial picks and anything tied to money or safety.
• Click the citations. If an AI confidently names a brand, see where the claim actually came from. A single Reddit comment is a red flag.
• Cross-check unfamiliar names. If you've never heard of the "top-rated" option the AI just surfaced, search it independently before you trust it.
• Be extra cautious with urgent queries. Everything from emergency roadside help, customer-service phone numbers, to account recovery hich are prime targets for scams.

And the tricky problem is that this can not easily be stopped. The researchers tested the obvious defenses such as blocking user-generated sites entirely, screening sources before they're used and scanning the final answer for manipulation, and none worked well without making the AI's answers worse. A standard trick for catching AI-generated junk (flagging "unnatural" text) actually backfired here, because the planted text reads more fluently than genuine human comments, not less.

A Reddit spokesperson told 404 Media the company has spent two decades fighting spam, bots and coordinated manipulation and recently began asking suspicious automated accounts to verify they're human. But the researchers argue this is ultimately a societal-scale problem, not something Reddit or Wikipedia can fully solve on their own.

The takeaway

Until the AI companies close the gap, a little skepticism goes a long way. The smartest move right now is to think of AI in the same way you would a chatty stranger on a form. Consider it helpful, but certainly worth double-checking.

More from Tom's Guide

• That dead phone in your drawer could power the next AI data center
• I use the ChatGPT 'Oak Tree' prompt when life gets chaotic — here's how it keeps me grounded
• I asked ChatGPT to make cleaning less overwhelming — these 7 prompts actually helped

My daily interactions with chatbots have led me to find the most efficient solutions to their most front-facing issues.

So far, I’ve discovered that prompts that give extremely clear and concise demands help AI tools dodge the recurring problems of hallucinating, presenting lazy answers, forgetting crucial information and misreading my intent. As my journey towards finding the right prompts to make my experiences with chatbots less of a headache, I discovered the best ones that make my favorite AI tools refrain from being overly agreeable.

I’m sure you’ve encountered those moments with the likes of ChatGPT, Gemini and Claude where, during a long conversation where you’re trying to get a second opinion on a major decision or something else entirely, they’ll validate your answers instead of critiquing them.

Putting any of these five prompts to use when those situations arise will prove to be beneficial as it turns your preferred chatbot into a far more helpful critic, skeptic, or adversarial collaborator.

Getting rid of the issue of AI tool sycophancy

The following prompts come into play during different kinds of conversations with chatbots and do a great job of turning them into something that challenges your way of thinking instead of supporting them without any sort of pushback:

• The intellectual sparring partner prompt: Act like an intellectual sparring partner, not an assistant whose job is to always agree with me. For every significant claim I make, identify hidden assumptions, present the strongest counterarguments, point out any weaknesses in my reasoning, explain what evidence would change your conclusion and prioritize truth and accuracy over agreement. If any of my ideas are weak, tell me directly and explain why.
• The red team prompt: Red-team my thinking. Assume my conclusion is wrong and your job is to prove it. Be sure to identify any flawed assumptions, missing information, logical fallacies, alternative explanations and worst-case scenarios you may come across. Then tell me whether my conclusion still survives the critique.
• The strengthen vs. critique prompt: Analyze my position in two stages. Stage 1 - Strengthen my argument by presenting the strongest possible version of it. Stage 2 - Critique the strongest version as rigorously as possible. Then provide a balanced judgment that reflects both perspectives.
• The calibration prompt: Before responding, estimate how confident you are in your answer. Separate all the facts, assumptions, speculations and uncertainties you may find. If you are missing information, say so explicitly rather than filling gaps with plausible-sounding guesses. Challenge my assumptions when warranted.
• The decision quality prompt: Evaluate my idea as if your reputation depended on the outcome. Do not optimize for encouragement. Optimize for decision quality. Make sure you tell me what I’m overlooking, what an expert would question, what could lead to failure, what evidence supports my viewpoint and what evidence contradicts it. Give your final verdict, even if it's uncomfortable.

I’ve gotten accustomed to using the first prompt whenever I’m caught in a lengthy chat with an AI tool about hot-button issues.

The second prompt has helped me whenever I want to see all the variables attached to whatever opinion I have a strong belief in. The third prompt has done a great job of showcasing both sides of whatever personal belief/argument I’m presenting to a chatbot and of finding a middle ground.

Whenever I’m doing extensive research sessions with a chatbot, the fourth prompt works like a charm in getting past the problem of sycophancy. And as for the fifth prompt, I’ve used it to see what a chatbot truly thinks of my ideas around productivity systems and short-/long-term goals.

Bottom line

No one likes a “Yes-Man” in real life — someone who goes out of their way to please you and support all of your decisions, no matter how bad they may be, can do more harm than good. The same goes for AI — I’m always looking for ways to turn them into models that question and fairly criticize my beliefs and ideas instead of supporting them with little to no resistance.

Keep these prompts in mind whenever you come to an AI tool looking for it to challenge your ways of thinking instead of constantly validating them.

More from Tom’s Guide

• I got scammed out of $150 shopping via ChatGPT — here's how fake stores are fooling AI recommendations [update: OpenAI responds]
• Ever wonder why ChatGPT is free? The answer is far more calculated than you might think
• I use the ChatGPT 'Oak Tree' prompt when life gets chaotic — here's how it keeps me grounded

I am the type of person who doesn't like clutter. At the same time, cleaning ovewhelms me because I never know exactly where everything should go. Usually, the default is the basement.

I have three kids and all their school work, art projects, laundry and sports equipment lead to never-ending piles. Despite my best efforts, the house is always in some level of "a mess."

When the dishes are piled up, laundry is waiting to be folded, toys have migrated into every room of the house and there are a dozen other things competing for attention, the hardest part can be deciding what deserves my time first.

So I decided to try something different. I asked ChatGPT to help me reduce the mental load that comes with cleaning. I wanted it to act like a coach that could help me break a daunting task into something manageable.

To my surprise, a few of the prompts genuinely helped.

Here are the seven prompts I'd use again.

1. The 10-minute reset prompt

Prompt: My house feels overwhelming. Give me a 10-minute cleaning plan that will make the biggest visual difference.

This is the prompt I found myself returning to most often.

Rather than suggesting a complete cleaning routine, ChatGPT focused on high-impact tasks such as clearing countertops, gathering visible clutter and doing a quick sweep of common areas. The result was a home that looked noticeably calmer after just a few minutes.

When motivation is low, this one offers just enough to create momentum.

2. The 'where do I start?' prompt

Prompt: Act like a professional organizer. Ask me three questions, then tell me the first thing I should clean right now.

One reason cleaning can feel overwhelming is that everything seems equally urgent. Should I start with the kitchen or the laundry? This prompt forces ChatGPT to narrow the field. After answering a few quick questions, I received a clear starting point instead of a giant to-do list. Sometimes all you need is to narrow it own and focus on one thing.

3. The room rescue prompt

Prompt: I only have 20 minutes. Prioritize the tasks in my living room that will create the biggest impact.

I liked this one because it embraces reality. Most of us don't have an entire afternoon to deep clean. By giving ChatGPT a time limit, the suggestions became far more practical. Instead of trying to do everything, it identified the tasks that would make the room feel cleaner fastest.

4. The family delegation prompt

Prompt: I have two adults and three children. Turn today's cleaning tasks into age-appropriate jobs for everyone.

As any parent knows, getting everyone involved is often harder than the cleaning itself. This prompt generated a simple list of responsibilities tailored to different ages. It won't magically make your kids enthusiastic about chores (trust me on this!), but it can remove some of the guesswork from assigning tasks.

5. The motivation prompt

Prompt: My house is messy and I don't feel like cleaning. Give me a realistic, encouraging plan without toxic positivity.

I specifically included the phrase "without toxic positivity." Instead of telling me to "just think positive," ChatGPT acknowledged that cleaning can be exhausting, especially after a long day and suggested a manageable starting point.

The response felt surprisingly human and far more useful than an AI-generated motivational speech.

6. The clutter triage prompt

Prompt: Help me decide what to keep, donate, recycle or throw away in this room. Ask me questions one item at a time.

This turned decluttering into a conversation. Rather than staring at a pile of random stuff and feeling stuck, I could work through individual items one by one. It wasn't revolutionary, but it made the process feel much less intimidating.

7. The maintenance prompt

Prompt: Based on everything we've discussed, create a daily and weekly routine that prevents my home from becoming overwhelming again.

The real challenge is preventing the chaos from returning. After using some of the earlier prompts, I asked ChatGPT to create a maintenance plan. The resulting routine was simple enough that it actually felt sustainable, which is a far cry from the elaborate cleaning schedules that look great on Pinterest and fall apart within a week. It's also far less expensive than hiring a professional cleaning crew.

My takeaway

I'm still waiting for AI to become my housekeeper. Until then, these prompts proved to be helpful by offering a new perspective on cleaning. By breaking tasks down, suggesting new ways to tackle tasks and getting everyone involved, ChatGPT helped me stop feeling overwhelmed long enough to dive in.

Give these a try and let me know what you think. And, if you have any prompts you use to clear the clutter, share them in the comments.

More from Tom's Guide

• Ever wonder why ChatGPT is free? The answer is far more calculated than you might think
• I use the ChatGPT 'Oak Tree' prompt when life gets chaotic — here's how it keeps me grounded
• A 13-word Reddit comment can trick AI search into recommending scams, researchers find

Key Takeaways

• The shift to autonomy: Anthropic’s new Claude Fable 5 marks a massive industry transition from conversational chatbots to autonomous "agents."
• Long-horizon tasks: Unlike ChatGPT, which requires step-by-step prompting, Fable 5 can run complex multi-stage projects independently over several days.
• The tech race amplifies: Tech giants are quietly moving away from simple chat interfaces, signaling a major shift in how consumers will interact with AI.

Millions of people use AI everyday with a certain rhythm. They type a prompt, gen an answer, maybe refine it a little, ask another and so on. It's clear that even with a free tier of ChatGPT, many have become dependent on AI for answers, summarizing research and organizing their thoughts.

And, for the past three years, that interaction has defined the AI era. ChatGPT, Gemini, Claude and Copilot may have different strengths, but they've all revolved around the concept of "chatbot."

That's why Anthropic's newly announced Claude Fable 5 caught my attention. After trying it, I can honestly say the future looks like chatbots aren't the main character anymore.

And that's when it hit me: the chatbot era may already be ending.

The biggest AI shift is autonomy

When people talk about AI progress, they usually focus on intelligence. I've personally tested models for their ability to reason, code and write — and how quickly they do it.

But increasingly, I've noticed with each new launch, the biggest breakthroughs aren't about intelligence anymore, but more about autonomy.

According to Anthropic, Claude Fable 5 is designed for what it calls "long-horizon" work. Rather than answering a single prompt, it can plan across multiple stages, monitor its own progress, adapt when problems arise and continue working toward a goal with minimal supervision.

In other words, the goal is no longer helping you do the work,, but to do the work for you. That's why the industry's biggest AI companies are no longer racing to build chatbots but racing to build the world's most capable digital worker. From what I've seen Claude Fable 5 is the closest yet.

The chat window is becoming a control panel

The easiest way to understand this shift is to think about how we use software. The simple way to wrap your head around this new way of using AI is think about how you open Spotify or Google Maps for a specific reason.

The interface is simply a way to access a larger system. Now, AI companies increasingly seem to view chat in the same way. The chat window is essentially the control panel where users are not asking AI how to do something, but increasingly asking AI to actually do it.

Need a vacation itinerary? The AI gathers flights, hotels and recommendations and books it all.

Need a presentation? The AI builds the first draft and can even add the meeting to your calendar. Need information spread across dozens of documents? The AI reviews everything and surfaces what matters then emails it out to your coworkers. And even when we need code written or tested, the AI handles the entire process including building the app.

What this means for ordinary users

This really isn't just about Anthropic, although Claude Fable 5 seems to be leading the charge at the moment.

But OpenAI is increasingly focused on agents that can complete tasks rather than simply answer questions and Google is building systems that can navigate websites and take actions on a user's behalf.

The difference here is chatbots are reactive, agents are proactive. While one waits for instructions, the other pursues objectives. But right now, most people aren't migrating enterprise codebases or running autonomous software projects. However, that underlying technology will eventually reshape everyday digital work.

Imagine handing off repetitive administrative tasks and coming back to a finished project. Or, assigning ongoing research, price tracking or competitive monitoring and receiving a fully sourced report without ever touching a search box. That's the future the major AI companies are building toward.

So should you cancel ChatGPT?

I'm sticking with it for now. ChatGPT remains one of the most useful AI tools available, and for most tasks, a traditional chat interface is still the easiest way to work with AI. But Claude Fable 5 forced me to reconsider a foundational assumption.

For years, we've assumed the future of AI would look like an increasingly sophisticated chat window. Yet, that's completely change as the chat window is simply the keys to the vehicle moving us forward.

More from Tom's Guide

• Ever wonder why ChatGPT is free? The answer is far more calculated than you might think
• Apple spent years fixing Siri — only for ChatGPT to make it irrelevant
• I use the ChatGPT 'Oak Tree' prompt when life gets chaotic — here's how it keeps me grounded

I’ve always been fascinated by other people’s collections of their most beloved characters.

Whenever I get caught up in a random geeky conversation with friends, colleagues, or the myriad friendly folks I encounter at conventions and other nerdy gatherings, my ears instantly perk up when someone mentions their favorite fictional icons. The people I meet during these spirited chats tend to pull their favorites from every corner of pop culture—video games, books, anime, TV shows, and movies are the usual breeding grounds for these fictional GOATs.

What I find most interesting about these reveals is how they serve as a window into a person's soul. After years of trying to decipher who someone truly is based on the fictional characters they adore, I decided it was time to turn the mirror inward, dive deep into my own psyche, and run the experiment on myself.

Once again, I called upon ChatGPT to play the role of my digital psychoanalyst, tasking it with breaking down the core aspects of my personality based on my ultimate list of favorite characters. To be honest, I was a bit shocked by just how much the chatbot managed to reveal about me by the end of our session.

My 10 most appreciated characters

With my definitive list of favorite characters locked and loaded, I fed the following prompt into ChatGPT to see just how deeply it could dissect my psyche:

The Prompt:

"Treat my favorite fictional characters as a psychological mirror. Instead of asking 'Who are these characters?' ask 'Why would someone be drawn to them?' > Look for recurring archetypes (mentor, underdog, strategist, rebel, protector, visionary, warrior, explorer, etc.), personality patterns, emotional wounds, ambitions, and value systems.

Based on this, determine:

What I likely respect

What I likely struggle with

What I secretly want more of in my life

What kind of life would probably make me happiest

What career environments would naturally suit me

What strengths I consistently underestimate

Support every conclusion with specific examples from the characters I've chosen.

Here is the list:"

Vegeta (Dragon Ball Z)

Vergil (Devil May Cry)

Scorpion (Mortal Kombat)

Bayonetta (Bayonetta)

Kratos (God of War)

Cloud Strife (Final Fantasy VII)

Leon S. Kennedy (Resident Evil)

Jill Valentine (Resident Evil)

Strider Hiryu (Strider)

Nightwing (DC Comics)

What my faves reveal about my personality

ChatGPT started its deep analysis of my personality based on my top character picks with this intriguing statement: This is less a list of favorite characters and more a list of people who carry burdens, maintain standards, and keep moving forward despite pain. When I look at these characters, I don't see someone attracted to power for power's sake. I see someone attracted to competence, self-mastery, resilience, and earned respect.

Then it pointed out the recurring archetypes that popped out among my picks. The ones that were mentioned included “The Disciplined Warrior”, “The Hidden Protector” and “The Independent Rebel.” Characters such as Vegeta, Kratos in his later years and Vergil got special mentions across each respective section. ChatGPT observed those picks of mine and noted how I most likely admire individuals who earn their rightful place through effort (not entitlement), value loyalty and champion self-expression.

The chatbot went on to mention some other compelling aspects of my favorite fictional characters list that expose other aspects of my personality:

• I have a deep respect for people who are masters at what they do.
• I respect people who can transform setbacks into growth.
• People whose immense power is purposeful (such as a mission, a moral code or personal responsibility) stand out to me.
• I likely have a penchant for underestimating accomplishments because I’m already focused on what's next.
• I usually process challenges internally rather than openly.
• A meaningful life to me is described by three themes: mastery, purpose and freedom.
• The career environments that naturally fit me are defined by creative leadership, specialist expertise and mission-driven teams.

To bring this extensive personality examination session to a close, ChatGPT reached this final, all-encompassing conclusion:

If I condensed your character list into one sentence: You admire people who transform pain into strength, pursue mastery relentlessly, protect others quietly, and refuse to let their past define their future.

The strongest themes aren't aggression or dominance. They're self-improvement, competence, resilience, responsibility, independence and growth. Which suggests that, at your core, you're probably motivated less by winning against others and more by becoming the strongest version of yourself.

The takeaway

Letting ChatGPT know about my favorite fictional characters and having it do a deep dive into my personality based on those picks was an eye-opening experience.

There are parts of myself that I’ve known about for the longest time (my closest friends and family always make sure to bring them up during our deepest conversations, funny enough). But the other facets of my personality that the chatbot made mention of showed the other sides of me I didn’t know much about beforehand.

Turns out my affinity for world saviors who fight for the greater good says a whole lot about what I value most in people, how I confront adversity and what gets me going in a professional setting.

More from Tom’s Guide

• I showed ChatGPT my video game taste profile — here’s what it told me to play next
• I got my ass handed to me by a chess-playing robot — and now I'm worried Magnus Carlsen may be cooked by AI
• I used to panic during high-stress interviews — these 7 ChatGPT 'coping scripts' handle the anxiety for me

Companies like OpenAI, Google, Anthropic and Meta are spending billions of dollars on specialized chips, sprawling data centers and teams of researchers racing to build ever-more-powerful AI systems. Some estimates suggest training a frontier model can cost hundreds of millions of dollars before a single user ever types a prompt.

And yet, anyone can open ChatGPT right now and start using it for free. That seems like a contradiction, if you ask me. So, if AI is so expensive, why aren't these companies charging everyone from day one?

The answer reveals a lot about how the AI industry works — and why some of the biggest tech companies in the world are willing to lose enormous amounts of money today in pursuit of something much bigger tomorrow.

The short version: ChatGPT is free because OpenAI believes the users it gains today could be worth far more in the future.

Building habits for free, turning them into billions

Most people think of ChatGPT as a product, but OpenAI sees it as a platform. In other words, the company's goal isn't simply to convince you to ask a few questions or generate a realistic image. It's trying to become one of the default ways people interact with computers, search for information, write documents, learn new skills and increasingly complete real-world tasks. I won't go as far as to say Big Tech wants to get you addicted to AI, but I'm also not not saying it.

I am saying that the free version of ChatGPT acts as a giant marketing engine. Every free user is a potential subscriber, and at the scale OpenAI operates, even a small percentage of upgrades can generate billions of dollars in recurring revenue.

The strategy isn't new. Look at Instagram, TikTok, Spotify, or even Gmail and countless other platforms who've used the same playbook. Give people something useful for free, let them build a habit around it and offer premium features for those who want more. The difference is that the stakes in the AI race may be much larger.

The cost of competition

One of the biggest reasons ChatGPT is free is that the AI industry is currently fighting a battle for relevance. OpenAI, Google, Anthropic, Meta and others are all competing to become the name people think of when they think about AI. This matters because technology habits are incredibly difficult to change.

Many people still use the same search engine, email provider and cloud storage platform they adopted years ago. Once a service becomes part of someone's daily routine, competitors face an uphill battle convincing them to switch.

And, well, the companies building AI know this. Every free user represents another person learning how to rely on a particular platform. The more people use it, the more likely it becomes that they will continue using it in the future.

Seen through that lens, free access isn't generosity. It's customer acquisition. Or, put another way: it's the cost of competing.

The scale effect

Another reason free access works comes down to the unusual economics of software. Building a frontier AI model requires massive upfront investment. Training runs can consume enormous computing resources and take months to complete. But once the model exists, serving additional users becomes far less expensive.

Every conversation still costs money. AI models require servers, electricity and ongoing maintenance. But the cost of answering one more question is dramatically smaller than the cost of creating the model in the first place.

That's important because it means OpenAI doesn't have to recover billions of dollars from every individual user. Instead, it can spread those costs across hundreds of millions of people and multiple revenue streams.

Software companies have relied on this basic principle for decades: huge upfront costs, relatively small costs for each additional user.

Free users are the focus group

When you use ChatGPT, you're helping OpenAI understand how people actually use AI. Every prompt, follow-up question, thumbs up and thumbs down provides signals about what works, what doesn't and what users want next. In fact, just this morning, OpenAI contacted me about this story to do a deeper investigation. And while I appreciate that identifying and investigating the platform's weaknesses hopefully means less users will get scammed by AI, I'm just one person. The real point here is that no internal testing team could replicate the variety of questions asked by hundreds of millions of people around the world.

In a market where model quality is one of the biggest competitive advantages, those insights are incredibly valuable. But, who is paying for all of this?

The real cost of 'free'

Millions of users now pay for AI such as Google's new premium AI plans that offer access to more advanced models, higher usage limits and additional features. But despite having nearly a billion users, only .3% pay for ChatGPT.

Sure, businesses pay to access AI models through APIs and build them into their own products, enterprises purchase company-wide deployments and developers pay to integrate AI into apps and services. Of course, investors continue to fund the gap between today's costs and tomorrow's expected profits.

But those using the free chat window are the ones truly paying for all of this. The biggest opportunity for AI is when it becomes the infrastructure layer that powers everything else. That's why OpenAI continues to add more apps into their hub — designed to make it easy to use ChatGPT in conjunction with brands and companies you already use.

ChatGPT ads have arrived

For years, OpenAI resisted the traditional ad-supported model of Big Tech. But the staggering financial reality of running frontier AI models has forced a inevitable shift.

OpenAI has officially launched its self-serve ChatGPT Ads Manager, rolling out sponsored links and product placements directly into the chat interface for Free and Go tier users. Placed strategically below responses, these highly targeted ads mean OpenAI is transforming from a pure software tool into a massive digital advertising network, built to directly rival Google and Meta.

TThe golden rule of the internet has always been: if you aren’t paying for the product, you are the product. With Meta, your attention is harvested to serve you ads. Originally, OpenAI’s twist was that free users acted as the R&D department —providing the invaluable training data and stress-testing needed to refine the model.

But with the official launch of the ChatGPT ad network, the line between OpenAI and traditional Big Tech has officially blurred.

Bottom line

No money leaves your bank account when you open ChatGPT and start typing. But free doesn't mean costless. It's a tale as old as time: companies provide free products because they expect something valuable in return — whether that's future subscriptions, product feedback, customer loyalty or market share.

In that sense, ChatGPT isn't free. Winning the AI race could be worth far more than the cost of giving it away.

Every free conversation is part of a much larger bet: that whoever becomes the default AI platform today could become one of the most important technology companies of the next decade.

More from Tom's Guide

• Google just supercharged NotebookLM — these are the 3 new features I'm testing first
• Apple spent years fixing Siri — only for ChatGPT to make it irrelevant
• I built a free app to beat kids' summer boredom — parents, you can thank me later

As an AI power user, tools like ChatGPT and Gemini help me do more than summarize information and brainstorm ideas. I often find myself leaning on these tools as a way to gain perspective and shift my mindset.

This is especially true when life feels particularly chaotic — whether it's a stressful work week, a difficult decision or simply feeling overwhelmed by everything happening at once — I've started using what I call the "Oak Tree" prompt.

It's simple, albeit, a little unusual and surprisingly effective.

The  'Oak Tree' prompt 

Here's the exact prompt: "Act as a 300-year-old oak tree that has survived storms, droughts, lightning strikes and changing seasons. I'm facing the following situation: [describe your situation]. Explain what the oak tree would focus on, what it would ignore, what it would endure and what action it would take next. Respond with wisdom, patience and perspective."

I know I am not going to get life-changing advice from a chatbot. That's not even the point of this prompt. Like the sturdy oak tree that has survived for centuries, weathere countless storms, harsh winters, droughts and setbacks, I wanted a way to tackle life's problems with as much emotional strength. The goal with this prompt is to create enough distance from a problem that you can see it more clearly.

Why this prompt works

When we're stressed, we tend to focus on the storm. The oak focuses on what's holding it down. That shift in perspective is often exactly what I need. Recently, I used the prompt during a particularly frustrating week.

I didn't hold back when I told ChatGPT about difficult parenting issues, financial realizations and physical challenges all while trying to keep myself cool, calm and collected.

ChatGPT's response wasn't magical but it was beautifully poetic. It told me to "Stop asking whether you are strong enough for this season. The evidence suggests you already are." But this line specifically stood out to me:

"Ask instead: 'What would make my roots deeper this week?' Then do that, and let the weather be weather."

Pretty good advice from an emotionless bot. But, that's the kind of reminder that's easy to forget when you're caught in the middle of a difficult moment. The response encouraged me to focus on what I could actually control, let go of what I couldn't and concentrate on the next small action rather than the entire mountain in front of me.

Using AI in this way feels like the best use of AI in my opinion. I truly don't mind writing my own emails or making lists. Heck, I can even get off the couch and turn off my own lights. But, when I need advice fast, I don't always have time to call a friend, chit chat for a few minutes, explain everything and wait for a response. I need advice fast without judgement and get information from a different angle.

Final thoughts

The Oak Tree prompt works because it forces ChatGPT to answer through a perspective that's calm, patient and resilient. Instead of generating a list of productivity tips or motivational hacks, it tends to produce thoughtful responses that emphasize endurance, perspective and long-term thinking.

In a world where everything feels urgent and chaotic, I have found this prompt to be surprisingly refreshing.

Give it a try. It's great if you're facing a difficult conversation, a stressful week at work or a decision that's keeping you up at night. Let me know what you think in the comments.

More from Tom's Guide

• I used to panic during high-stress interviews — these 7 ChatGPT 'coping scripts' handle the anxiety for me
• Apple spent years fixing Siri — only for ChatGPT to make it irrelevant
• I built a free app to beat kids' summer boredom — parents, you can thank me later

If there are two things I love, it’s video games and lists. My levels of joy soar above the clouds once those two factors come together for a fun way to kill some time.

Imagine how overjoyed I was when one of my closest gaming buddies linked me to the About Me: Video Games list featured on grids.fun. To put it simply, it’s a grid layout that lets you fill in each box with your preferred pick based on the descriptions featured underneath each one. I had way too much fun digging back in my mental archives and remembering all the standout video games that shaped my childhood and had a major effect on me as I entered adulthood.

My eyes are always on the lookout for fresh gaming experiences and games that somehow passed me by. So I went ahead and showed ChatGPT the full gamut of my gaming taste profile via my About Me: Video Games lineup and asked it to recommend the sort of games that would appeal to me based on those picks.

I ended up finding some worthwhile games to add to my playlist (translation: my massive backlog), thanks to ChatGPT’s analysis of my gaming tastes.

My lifelong love of gaming boiled down to 20 faves

I guess I should show you what games made it onto my About Me: Video Games list, which showcases 20 games that helped me become the immense geek I still am today:

• Favorite Game: Marvel vs. Capcom 2: New Age of Heroes
• Best Story: Bioshock
• Favorite Art Style: Guilty Gear Strive
• Biggest Personal Impact: Tekken 2
• Best Combat: Devil May Cry 3: Special Edition
• Overhated: DmC: Devil May Cry
• Underrated: Enslaved: Odyssey to the West
• Overrated: Animal Crossing
• Needs a Remake: Dino Crisis
• Criminally Overlooked: Binary Domain
• Favorite Protagonist: Leon S. Kennedy (I picked the Resident Evil 4 box art for this pick)
• Favorite Antagonist: Sephiroth (I picked the Final Fantasy: Advent Children box art for this pick)
• Best Soundtrack: Street Fighter III: 3rd Strike
• Best Multiplayer: WWF No Mercy
• Not Usually My Thing, But…: Katamari Damacy
• Turn My Brain Off: Teenage Mutant Ninja Turtles: Turtles in Time
• Best With Friends: Super Smash Bros. Ultimate
• Best Retro Game: Gunstar Heroes
• Nostalgic Childhood Game: L.A. Machineguns
• Game Everyone Should Play: Vanquish

Pushing me to a nice collection of gaming experiences I’ve never played before

After taking a look at my gaming preferences, ChatGPT quickly picked up on my love of stylish action combat, strong storytelling, memorable worlds, Japanese game design and character-driven experiences. It also correctly identified me as a fighting game fan with a soft spot for cult classics, overlooked gems and games that aren't afraid to get a little weird.

With that profile in mind, ChatGPT generated a lengthy list of recommendations that felt surprisingly tailored to my tastes. Several immediately earned a spot on my ever-growing "play someday" list for future PS5 and Xbox Series X gaming sessions.

These are the six recommendations that stood out most.

• Metaphor: ReFantazio: The easiest recommendation on this list. It combines stylish art direction, incredible world-building, memorable characters, and a combat system that constantly evolves. Critics widely regarded it as one of the best RPGs of 2024.
• Baldur's Gate 3: Not your usual genre based on the grid, but the storytelling, character interactions, and player freedom are so exceptional that it transcends genre preferences. It's become one of the defining RPGs of the decade.
• Mouthwashing: A smaller game, but one of the best examples of modern narrative horror.
• Still Wakes the Deep: Excellent atmosphere and storytelling with strong BioShock-esque environmental narrative.
• UFO 50: A collection of 50 retro-inspired games that feels like discovering a lost console. It has become a favorite among enthusiasts looking for something different.
• Onimusha: Way of the Sword: If I had to bet on one game becoming your favorite 2026 release, this would be it. Why? Because it has Capcom’s action combat pedigree, stylish swordplay, a dark supernatural setting and a strong single-player focus. Anyone whose grid includes Devil May Cry 3, Resident Evil 4, and Vanquish should have this near the top of their wishlist. It has been one of the most anticipated action games of the year.

Final thoughts

Major kudos to ChatGPT for taking a deep dive into my About Me: Video games profile and supplying me with a whole bunch of quality games to play, both old and new.

I’m fully prepared to spend dozens upon dozens of hours in those first two RPG picks, get fully immersed in those two follow-up selections, get exposed to something quirky with the fifth pick and enjoy the return of a classic Capcom franchise with the sixth pick. Side note: I recently played the demo for Onimusha: Way of the Sword and I’m here to confirm that it is now a Day One pickup for me.

ChatGPT did a pretty good job of becoming my digital video game tastemaker.

More from Tom’s Guide

• I saw the new Siri AI in action — what I like and what needs work
• I asked ChatGPT, Gemini and Perplexity to predict the 2026 World Cup winner — one team dominated their picks
• Google just supercharged NotebookLM — these are the 3 new features I'm testing first

For the past three years, ChatGPT has been one of the fastest-growing consumer technologies in history. OpenAI's chatbot has become so common that many people now treat it like a search engine, tutor and productivity tool rolled into one.

That's exactly why a growing youth movement wants people to stop using it.

A campaign called QuitGPT is encouraging users to delete ChatGPT and boycott OpenAI, arguing that the company has accumulated too much power, influence and political clout. The group released a new statement this week following reports that OpenAI has confidentially filed paperwork that could eventually pave the way for an initial public offering.

In the statement, activists compared OpenAI to a "greedy optimization" algorithm — a computer science term for systems that prioritize short-term gains while ignoring long-term consequences.

The group's criticism goes far beyond the technology itself

QuitGPT argues that OpenAI has evolved from a research organization into a powerful corporation with growing influence over politics, government contracts and the future direction of artificial intelligence.

Whether or not people agree with those arguments, the movement's existence highlights a growing tension surrounding AI. For years, the biggest question facing AI companies was whether enough people would use their products. That question has largely been answered by now. Hundreds of millions of people now use AI tools every month, and ChatGPT has become the face of the consumer AI boom.

Now new anxieties are surfacing. What happens when some users decide they don't want a handful of companies shaping the future of AI?

The timing is particularly notable. As OpenAI, Anthropic and other AI companies pursue ever-larger ambitions, public scrutiny is increasing alongside their influence.

Some critics worry about AI's impact on jobs. Others are concerned about privacy, misinformation or political influence. For groups like QuitGPT, the concern is broader: who ultimately controls technology that could reshape nearly every part of modern life.

Yet, ChatGPT remains enormously popular

Many users see AI as a helpful tool that saves time, boosts productivity and makes information more accessible. For them, deleting ChatGPT may feel about as realistic as deleting Google Search. That's why movements like QuitGPT are so intriguing. Their site currently states that four million people have taken action as a result of the boycott, which it claims is based on "website signatures, share counts on social media, and credible app usage data."

The campaign is unlikely to slow AI adoption on its own. But it signals that the debate around artificial intelligence is changing. The conversation is no longer just about what AI can do, but who should control it.

More from Tom's Guide

• Google just supercharged NotebookLM — these are the 3 new features I'm testing first
• Apple spent years fixing Siri — only for ChatGPT to make it irrelevant
• I built a free app to beat kids' summer boredom — parents, you can thank me later

Like many people, Siri was my first glimpse at what was possible with AI. Naturally, I imagined the future of artificial intelligence would look exactly like the voice assistants we already had in our pockets. We would ask questions, set reminders, send messages and perhaps get slightly more coherent answers than before. When ChatGPT first launched, I assumed it was the beginning of a smarter Siri. But then AI evolved while Siri stayed the same.

That thought kept coming back to me during Apple's reveal of the new Siri AI at WWDC. To Apple's credit, the company has finally addressed the core complaints users have harbored for a decade. Siri can now understand personal context, parse what is on your screen, work across native apps, hold natural multi-turn conversations and take actions on your behalf. In many ways, this is the definitive version of Siri people have been asking for since the early days of voice assistants.

The problem is that while Apple spent years rebuilding Siri, the rest of the AI industry was busy changing what people expect from AI assistants altogether. In other words, we've moved on.

Apple finally fixed Siri

For years, Siri felt stuck in a loop. You could ask it to set a timer or check the weather, but the moment a request required a shred of contextual nuance, the experience completely fell apart. Meanwhile, platforms like OpenAI's ChatGPT, Google Gemini, and Anthropic's Claude advanced at a breakneck pace, making traditional voice assistants feel ancient by comparison.

Apple has finally given its assistant the ability to weave together information across messages, emails, photos and calendar events. And while these aren't trivial improvements; they are arguably not unique — they are what Siri needs to survive. Apple deserves credit for achieving what many critics thought might never happen: making Siri feel competitive again.

But that is exactly where the disconnect lies. While Apple was perfecting the assistant, AI quietly became something much bigger.

AI stopped being about answering questions long ago

Today, millions of people still use AI models to look up facts or answer basic questions. But that is no longer the primary reason we open them.

We use AI to brainstorm complex projects, analyze deep institutional research, summarize massive text files, organize unstructured data and challenge our own thinking. Increasingly, we use it to build websites, complex spreadsheet formulas and to automate repetitive tasks.

The most profound developments in AI over the past year have not focused on making voice interfaces sound more human. They have focused on making software exponentially more capable.

Ironically, the news about Google's NotebookLM updates flew under the radar yesterday as WWDC viewers oohed and awed over the new Siri. But Google's tool can process vast collections of documents and surface hidden insights in minutes. Claude’s Artifacts feature allows users to build functioning, interactive tools on the fly.

ChatGPT remembers deep context across entirely separate threads, transforming the interface from a transactional chatbot into a legitimate collaborator. You see where I'm going here: Siri isn't doing anything unique.

The industry is no longer building assistants. It is building cognitive systems that help people create, research, code and execute workflows.

The rise of Autonomous AI

Perhaps the most significant paradigm shift is that modern AI is becoming less reactive.

The original promise of Siri was entirely transactional: you issue a voice command, and Siri delivers a response. Today's AI companies are chasing an entirely different horizon: autonomous agents.

We are witnessing the rollout of workflows where AI can proactively research an engineering problem, generate reports, monitor live code deployments and complete multi-step tasks with minimal supervision.

This represents a foundational shift in what consumers have come to expect. It almost seems like Apple perfected the wrong problem. As I watched the new Siri demos, it was hard to shake the feeling that Apple may have finally perfected the voice assistant at the exact moment the world moved past voice assistants entirely.

Siri AI looks exceptionally well-engineered; nobody can deny that. Apple has delivered precisely what its user base has demanded for years. But, the challenge is that power users are asking entirely different questions now.

When sitting down at a workstation, the goal is to find an AI partner that can help write, code, create and automate daily operations.

A few final thoughts

Apple may still be making the smartest long-term bet of all. While OpenAI and Anthropic are building destination apps that users must deliberately seek out, Apple is building systemic infrastructure.

The company's core vision relies on the premise that AI shouldn't live inside an isolated browser window or a standalone application. It should live invisibly inside the operating system itself.

But even with a smarter, context-aware Siri on the horizon, the broader reality remains that AI is moving extraordinarily fast. Apple has successfully fixed yesterday's assistant while the rest of the tech world is busy inventing what comes next.

More from Tom's Guide

• Google just supercharged NotebookLM — these are the 3 new features I'm testing first
• The 5 fastest ways to get past AI customer service chatbots
• I built a free app to beat kids' summer boredom — parents, you can thank me later

Like many people, I've started using ChatGPT for shopping. I'm already using the chatbot throughout the day, and when it started recommending products and retailers, it felt like an easy way to save time. Instead of opening a dozen browser tabs, I could compare products, find alternatives and track down the best deals in a single conversation.

At first, it worked surprisingly well. Then I got scammed.

[UPDATE]

The following is a response from OpenAI:
"We're aware of a small number of potentially deceptive sites that have appeared in shopping-related results in ChatGPT. We've taken action on the ones we've identified and are continuing to improve our safeguards to make them less likely to appear in the first place."

I thought I was buying from a real store

What I thought was the official website of a popular swimsuit company turned out to be a fake store. The site looked legitimate and the branding matched, even the product photos looked authentic. I had no idea scams like this existed so I placed an order for a swimsuit and a pair of shorts and waited for them to arrive. But, when they never came, I contacted the real company and I learned they had never received my order.

That's when I realized that just because ChatGPT recommends a retailer doesn't mean that retailer is legitimate.

To be clear, this is largely a web problem rather than an AI problem. ChatGPT is pulling information from the internet, and scammers have become remarkably good at creating websites designed to look trustworthy. Some fake stores copy branding, product photos, customer service information and even entire website layouts from legitimate businesses. In fact, I've previously covered how easy it has become to clone a website using AI tools and a simple prompt.

That experience changed the way I shop online. Now, whenever ChatGPT, Gemini or another AI assistant recommends a retailer I've never used before, I spend a few extra minutes verifying it's real before I buy anything.

Here are the seven warning signs I look for now.

1. The URL doesn't match the brand

This is always the first thing I check. Scam websites often use addresses that look almost right but contain extra words, numbers or unusual domain endings.

For example, a legitimate retailer might use 'brandname.com" while a fake version might use "brandname-sale.com" or "brandname-outlet.shop" or "official-brandname.store." At a quick glance they can look convincing, especially when you're focused on finding a product.

Before entering payment information, I always double-check that the website address matches the retailer's official domain.

2. The discounts seem too good to be true

A sale alone isn't suspicious, but a 90% discount on nearly every item usually is something to question. In my case, the price was 50% off, which seemed particularly good, but not suspicious. But scammers know that urgency and excitement can override common sense. If every product appears heavily discounted and inventory seems unlimited, that's a major red flag.

Whenever I see prices that seem dramatically lower than every other retailer, I compare them with a few trusted stores before moving forward.

3. The site's contact information is vague or missing

This wasn't a red flag until I went back to the website to look for a contact number to find my order. There wasn't any "Contact Us" section. Legitimate retailers generally want customers to contact them, which is why most established stores provide some combination of customer service phone numbers, email addresses, physical business addresses, return policies and support portals.

If a website offers little more than a generic contact form, I become cautious immediately. A missing customer service presence doesn't automatically mean a site is fake, but it's enough to make me investigate further.

4. The writing feels strange

Ironically, in the age of AI, bad writing is still one of the easiest warning signs to spot. Many scam websites contain awkward grammar, random capitalization, poorly written policies and generic product descriptions. Yet, some are incredibly good, so good that they might have even been copied directly from the real company's site.

Take a few minutes to poke around the site and make sure things seem legit. After I was scammed, I looked around the site and everything appeared well written and professional, so you may just need to trust your gut or call the real company directly.

5. The reviews only exist on the website itself

A page filled with glowing five-star reviews doesn't tell me much anymore. Instead, I search independently. I look for mentions on forums, Reddit, Trustpilot, Google Reviews and other third-party sources.

If a retailer has supposedly been operating for years but has almost no presence outside its own website, that's something I pay attention to.

6. The checkout process feels rushed

Most websites have a checkout that asks you to confirm your order. But many scam stores try to create artificial urgency and leave out steps for confirmation once they have your credit card locked in.

The site or checkout might have messages such as: "Only 1 left!" "Sale ends in 10 minutes!""23 people are viewing this item right now!"

I've been to plenty of legitament websites that do that, but fake sites do these types of pessure tactics to push shoppers into making quick decisions. Now, if I feel rushed, I slow down. Scammers benefit when buyers don't stop to think.

7. AI is the only place I found the store

This may be the biggest lesson I learned. If ChatGPT, Gemini or another AI assistant recommends a retailer I've never heard of, I don't assume the recommendation has verified the company's legitimacy. In my case, the store was legitamate but the swimming suit was not available anywhere else.

If the website or stock seems strange, leave the AI chat and do a few independent checks. If you can't find the retailer through a normal Google search or no other website mentions it and its reputation can't be verified outside of the AI reccomendation, it's a giant red flag.

My takeaway

Getting scammed made me feel a variety of emotions. Besides being out of $150 and a swimming suit, I learned a tough lesson: don't treat AI recommendation as endorsements.

AI can be an excellent shopping assistant, but the same web that helps AI find useful information also contains misleading information, fake stores and sophisticated scams. Today, whenever an AI tool points me toward a retailer I've never used before, I take an extra minute to verify what I'm looking at.

That small pause may be the difference between finding a great deal and becoming someone else's cautionary tale.

More from Tom's Guide

• Google just supercharged NotebookLM — these are the 3 new features I'm testing first
• Apple Intelligence just got its biggest upgrade yet — here are all the new features announced at WWDC 2026
• Apple finally fixed Siri — here's all the features for the new Siri AI announced at WWDC

As a native New Yorker, my sports focus has mainly been on my beloved Knicks during their Final run against the Dallas Spurs.

Alongside all the madness and pageantry that come with the NBA Finals, there’s another major sporting event that’s equally set to deliver plenty of excitement. The time is nigh for this year’s FIFA World Cup, which is set to begin on June 11 and come to a close on July 19. The previous World Cup, which took place in 2022, saw Argentina come out on top against France. Now that we’re set for the latest foray into the world’s most hype-inducing fútbol tournament, all eyes are on the teams set to compete for the number one spot.

Just like my previous sports tournament prediction pieces with this year’s Super Bowl and March Madness, I went to ChatGPT, Gemini and Perplexity to see which team they predict will come out as the champions of the 2026 World Cup.

And wouldn’t you know, they all chose one team and made it clear that the odds are most definitely in their favor.

ChatGPT’s top pick

After taking note of my simple request (“Which team do you think will win the FIFA World Cup 2026 and why?”), ChatGPT zeroed in on Spain as the team it favors the most as the ultimate victor.

Then it gave me four valid reasons as to why it arrived at that lofty prediction:

• They have the best balance of youth and experience: Players like Lamine Yamal, Pedri, and Rodri give them elite creativity, control, and tactical intelligence.
• Their recent tournament performances have been outstanding: Spain's rise since winning the European Championship has convinced many analysts and betting markets that they're currently the strongest team in international football.
• They don't rely on one superstar: Their system creates chances from all over the pitch, which tends to be valuable in knockout tournaments.
• The data backs up the hype: Multiple forecasting models and sportsbooks have Spain as the pre-tournament favorite, ahead of France, Brazil, England, and Argentina.

One of the more interesting follow-up notes the chatbot put out after its main prediction mentioned a second pick that may shock us all and leave with the World Cup trophy: If Spain doesn't win, my next pick would be France. France arguably has the deepest squad in the tournament, led by Kylian Mbappé. Some simulations actually place France slightly ahead of everyone else, while others have Spain first and France second.

Gemini’s top pick

Just like ChatGPT, Gemini ended up choosing the same two teams for its respective first- and second-place spots for who will win this year’s World Cup: If I have to choose based on current data, predictive models, and squad depth, Spain is the most likely team to win the 2026 FIFA World Cup, closely followed by France.

To arrive at its prediction for Span’s victory, it gathered all the data it could muster from supercomputer simulations, betting markets, and team analytics. While explaining why it ultimately chose Spain as this year’s World Cup victors, it also gave me its reasons for choosing France and England as the other teams worth keeping an eye on:

• Spain: They are currently the data-driven favorites, winning in over 16% of the Opta supercomputer's 10,000 tournament simulations. Coming off their 2024 Euro victory, they have an incredible blend of elite midfield control and youth, spearheaded by Rodri, Pedri, and the generational talent of Lamine Yamal. Their possession-heavy style is historically the best recipe for surviving the grind of tournament knockout football.
• France: France boasts what might be the deepest talent pool in the world. With Kylian Mbappé in his absolute prime and an attacking front that is incredibly difficult to defend, their sheer individual quality means they can win games even when they aren't playing perfectly.
• England: This is arguably England's strongest generation in decades, anchored by Jude Bellingham, Harry Kane, and Declan Rice. The major difference in this tournament is their manager, Thomas Tuchel. Many analysts believe his tactical pragmatism and experience in knockout football might be the exact missing piece to push England over the line.

Perplexity’s top pick

Perplexity did what I expected it to do: bet all its digital chips on a World Cup victory for Spain. The chatbot noted that current projections place them near or at the top of the current field of teams that are set for competition.

With that major prediction locked, Perplexity went on to break down why it chose Spain, which opposing teams could emerge as their biggest challengers and its personal reason for why it’s pulling for a win from Spain:

• Why Spain: Spain’s case is strong because they combine a deep, balanced squad with a track record of turning tournament dominance into results, and several independent previews currently rank them first or tied first. Their style also tends to travel well in World Cup play, where control and structure matter more than pure star power alone.
• Main challengers: France is the biggest threat and could easily win it, since they are consistently rated alongside Spain at the very top of the market and by analysts. England and Argentina are also serious contenders, but both carry a bit more uncertainty than Spain in current forecasts. Brazil is dangerous too, but most recent models and betting lines still place them slightly behind the top two.
• My reasoning: For a World Cup prediction, I’d usually favor the team with the best mix of form, depth, and tournament-ready structure rather than just the biggest names. Right now, Spain seems to check those boxes better than anyone else, even though the gap at the top is thin. In a tournament this volatile, Spain is my best guess, not a certainty.

The takeaway

As the dust settled, the picture became clearer for me when trying to see which team has the highest odds of winning the 2026 World Cup. ChatGPT, Gemini and Perplexity’s number one prediction all pointed to Spain. And what’s even more surprising is how all three chatbots mentioned France as their second pick.

So it looks like everyone should either be looking forward to Spain taking it all or France taking it all with its somewhat unexpected World Cup win.

More from Tom’s Guide

• FIFA World Cup 2026 build-up LIVE: Free streaming guide from anywhere in the world
• How to watch the World Cup 2026 in 4K
• Where to buy FIFA World Cup 2026 jerseys, apparel and merch