Club Benefits
If you’ve ever felt like AI is getting a little too convenient, you’re not alone — and now there’s a tool designed to push back.
It’s called Slow LLM, and instead of making ChatGPT or Claude better, faster or smarter, it does the exact opposite. It deliberately slows them down. And, let me tell you, it's not just by seconds. I gave it a try and it was enough to make the experience feel extremely frustrating.
Why would anyone do this to themselves, you may be wondering? Well, people are building tools like this to add friction back into AI, forcing users to slow down and think instead of outsourcing everything to a chatbot. It’s also a form of quiet protest against how quickly AI is becoming embedded in everyday life, raising concerns about overreliance and lost critical thinking skills.
What Slow LLM actually does
Created by artist and programmer Sam Lavigne, Slow LLM works as a browser extension that interferes with how AI responses load. Instead of breaking the chatbot, it introduces lag — stretching out responses so they arrive noticeably slower than usual.
Article continues belowThe result is that users can still use the chatbot, but it almost becomes annoying to do so. When I tried it, I ended up reverting back to Google for my search. In some cases, the tool can even be deployed across an entire network using DNS settings, meaning multiple users could experience slower AI without realizing why.
The real reason behind a tool like this comes down to resistance. Not quite the QuitGPT movement. the idea is if using AI becomes less convenient, people may rely on it less.
That taps into a growing concern around AI use — not just what it can do, but what it might be quietly replacing such as critical thinking, problem-solving, creativity and even everyday decision-making.
Are we getting too comfortable letting AI do the thinking for us?
As I was using this tool, it did make me question whether we are getting too comfortable with using AI for...everything. While Slow LLM is positioned as an artistic and philosophical experiment, there’s a more practical side to consider.
Since it can be deployed at the network level, it could theoretically be used in shared environments — like schools or offices — without anyone knowing it’s there.
There’s also a security concern. Routing traffic through third-party DNS tools always carries some risk, and even the creator has warned users to be cautious or self-host if they experiment with it.
But this taps into a bigger message about how AI is being embedded into everyday life as the tools get smarter and faster. This web tool flips that narrative completely.
The takeaway
I'm an AI power user and have already removed the tool, but it was certainly interesting to try. I don't see Slow LLM going mainstream — most people don’t want to make AI harder to use when the entire industry is racing to make it faster, smoother and more embedded in daily life.
Between the QuitGPT movement and now this, for the first time, we’re not just seeing people build better, more powerful AI, we’re seeing a quiet shift in the opposite direction, where some users are starting to question how much they rely on it in the first place.
Instead of optimizing for speed and convenience, tools like this introduce bottlenecks on purpose — forcing a pause, a second thought, a moment of “do I actually need this?” This trend isn't going anywhere and will continue to be an interesting one to watch.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom's Guide
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
