2. O’Reilly on O’Reilly ETech

O’Reilly on O’Reilly ETech

The people who developed agriculture were hackers, and so were Archimedes, the ancient Romans who worked out how to build with concrete; and the Wright brothers who started out as bicycle mechanics, Tim O’Reilly himself said.

“ETech is about these edges where we’re making up things as we go along,” O’Reilly said.

O’Reilly also doesn’t think great discoveries are usually made by people who set out to achieve a breakthrough. “Often the hackers and the people on the edge of the next revolution don’t know it’s the next big thing they’re working on - they just know that they love it,” O’Reilly said. “When you follow hackers you just know something interesting is happening.”

O’Reilly is trying out personal genomics himself, getting a personal DNA analysis through a site called 23 and me, where he’s meeting people who share genetic characteristics. He pointed out that Google is using “the basic scientific methodology of gathering huge amounts of data” so we can expect interesting discoveries as they start to mine the data. But he thinks the next big thing will come from the combination of social networking, machine learning and hardware. “Web 2.0 apps are bionic apps made of people connecting the computers to get results they couldn’t do alone,” O’Reilly said. “What comes next is when these collective intelligence apps are driven not by people typing on keyboards but by collecting information from sensor networks. When ambient computing becomes pervasive we can watch how sensor networks change the applications we use.”

The example of so-called collective intelligence apps you’ll see first, O’Reilly said, is the Dash navigation system: inside it’s an OpenMoko mobile phone and it uses swarming behavior to spot where other Dash users are stuck in traffic. A company in the UK is using the details of cell phones connecting to base stations to measure foot traffic in shopping malls. And in Asia, some ATMs have cameras that use facial recognition to tell if you’re old enough to have a bank account before they pay out your money.

Elephants Never Forget

Programming systems that process a wide range of real-world information might be easier in Elephant 2000, a programming language being written by John McCarthy, the man who first came up with the term Artificial Intelligence. He’s already responsible for LISP, a language designed for AI but also used for Web systems like Yahoo’s first advertising platform. It’s called Elephant because “elephants never forget.” Programs will remember what has happened, so if the software knows you bought a ticket for a plane flight, then it will carry on believing that you have a reservation unless you cancel it. Thermostats would believe that a room was too hot or too cold; it’s a more natural way of thinking about systems than reducing everything to procedures and variables.

Elephant 2000 is designed for programs that will interact with real people – what McCarthy calls “the borderline of AI.” He refers to programs as if they were a conversation, with questions, answers, offers, requests, permissions – and promises that create obligations. “Computer programs will and do increasingly operate in a social context or a business context where they buy and sell, and therefore make commitments on behalf of their owners or the people who have paid to use them,” McCarthy said. “So they should make only the promises they are authorized to make and they should keep the promises.” However Elephant is a work in progress rather than a language developers can use today (and McCarthy didn’t have time to go into all of the details about how the language would work).