I distinctly remember, some time in the nineties, when I was still in school, when the staff of the school’s library was tasked with introducing us to a new addition to the library’s facilities. Along with the books, and the media library of VCR tapes, and the microfisch archives, there was now a few computers hooked up to this new thing – the internet.
Some of us already knew all there was to know about it, but the school wanted everyone to be on the same page. So the staff had to scramble and find out what this new thing was.
Similies abounded: “A website is like a leaflet.” “A newsgroup is like a newspaper” (hint: it wasn’t). “Yahoo is like a phone directory” (again, it wasn’t). And if you had the bandwidth, you could maybe even listen to radio over the Internet.
Story time, kids: At that time before Google, popular search engines were Lycos, Yahoo and Altavista, and in addition to the search box, they would feature a catalog, basically categories and subcategories, where the major websites were listed. That is how humanity used to organize information, and for a while they continued to do so, until it became obvious that this was not going to work.
And then there were other use cases for the Internet that didn’t neatly fit into existing patterns. What is IRC like? What is a forum? Those concepts were new, and didn’t have a clean analogy to the analog world.
And there was a widespread hesitation about using search engines. What can you find, if you search for it? And what keywords are you supposed to type? At first, not much! Then it occurred to you: you could find music lyrics online. Or cooking recipes. Or information about science and technology. The bubble of what your mental model of the Internet was, expanded just as fast as the Internet itself expanded. We know today that it expanded to envelop almost all social interaction, all shopping, all dating, all media and all entertainment, an idea that at the time was laughable. And it did so because it was based on a new building block: low-cost long-distance communication. This was something that had never existed before.
There is a lesson to be learned here about technological adoption. AI is at least as transformative a technology as the early web was. It’s up there in importance along electrification, automobiles, agriculture, since its use cases are not limited to this or that area, but are ubiquitous. We now have the new building block which are the models, and we are now seeing the first agents, the first integrations.
And the lesson is this: A car is not a faster horse. An online chat is not like a written telephone call. Electricity is not better fire. Telepresence is not like travelling. Yes, new technologies help you do the same things faster and easier, but they also open up new avenues, make new goals reachable.
We already saw this with chat-bots: In the beginning, people were just using them to build things faster; to continue the status quo with increased productivity. But today we are finally seeing some engineers realize that they can set new, more ambitious goals.
We are lucky enough to live through one of the big inflection points of history. So let’s not shy away from setting big goals. The people who will benefit the most from this shift, are not the ones who will continue to build health trackers and productivity organizer apps, but faster. They will be the ones who realize that they can now build what before was unthinkable.
As a software engineer, I am trying to train my mind at this new reality. I am building a game, but I am building it in a way that would not have been possible before AI. While in a certain sense it’s nothing really special, at the same time, it’s not like anything that I have ever seen before. I don’t want to say more right now, except that it’s close to being released and it’s looking good. More info soon.