Think of it as a pipe with intelligence coming out of it. And you can convert that intelligence into all sorts of things. You can turn it into actions, into summaries, into translations, into automations of the stuff you’d normally do by hand. It’s raw intelligence, and it’s sold by consumption, not by subscription.
What do you actually do with it?
First question: what do you want to do with this intelligence? Go looking for the things you do most often, the ones that frustrate you, the ones where it would be nice to have a button that just does the whole thing for you.
And this button also comes with language skills. So you don’t have to do what programmers do, spelling out every last detail to the computer, because the computer now understands language. Through this intelligence that arrives on tap, you can make it do the most varied things, programming it the way you’d talk to a colleague or a friend: “in these cases do this and this and that, in this way, and when you’re done maybe ask me to confirm, or ask me these questions.” That’s the scenario in front of us.
The ground is shaking under software
I think the world of programming is in turmoil, because this stuff hit us first, but it’s coming for everyone, folks. We’re in a phase where the whole thing is being shaken at its foundations. It takes ten minutes to do what used to take you a month.
The famous software-as-a-service model is collapsing. Those subscriptions at 10, 20, 30 euros a month for software that serves some specific purpose. It’s collapsing for two reasons.
One: it stops making sense to pay for narrow, specific software, when you can build it from scratch without knowing how to program and cover most of your needs yourself.
Two: the subscription logic doesn’t hold up with artificial intelligence. And why not? Because the marginal cost of more intelligence is an energy cost, not the old-style software-and-hardware cost, where you could go from 10 to 100 to 10,000 users with ever-lower costs to serve each one. That ever-decreasing marginal cost was the key to scale. Now, every time you prompt an AI, every time you send it a message, somewhere out there you burn the energy equivalent of three or four microwave ovens. That does not scale the way technology has scaled until now.
Confusion is opportunity
So we’re facing a scenario that is, on one hand, unpredictable, and on the other, extremely interesting. Precisely when everything is this fragmented, this much a work in progress, that’s when you get the wildest ideas, but also the best ones, the ones that bring real satisfaction. Where there’s confusion and fragmentation, there are also lots of possibilities. There’s also the need to reinvent yourself a little, to sit back down and study, to understand what’s happening.
