Where I’m coming from

I’m a designer who works with code, though I cannot call myself a programmer. I’m not a tech bro, at least I try not to be. I’m not an AI evangelist. In fact, I began very firmly in the AI skeptic camp. But after playing with and trying to understand what it is we’re dealing with here, I’ve come to see that there is something there — it may not be as fully baked as Sam Altman et al would like us to believe, but I have to suspect that we are only seeing the consumer grade version of something far more astonishing. So, yes, I think there is a there there, but I’m starting from the assumption that there is a level of hype that needs to be cut through first.

I think the best thing I can do here is to try to describe what I’m seeing in terms most people who aren’t following this stuff as closely can understand. That is, from the perspective of an English major (see Ma, it turned out to be somewhat useful!). In my view, if you’re not building your own GPT from scratch, you’re just as much of an imposter as I am. And you shouldn’t have to do matrix multiplication in order to use these things, just like I barely understand how my car works but I can appreciate the way a BMW handles or if I’m driving a lemon.

I’ve taught classes to college kids about how computers work and how the Internet works, so I am approaching how “AI” works in the same way. Most people can’t be bothered to use the free AI tools out there and they’re not even seeing what it can do, or they’re using ChatGPT only and missing the wide range of other frontier and open-source models that are proliferating, or they’re conscientious objectors who are opposed to it for ethical and environmental reasons (all of which, in my opinion, are completely valid). The tech barriers are there, just as they were during the Web 1.0 years.