AI, Wine, and the Bellend Principle of Human Creativity

My friend Nick asked me to explain AI to him. This is what I told him

In two weeks time I’m speaking at a conference on AI, Wine, and The Law at The University of Arizona. I was chatting about it with a friend

Nick Parker

of the utterly, utterly brilliant

Tone Knob

Substack who said of AI “Could you explain it to me?!”

This is what I wrote back.

Deep down you already know the truth. The most important thing about learning about AI is that there are lots of people who rely on you not knowing about AI. The boosters don't want to admit they've put their cash (in the billions) and social capital into something that they've overpromised. At least for now. And the skeptics don't want to admit they're skeptics because it means they'll be disrupted when change does happen.

Trying to "understand" it is hard because it's based on very complex maths (although this shows you can understand some of it by analogy). But also because it's a case of shifting sands. In the last few weeks the cost of building an equivalently powerful ChatGPT model has gone from $100M, to $5M, to $50. Literally fifty dollars. I spoke to a guy yesterday whose friend had invested $300M in a