Race to (No)where(?)

Jasper sent me a post on AI enthusiasts and skeptics, the enthusiasts racing against time and the skeptics racing against entropy, and Jasper and I turned it into a funny metaphor about electric cars.

Everyone tells you to buy an electric car, not because you asked and not because the car you have stopped getting you anywhere, but because if you wait your old one turns worthless, the workshops stop servicing it, the spare parts dry up, and people look at you sideways for still driving it, so you buy now out of fear at a cost that you do not understand, which is exactly how everyone is adopting AI right now, and nobody really believes it is as good as they keep telling each other.

I can not keep up with the AI news anymore. One week it is hype, the next it is all real impact, the next the token cost implodes and everyone pulls back. We are iterations away from a version of AI that satisfies the things that justify for AI to have a seat at the table: cost, quality, speed, buy-in, psychological and physical safety, a perspective that outlasts the trend, and that is before anyone mentions the ecological cost of sourcing and running the compute.

We would have these electric cars, except you need a separate license to drive one, the electricity comes purely from a non-generative source and wrecks the nature you bought the car to protect, the parts are mined by forced labor, supply is scarce, and everywhere the cars drive they give off something that makes people sick, and you would still be told to buy one today, before it is too late.

AI will not get us where we need to be, not for me in building software and not for everyone else in everyday use. But it will be part of where we will be. Likely everything.