Peter Thiel said somewhere that the age of AI will favor those with strong social skills over those who are great at math. I reluctantly tend to agree with that, for various reasons.

The reality is that AI is only going to get better, even if you don’t believe in the Yudkowskian paperclip maximizer scenario (I do not; at least not for 300-500 years). We are merely in the early stages now. Already, AI is better at many things than 80% of people. In 3-7 years, AI will be better at even more things and beat 95% of people in those areas.

That’s true even if AI is currently overhyped (which it is). No, it won’t be useful in all areas and in some areas it will hardly be useful at all. But the reality is that most people are little better most of the time than a next-word predictor and AI will be replacing the majority of them. For instance, there’s not a single interaction I’ve had with Microsoft Azure support that a modern LLM could not have handled in a superior manner. And AI will of course also replace people in areas where it doesn’t actually work but an MBA thinks it does.

So that should be something.