Matisyahu's Blog

One of the biggest myths perpetuated by the neoliberal orthodoxy is the idea that markets are rational because those participating in the market are motivated by rational self interest. If that orthodoxy was real rather than being an article of faith akin to some sort of religious conviction then there wouldn’t be crashes on a near routine basis of almost every 7-10 years. Well, here is an example regarding Apple’s share price in regards to AI (link), there was a survey not too long ago where they looked into the buying habits of iPhone purchasers and what they prioritise when purchasing a phone and want to know what was at the bottom of the list of features? at 7% ‘thinness’ of the phone and AI functionality were dead last or close to last on the list of priorities.

If there is one thing the AI bubble proves it is the idea that a small group of people have far too much money and they’ve convinced themselves, after huffing on their own farts for decades, that they can do no wrong and everything they touch will become a winner. Well, I think that when the correct comes I think there are going to be many left wondering why they were so foolish but I do hope is that governments of the world put their foot down and say ‘play stupid games, win stupid prizes’. I think it is abundantly clear that LLMs are a dead end and this idea of “we just need to make them bigger and that will make them better” has been well and truly debunked by people who are a lot smarter than I am. Are there great things occurring in the field of AI? sure, and that will continue, what will crash and burn will be the fixation on LLMs and the realisation that there is a law of diminishing returns because as the models get bigger the improvements become less and less to the point that they’re just giant money and power sucking operations that aren’t going to get better.