Assume there's a new technology and you decided not to act upon it since you were betting on it not being quote-unquote "the next thing". Via the law of excluded middle there's two categories of outcome:

• You were wrong, in which case you lost the chance of being early and all the perks tha came with it.
• You were right, in which case:
  • People who bet the other way did not technically lose everything; even if they didn't manage to gain any new connections out of their ventures, at the very least the failure could be converted into knowledge and experience through thorough introspection.
  • You gained nothing by not doing anything. Just because your conclusion is correct - or more accurately, exactly because your conclusion is correct, - you have no way to know even if your thinking process is wrong. Also, even if you can claim that you dodged the failure because you're smart enough to see the failure coming, it's hard to convert this "expressed smartness" into tangible gains (e.g. reputation) unless you're already a known "smart enough" guy.

From this perspective, the best strategy seems to be always following the latest trend in tech. I seriously wish the world weren't like this because it's fucking tiring, but what bargain chips do I really have against God on this front after the Fall of Men?… Some people would say the failure of NFT cannot be explained this way, but NFT was immediately and obviously stupid, and this time with LLM it's not.

Some people, I believe, fails to predict the effect the advent of LLM would have not because they're wrong on the entirety of the issue but because they're correct on the wrong part of the issue. I believe they have made two judgements, both of them I also believe to be technically correct:

• LLM in its earlier stages (especially during pre-agent days) wasn't necessarily immediately useful and subject to more errors.
• LLMs will not lead us to AGI even if at its most advanced form.

But LLMs don't need to be able to become AGI to have a massive effect on a lot of things - there exists a middle spot where conclusions based (and only based) on the two judgements above fall apart, and because the two judgements on the extreme cases are so obviously correct, people tend to not think about the possibility of having such a middle spot. The fact that luddites have had complete victory when NFT turned out to be a truly entirely stupid thing and the fact that most hype around LLM AI are done by the people who also did (and still do) blockchain and cryptocurrencies have probably strengthened people's belief on their own judgements on LLMs as well.

---

Back

Last update: 2026.3.28