As a real dummy on the subject, maybe you could help me understand where vector search tends to fall over?
I use it to retrieve tool functions by description and it has worked very well for me, but I expect I'm in the "very simple use cases" category that you mentioned.
I assume a bunch of them print pretty reliably, like Call of Duty, when you're not using them as a loss leader on an expensive subscription that nobody wants.
> If the best we've got for convincing people to learn to code is that it's like math notation (the most hated part of math [...]
That's funny. I've told a mathy friend that I've sometimes wondered if I could have grown up without the whole, "... except I suck at math", and I think that's why.
I don't struggle with the problem solving. I've watched people reinvent chunks of "difficult" math in code without realizing or caring that they've done it.
I've started to think that math might actually be awful on purpose.
My spidey sense goes off when we refer to open weight models from Chinese companies trained against frontier models as OSS... as if it was some Torvalds types maintaining a public git repo.
I think those open weight models exist because openai, anthropic, and google rule the roost. And I don't think things would continue as-is if those guys implode.
Years of crime and corruption caused long perception and union density declines from the 50s to the late 80s, when hundreds of officials finally got RICO’d, some unions fell under government supervision, etc.
They still didn’t really give up on crime and corruption, but they never really recovered in the US.
It's so distinct... it can't be an amalgam of the most popular design choices, can it?