@screwlisp LOL, I hope my “responsible use of AI” comments didn’t offend the regular audience members too badly! I wasn’t able to see the chat at the time.

Like @someodd said, and I agree, you can use LLMs for type checking, and possible also for producing heuristics that guide a sat-solver. For example, LLMs are already being used to find zero-day vulnerabilities by both attackers and defenders. As another example, using an LLM in a place similar to where you would use refinement types with a sat-solver can work as long as you have a classical type checking algorithm to guide your explorations of the design space.

LLMs are an algorithm like any other: they can be overused and abused (e.g. slop coding), they can be used by bad people to harm good people, but these algorithms do have legitimate uses. And with more research, like what we see happening in China right now, the energy cost of these algorithms may drop quite a bit in the near future, so at least the problem of data centers destroying the planet and using up all the fresh water could theoretically be resolved.

I hope soon the hype around LLMs will die away, and people will stop seeing these things as proto-gods and start seeing them as just another tool you can use to solve certain specific software engineering problems.

@dougmerritt @havoc @mdhughes

@ramin_hal9001 @mdhughes @someodd @screwlisp @dougmerritt It seems that the current "AI" vibe is to build datacenters to process surveillance data on citizenry, backed by government bailouts for the ultra wealthy, to essentially establish a permanently rich caste who can surveil the citizenry at will.

It isn't "just a tool" it's a technofedalist wet dream.

@ramin_hal9001 @mdhughes @someodd @screwlisp @dougmerritt Look at the amount being spent on datacenters. The only way this investment makes any sense at all is if you can use it as a means of social control AND you don't expect to take losses.

Our debt is 39T right now. "If the bubble pops, think of all those retirement funds that will fall!" - the only way you'd make these bets is if you think someone else will pick up the tab. WE will pick up the tab. We'll pay for our own nightmare.

@havoc I think you are conflating the algorithm with how it is being abused by techno-fascists. I am not denying that the abuse is real, and is a horrible problem. But the fact of the abuse is a different thing from fact of the algorithm itself.

@mdhughes @someodd @screwlisp @dougmerritt

@ramin_hal9001 @mdhughes @someodd @screwlisp @dougmerritt The algorithms themselves are just linear algebra. I don't have a hate hard-on for linear algebra.

It's pretty clear what the social consequences of what has been enabled are going to be.

@havoc you are absolutely right, and I fully agree.

I also think it is true that I can use an LLM to find bugs in my code, I have tried it and it works.

@mdhughes @someodd @screwlisp @dougmerritt

@someodd @ramin_hal9001 @mdhughes @screwlisp @dougmerritt Ironically the Chinese might save us, weirdly enough.

By releasing models that run on commodity hardware, they destroy the business case for the "infrastructure moat" that the techno-feudalists were relying on. If the AI bailout ask is "too much, too fast" it likely won't succeed politically, and thus will backfire, leaving the ultra-wealthy far worse off.