[andy99]

All current "guardrails" are silly censorship / political correctness stuff, or for business appropriateness. They are also trivially circumvented. There is no "threat" from the un-shielded capability of current or foreseeable ML models.

[whimsicalism]

none of yall saying this “foresaw” anything even in 2020 when it was obvious.

i’ve been listening to skeptics be wrong about future capability predictions for 4 years now and the confidence doesn’t seem to be waning at all. i have no clue what the future brings but your confidence is misplaced

[astrange]

I don't think this was obvious in 2020, it wasn't a popular research direction until InstructGPT came out in 2022.

I have continued to have the same opinion of it not being a problem since then. Especially since the AI doomer theories are based on 90s ideas of GOFAI that isn't even how GPTs work.

LLMs are a pretty neat impossible thing, but we'd need a few more uncorrelated impossible things for it to be "dangerous".