As a counterexample, I've been seeing more "safety rejections" from Claude. Unlike search, being unable to ask _anything_ about botulinum, or details about the recent Copy Fail vulnerability (without giving my fingerprints to Anthropic to become a "verified security researcher") we're only just beginning to see the ways LLM can be used to distort information and its availability.
replyMy grandfather was one of the first people in Canada to own a commercially available chainsaw.
replyLet me tell you - it didn’t take 30 years for people to figure out that chainsaws were useful.
Yet most of the kids on HN think they write better code than ai and that it’s completely useless and has no place.
replyThat's fine if we aren't destroying existing products to replace them with AI.
replyPeople can already use AI mode in google search if they want. "It'll be better later" is a shit reason to kill one product for it.
No. It is already good enough. But complaining that wording is so generic is hilarious. It just shows how little people on HN understand ai at all.
replySo you started with ‘highly doubtful’ as a comment, got given lots of examples and instead of assimilating that info you closed your eyes put your fingers in yours ears and said “oh who cares?’ - you’re on team AI regardless eh? That’s fucking weird mate.
replyNa. Wasn’t given any good examples. People just whining about the same stuff because “oh no I got information that’s former in the same structure that I can tell it’s AI and it makes me feel bad”
reply