new | show | ask | jobs Github

p-e-w 2 hours ago

Noone can seem to remember anything with certainty when asked to actually substantiate these claims.

▲

computerex 2 hours ago | parent | next [-]

There was this: https://www.reddit.com/r/Anthropic/comments/1sl5wfh/the_degr...

▲

prodigycorp 2 hours ago | parent | prev | next [-]

I’m typing from my phone and im not going to review the semantics of Anthropic’s storied history of performance issues.

It’s not just ant. There are so many small knobs that providers can claim isn’t nerfing but “load management” or “improving user experience”. One example from OpenAI is reducing juice to reduce time to first token.

	▲	winwang an hour ago \| parent [-]
		You can just have your agent find the evidence, review it, copypaste it.

▲

erinnh 2 hours ago | parent | prev [-]

I mean there is a direct link two comments down from here from 30 minutes before your comment: https://news.ycombinator.com/item?id=49902477

	▲	p-e-w an hour ago \| parent [-]
		That’s NOT Anthropic admitting to “nerfing” their model as claimed above (which implies intent), that’s a regression which they quickly fixed. Christ this forum has become intellectually dishonest.