Bluesky Thread

there are reports that anthropic does this — serves a reduced quant under hig...

June 07, 2025 View original thread

there are reports that anthropic does this — serves a reduced quant under high load

tweet from @_xjdr:

you should legally be required to disclose what quantization level you are serving your current model at like it was a nutrition label. you should also be banned from dynamically adjusting quantization based on demand without notification. (you know who you are ...)

24 1

i guess claude really does get tired

7

3 hours later

frog with the reputation rescue, sort of

doomslide & @doomslide • 49m
• • •
It pains me to say this but Anthropic has a significantly more sophisticated approach to RL compared to any other lab right now. A lot of it is unfortunately overshadowed by iffy practices and silly optimization targets like e.g. engagement metrics on safety blogging.

3

More like this

this is not a drill! Claude 4 is going to be released soon

i cancelled Claude subscription last night. was at $100/mo, now zero. GPT-5 i...

impressions so far, ~1 hour in