ngl code-llama has been better

even claude-2 sometimes

gpt-4 is not as good as it used to be

janitor for /zk and some more stuff here linktr.ee/randomishwalk

just annoyed gh copilot kinda enshrines gpt 🤢

Hi! My name is Daniel. I am building bountycaster.xyz

Ex Phantom, 0x Project engineer.

From Italy originally, currently live in NYC

hm. aside from self deployment, poe[.]com is a nice easy button for using llama and comparing inference quality across a bunch of models

there’s probably heavier duty stuff out there but this works for me

Mobile and web developer, building re-public.io, journodao.xyz, demo.factland.org. Ex-WIRED. Writer, dadder, musicer. @developer_dao.

RLHF (leading hypothesis i think) and other stuff

Just noticed this too. Weird that it’s getting worse. What would make that happen? I thought it was ChatGPT specifically and they were making adjustments behind the scenes that were degrading it

when you say it is not as good as it used to be, what do you mean specifically? for example the code doesn't compile now, the logic behind the code is wrong, or something else?