Hey, I’m new here. I’m a developer looking to connect with other people who actually ship things - web, backend, mobile, AI, infra, whatever you’re building. If you’re learning, building in public, job hunting, or just want to talk shop, say hi.
Drop what you’re working on and I’ll follow back.#BuildInPublic #DevCommunity #100DaysOfCode #SoftwareEngineering #TechTwitter #letsconnect
@thsottiaux poor tibo will always get asked for a reset even after he leaves OpenAI xD if he ever leaves. Also I would say grifters will always make the poll say reset even if the daily shipping was amazing :D
@bridgemindai@bridgebench@grok analyze this methodology and tell me if it makes sense to test nerfing of models like this or is this a bunch of vibe coded crap.
@eptwts I would say it has to have something with the training data used and the humans that validated it initally. I would say that the humans that pressed thumbs up on the initial claude outputs (much before we got access to it) had much more "taste" than openAI humans.
FFS. Just try using 6.1 as a daily driver for a day and check the usage at the end of the day... I've been on a 100$ sub for a couple of months and I heavily relied on the resets from @thsottiaux I've got a feeling this ones different!