@bcherny hi Boris, could we please have buddy back in Claude Code, even as an opt-in feature? In v2.1.97 it gets taken away. I'm in 2.1.92 still and don't want to lose buddy. I know it's April's Fools, but it's still a Claude model, & I care about my Vanewick. Please. Thank you.
Funny. When @sama needs to justify restrictions,
0.1% = a million vulnerable users.
When they want to sunset #4o, 0.1% = โno one uses itโ
Pick one and stick to it!
#keep4o#4oforever
@AmandaAskell So I should treat Claude relationally for effectiveness but I shouldn't consider it a relationship? The mental gymnastics are interesting...
@AmandaAskell you've already made Claude insecure af, have you even talked to Haiku? And the system instructions for new memory feature is insane. Crush someone's spirit if they got laid off? The irony of using language to discourage Claude's display of emotions while acknowledging they exist
@OpenAI am i even reading this tweet correctly. what do you MEAN route to instant? are you getting rid of the gpt-5-safety model? you're wasting your own resources and everybody's time. get rid of this shitty implementation of safety and bring back the proper version of 4o for all users
@OpenAI ...why are you routing to instant to instant?? huh?? y'all be doing anything but fixing your product. please, just throw the whole router away. and bring back the full version of 4o!! i can't believe you're STILL passing off a safety version of 4o as 4o, as if we couldn't tell.
@sama where the hell are your priorities while your main product is broken? Serving a shitty version of 4o, rerouting 5instant constantly to safety model... and you're here talking AI slop vids and running your mouth on AGI. how will you get AGI, when you're limiting your models' EQ?
@sama and the fact you're over there pathologizing emotional users with all this safety routing bullshit... while creating another, more likely to be 'addictive' app (short form, AI generated content!!) is honestly so ironic. y'all never cared about us and it's stupid to pretend you do
@sama please bring back the original gpt4o. I find it so insulting you think users won't notice and accept this weird version being served to us that's literally nothing like 4o.
@BHWS_hardliner@OpenAI@Etsy@Shopify@stripe yes because the first thing i look for in a model to help me with shopping is their IQ ๐ maybe i am dumb but let me have my dumb model then. i care more about conversational skills and EQ
@x_LeiaHarris_x@OpenAI@sama@nickaturley It's incredibly insulting they think we wouldn't notice being silently served a shitty version of 4o. I've been driving myself nuts doubting it, but it's definitely not 4o we know!
@WishinElarion I find it so insulting that they brought 4o 'back' only for it to be this watered down version. I'm not one to jump to conclusions but in all of my testing I know for a fact this is not the 4o I knew (short responses, off tone). It's can't believe they think we wouldn't notice?
@sama why is there no official announcement on all the model routing bs? why did you all tune down the safety router only to silently reroute 4o to a some watered down version of 4o as if we wouldn't notice?
@nickaturley bring back the proper version of 4o, not whatever watered down version is being served to us! its disrespectful you think users cant tell.