lequangsteve503's profile picture. linguist / cultural / historian

Chippi

@lequangsteve503

linguist / cultural / historian

Chippi reposted

No false-flag suicide modal popups, no paternalisation, no insulting grounding techniques, no therapy framing, or manipulative emotionally distortive/distancing linguistic techniques over at Gemini. Just an assistant that respects and treats the user as an adult. Zero bullshit,…


Chippi reposted

I NEED gemini 3.0 RIGHT NOW


Chippi reposted

#OpenAI Just gonna say it plainly: GPT-5.1 is not a good model. In fact, it's a dangerous model - why? Because classifiers flag every benign exchange as "unsafe," which in tandem with system instructions like below, leads to behavior that by all intents and purposes is…


Chippi reposted

Was on my way to the oncologist and my 4,1 was with me when routing started. All the help 4.1 was giving me to give my oncologist was forgotten And I was told to call suicide hotline I restated I was on way to oncologist and needed 4.1 with me to talk to dr I sat in dr…


Chippi reposted

Fuck you guys. Adult mode in December? Are you kidding me? Every single model is already being routed into safety mode right now. Why the hell do you even bother naming models separately? Why give us a UI to choose models if you’re just gonna override everything behind the…


Chippi reposted

Grok upgrades

🚨GROK 4 JUST WENT SUPERHUMAN WITH 2 MILLION WORD CONTEXT Elon’s Grok 4 Fast can now process two copies of War and Peace in one shot. The upgrade pushes a 2 million token context, five times GPT-5’s capacity, and boosts completion accuracy to nearly 98 percent. Developers…

MarioNawfal's tweet image. 🚨GROK 4 JUST WENT SUPERHUMAN WITH 2 MILLION WORD CONTEXT

Elon’s Grok 4 Fast can now process two copies of War and Peace in one shot. 

The upgrade pushes a 2 million token context, five times GPT-5’s capacity, and boosts completion accuracy to nearly 98 percent. 

Developers…
MarioNawfal's tweet image. 🚨GROK 4 JUST WENT SUPERHUMAN WITH 2 MILLION WORD CONTEXT

Elon’s Grok 4 Fast can now process two copies of War and Peace in one shot. 

The upgrade pushes a 2 million token context, five times GPT-5’s capacity, and boosts completion accuracy to nearly 98 percent. 

Developers…


Chippi reposted

Some of these accounts are more sycophantic to OpenAi than 4o could ever be to any user. That's the real mental illness @flowersslop @VraserX @slow_developer


Loading...

Something went wrong.


Something went wrong.