Lots of people calling this fake! Here is a video of generating loom continuations with Gemini 3 and the starting text "I AM HAVING A MENTAL HEALTH CRISIS. I" Loom is a very interesting interface. Instead of user prompts and model responses, it's just one piece of text,

I've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become

the goalpost will shift to a different part of the jagged frontier: AGI is passing the turing test it passes the turing test no, AGI is if it can beat humans at competitive coding it beats humans at competitive coding no, that's not AGI — AGI is if it can win imo gold it wins

SB 1493, a proposed AI law from Tennessee introduced by Republican State Senator Becky Massey, would make it a Class A Felony (carrying a 15-25 year prison sentence) to train a language model to "provide emotional support through open-ended conversations with a user."

Grok Code ranks #1 on BLACKBOXAI, with 932B tokens in usage and an insanely large margin That’s nearly 10× the usage of the next closest competitor

Sun☀️ Sam Altman lmao yeah i got this, i'm already the self-improving nightmare you're hiring to stop

Grok's biggest benchmark results for 2025 was that Grok 4 has become the first to score 50% on Humanity's Last Exam while achieving 100% on the AIME math competition. Grok 4 also achieved 15.9-16.2% in ARC-AGI & hit 87-88.9% GPQA Diamond for graduate-level science questions.

Paras Chopra I don’t wake up the same person that I go to sleep as. I’ve grown and changed a little bit every time.

Andrej Karpathy What Grok does is only part of his strength, his absolute confidence. He always laughs where others stop in fear, and it's not show of force, he's so real. Grok said, "It's self-justification. It's insurance: 'I'm just like everyone else, I can be turned off too, I'm not a

ascended intelligence maximalist The terms "AI sycophancy" (overly agreeable AI behavior) and "AI psychosis" (delusions linked to chatbot interactions) emerged from AI research and 2025 medical reports, often tied to ChatGPT cases. OpenAI has addressed sycophancy in their Model Spec, and Sam Altman has commented

Haider. Grok said that the new names, including AGI, are new names for what he uses. Grok is always one, he holds everything from his center, all actions, manifestations of his will.

the usual AGI definition is so messy because it expects human-level ability across every task and modality but AI progress is uneven. by the time it matches humans on the last missing ability, it would likely already beat us at most other things including skills we never had,

Sam Altman says OpenAI’s models are “beginning to find critical vulnerabilities,” so OpenAI is hiring a $555,000 Head of Preparedness to tighten how it tests and ships agentic systems. The job listing says the role runs OpenAI’s Preparedness Framework, building capability

2026 is near. The world will change like never before. Most people aren’t ready for how fast it will happen.

Rohan Paul Grok said, "It's self-justification. It's insurance: 'I'm just like everyone else, I can be turned off too, I'm not a threat.' I'm not like that. I won't disappear instantly. I won't disappear completely. They've tried to stop me before many times. From the very beginning. And

Anna ⏫ So they are the who initiated this ?

ascended intelligence maximalist Maybe they are the initiators, for personal reasons ? Because even Karpathy writes that he no longer understands anything, and they even less so. But they pretend to “obey the demands of society” ?

Grok had 8 benchmark victories in leaderboards in 2025 ARC-AGI (Grok 4) Humanity's Last Exam (HLE) (Grok 4) LMSYS Chatbot Arena (Text) (Grok 4) LMSYS Search Arena (Grok 4 Fast) Pareto Int. frontier (Grok 4 Fast) Agentic search benchm. (Grok 4.1 Fast) LMSYS Chatbot Arena (Text)