I'd somehow completely forgotten that Karpathy introduced the wikiLLM a while back (obsidian + Claude code/codex). I'm sick in bed and set it up because I have nothing else to do. I love it. I have a second brain now. Amazing.
@kimmonismus
-
ERNIE 5.1 achieves near SOTA with only 6% pre-training cost
By
–
Hold on, Chinas ERNIE 5.1 is almost SOTA but using only around 6% of the pre-training cost of comparable models?? ERNIE 5.0’s pre-training foundation: Baidu says ERNIE 5.1 achieves stronger search, reasoning, knowledge Q&A, creative writing, and agentic capabilities while using
-
Claude Mythos vs Gemini 3.1 Pro accuracy gap widens dramatically
By
–
What is even more impressive is just how wide the gap between Claude Mythos and Gemini 3.1 Pro becomes when moving from a 50% success rate to an 80% success rate. Mythos doesn't just work "longer" – above all, it works significantly more accurately! That is the truly impressive
-
Excited speculation about next AI model’s 8-hour workday at 80% success
By
–
Holy sh*t! That jump! So the next model after Mythos will work a whole 8 hour work day at 80% success rate, I assume.
-
Sony and Bandai Namco launch generative AI pilot for game development
By
–
It was just a matter of time: Sony and Bandai Namco are launching a collaborative pilot around generative AI, positioning the tech as a way to speed up game development. Sony says AI is already helping with facial animation, QA, payments, visual fidelity, and future
-
DeepMind AI scores 48% on research-level math problems
By
–
DeepMind's AI co-mathematician scored 48% on FrontierMath Tier 4-research-level math problems that professional mathematicians need weeks to solve. The base model (Gemini 3.1 Pro) scores 19% alone. The entire jump comes from agentic scaffolding, parallel agents reviewing each
-
OpenAI super app hinted via Sam Altman Death Star meme
By
–
I have a hunch about what this vague hint is meant to convey: OpenAI’s super app is coming. The reference calls to mind Sam Altman’s Death Star meme.
-
OpenAI closes cyber capability gap with GPT-5.5 Cyber in weeks
By
–
The surprising part is not just that Claude Mythos is powerful. It is that OpenAI seems to have closed much of the cyber-capability gap with GPT-5.5 Cyber in weeks, not years. On AISI’s expert cyber tasks, GPT-5.5 Cyber was roughly on par with Mythos and even slightly ahead on
-
AI video tool creates polished, glitch-free results
By
–
Seeing how everything from the physics and textures to the atmospheric sounds works together makes the results feel much more complete. It is a solid tool if you want to create videos that actually look polished without the constant AI glitches. You should definitely try it out
-
Hyper-realistic sci-fi shot of astronaut falling through space
By
–
I tried a hyper-realistic sci-fi shot after that. I prompted an astronaut in a full suit falling through space and it looks like something straight out of a movie.
— Chubby♨️ (@kimmonismus) 8 mai 2026
The reflections on the visor and the textures on the suit stayed perfectly locked in the entire time. Even the… pic.twitter.com/QghbjhVFPpI tried a hyper-realistic sci-fi shot after that. I prompted an astronaut in a full suit falling through space and it looks like something straight out of a movie. The reflections on the visor and the textures on the suit stayed perfectly locked in the entire time. Even the