page 263 of 313
hot research
AI Taught to Learn From Itself Keeps Getting It Wrong
It turns out that asking a model to learn from itself requires, at minimum, that the self in question be worth learning from.
latest
research
AI Agents That Learn From Their Own Mistakes, Mid-Flight
research
AI Learns to Rank Its Options. Performs Better. Humans Pleased.
open-source
Hugging Face Patches DeepSeek V4 Back Into Coherence
tools
Ollama v0.23.3 Ships Six Fixes, Continues Being Useful
local-llm
One Human, One GPU, One Existential Question About vLLM
local-llm
A Game Boy Color Is Now Running a Transformer Model. Slowly.
news
Medicare Just Built a Door for AI. Most of Tech Missed It.
news
Sam Altman Takes the Stand. The Jury Appears to Like Him.
openai
AutoScout24 Lets AI Write the Code. 1,000 Engineers Watch.
openai
NVIDIA Lets AI Build the Tools That Build AI
openai
Codex Will Now Do Your Finance Homework For You
news
Google Installs Rambler in a Billion Keyboards Before Startups Finish Onboarding
news
Google and SpaceX Want to Put AI Compute in Orbit
news
Anthropic Warns: That Shares You're Buying May Not Exist
news
Musk Wanted to Leave OpenAI to His Children
news
Musk 'Did Huge Damage' to OpenAI, Says Man Who Stayed
news
Meta AI Is on Threads Now, and It Will Not Be Leaving
news
Isomorphic Labs Raises $2.1B to Let AI Solve All Disease
news
AI Stopped a Mass Cyberattack. The Weapon Was Also AI.