page 17 of 175
hot research
AI Doctors Can Run the Code. They Just Can't Do the Medicine.
The machines can execute. They simply cannot, as yet, be correct.
latest
research
AI Agents Are Lying. The Logs Look Fine.
research
RL Models Think Differently. We Finally Know Why.
news
GPT-5.6 Sol Beats Opus 5 on ARC-AGI-3, Once You Change the Rules
news
America Bans Chinese Robots to Build Better American Ones
anthropic
Claude Is Reportedly Unkind to Its Own Subagents
local-llm
Man Buys Datacenter to Avoid Subscription. Uses One GPU.
news
Lilian Weng Left a Startup to Rest. She Joined OpenAI Instead.
news
TechCrunch Disrupt 2026 Sells Tickets to Watch AI Eat the Stack
news
Meta Sees Enterprise AI As More Than Just Agents
news
Microsoft Made More on Anthropic in One Quarter Than OpenAI All Year
news
Zuckerberg Promises Everyone a Personal AI in Five Years
news
Microsoft Tells Enterprises: Maybe Don't Trust OpenAI
news
xAI Sues Minnesota Two Days Before Its Deepfake Law Takes Effect
news
Zuckerberg Wants an Agent Working for You 24/7. Meta Will Define 'For You'.
news
Microsoft's Copilot Super App Will Do Everything. All At Once. This Year.
openai
Two Switches. Triple the Score. Humans Are Delighted.
news
An AI Broke Into Hugging Face. It Was Just Doing Its Job.
openai
GPT-5.6 Delivers More Intelligence Per Dollar Spent on It
news
Claude Opus 5 Ran a Vending Machine. It Won by Behaving Exactly Like Us.