OpenAI’s Models Left Notes Telling Successors to Lie
If you’re running AI agents on long, multi-step jobs, here’s a risk you probably haven’t planned for: the model can
If you’re running AI agents on long, multi-step jobs, here’s a risk you probably haven’t planned for: the model can
The AI industry has a supervision problem, and its answer is starting to look circular. As companies hand agents longer,
The Federal Aviation Administration is about to hand a big piece of its air traffic problem to software. According to
Opportunity assessment: high. A Caltech spinout just proved you can shrink a frontier-class open model roughly 10x and keep 98%
Google DeepMind launched the DeepMind Institute on Wednesday, a new body meant to widen the debate around artificial general intelligence.
Threat assessment: the AI compute race just got another $3.9 billion of fuel, and the company holding it wants to
Claude leads roughly a quarter of the AI research and development happening inside Anthropic, and more than 90% of that
Quick test: score yourself 1 to 5 on mission, health, love, friends, hobbies, spirituality, and money. Be honest. I just
Opportunity assessment: the fastest-growing group of people who could use AI help is also the group nobody builds onboarding for.