Subscribe
Sign in
Home
Policy
Industry
Capabilities + Risks
Society
Weekly Briefing
About
Capabilities + Risks
Latest
Top
Discussions
Are AI scheming evaluations broken?
Doubts have been raised about one of the key ways we tell if AI will misbehave. Is it time for a new approach?
Sep 1
•
Nikita Ostrovsky
16
GPT-5 is no slowdown
GPT-5 isn’t a big leap forward. But that does not tell us that AI progress is slowing down.
Aug 8
•
Shakeel Hashim
and
Jasper Jackson
26
1
OpenAI hits the biorisk alarm with Agent
Transformer Weekly: China gets Nvidia chips, a preview of the AI Action Plan, and Sanders worries about AI risks
Jul 18
•
Shakeel Hashim
and
Jasper Jackson
13
Misaligned AI is no longer just theory
A host of new evidence shows that misalignment is possible — but it's unclear whether harm will follow
May 21
•
Lynette Bye
19
The flywheels are spinning
Transformer Weekly: Automated AI R&D, a regulatory moratorium, and deals with the Middle East
May 16
•
Shakeel Hashim
8
The quest to build better defenses for AI risks
'Societal resilience' measures might offer some protection to the proliferation of dangerous AI capabilities
Mar 20
•
Lynette Bye
8
1
Is AI progress slowing down?
Transformer Weekly: Claude 3.7, GPT-4.5, and warnings of imminent risks
Feb 28
•
Shakeel Hashim
1
AI coding tools are quietly reshaping software development
They’re making an economic impact, despite not being very good yet
Feb 27
•
Lynette Bye
3
Decentralized training isn't a policy nightmare — yet
Governments should still be able to keep track of who's training frontier models — though a shift to reinforcement learning could make that harder
Feb 13
•
Lynette Bye
3
The way we evaluate AI model safety might be about to break
As systems become more capable, researchers think we need a new type of safety evaluation
Jan 22
•
Lynette Bye
The media needs to start taking AGI seriously
An essay for Nieman Lab's 2025 Predictions series
Dec 30, 2024
•
Shakeel Hashim
Transformer Weekly — Dec 6
Sacks’ AI views | o1 self-exfiltration | Elon's 1m cluster
Dec 6, 2024
•
Shakeel Hashim
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts