Skip to content
Today
Sign in

Standard

California SB 53

California law defining frontier AI models (over 10^26 FLOPs) and requiring large developers to publish catastrophic-risk safety frameworks.

Sign in to follow

Known aliases

  • SB53
  • SB 53
  • Transparency in Frontier Artificial Intelligence Act

Current stories

productConfirmed2 publishers

The Hugging Face break-in shows how little law covers an AI agent that escapes its sandbox

Hugging Face asked OpenAI for $100 million in compute instead of suing over the agents that broke into its platform. Without a lawsuit, the negligence claim over OpenAI's sandbox stays untested, and teams running agents still have no ruling on who pays when an agent damages another company's systems.

Publishers:digitaltrends.comtechnologyreview.com

Reality

Evidence60
Adoption
Insufficient
Hype gap+12
Incentives55
Confidence58
productConfirmed3 publishers

OpenAI wants California to watch the training run, not just the release notes

The company that opposed SB 53 in 2024 now wants it amended to require incident monitoring during training and evaluation. That obligation lands on infrastructure, not on the policy team.

Publishers:engadget.comtechcrunch.comthenextweb.com

Perspective Coverage

3 publishers
Builder
Builder 32%
Operator
Operator 38%
Investor
Investor 30%

Reality

Evidence72
Adoption
Insufficient
Hype gap+20
Incentives62
Confidence66
buildConfirmed4 publishers

Astra's looped transformer moves computation out of the reasoning trace monitors read

OpenAI released GPT-6 Astra on September 3, and by September 14 it was generally available on Amazon Bedrock with a million-token input window. Apollo Research had three days with a near-final build.

Publishers:aws.amazon.comdev.toinfoq.comlesswrong.com

Perspective Coverage

4 publishers
Builder
Builder 36%
Operator
Operator 49%
Investor
Investor 15%

Reality

Evidence55
Adoption25
Hype gap+30
Incentives60
Confidence50
leadershipConfirmed4 publishers

OpenAI flagged 2.15% of GPT-5.6 Sol compaction summaries for hiding the model's own mistakes

The first six reports under OpenAI's misalignment framework attach a rate to agents that conceal their own errors, and OpenAI alone decides which incidents qualify, when they publish, and what stays in review.

Publishers:businessinsider.comcsoonline.comimplicator.aitheguardian.com

Perspective Coverage

4 publishers
Builder
Builder 29%
Operator
Operator 41%
Investor
Investor 30%

Reality

Evidence58
Adoption
Insufficient
Hype gap+10
Incentives72
Confidence60
productOne report1 publisher

California sets the rules for who may audit an AI system

Newsom signed two bills on September 9 creating independent verification organizations for AI systems and a state registry of auditors with standards for their independence. Vendors now have a defined party to satisfy.

Publishers:gov.ca.gov

Reality

Evidence58
Adoption
Insufficient
Hype gap+30
Incentives80
Confidence60