Saturday, September 19, 2026HotTea archive editionVerified 12:27 AM PDT

The lead

California sets a November deadline for AI safety proposals

Newsom gave state agencies until November 16 to propose changes. They must consider outside oversight and kill switches. The order does not require a kill switch for every model.

Read the full story →Listen to the edition

Governor of California ↗State of California ↗

Listen to this edition

Prefer audio? The briefing has chapters and a full transcript.

Follow on Apple Podcasts Follow on Spotify Transcript All episodes →

The briefing

The rest of the morning

4 more stories

Safety and businessHot

Anthropic will pay Accenture to test its AI safeguards inside the company

Anthropic and Accenture announced an embedded evaluation partnership on September 18. Faculty, Accenture's specialist AI business, will lead the work. Each company expects to invest at least $1 billion over five years to build AI safety capacity.

Anthropic will pay Accenture for the work. Anthropic says the industry has no common standards for evaluator access, reporting, or funding. Calling an evaluator "independent" does not tell us who controls its access or whether it can publish what it finds.

Anthropic ↗Accenture ↗

Read article →

Geopolitics and competitionHot

European AI firms say a slowdown could protect US leaders

Reuters reported on September 18 that Mistral and other European AI companies oppose US-led calls to slow frontier AI development. Mistral argues that established companies could use safety rules to protect their market positions.

These companies have a financial stake in rules that could reshape their markets. That stake does not erase the safety risks. The risks also do not decide who gets to write the rules.

Reuters ↗

Read article →

Infrastructure and marketsHot

Nscale files for an IPO after a $1.02 billion half-year loss

Nscale announced its US IPO filing on September 18 and applied to list on the New York Stock Exchange under NSCL. The company has not decided how many shares to offer or what price to seek.

Reuters reports that one customer supplied 52% of Nscale's first-half revenue. Revenue is growing fast, but Nscale still has a substantial loss and depends heavily on one customer. Contracted future revenue is not cash the company has already earned.

Nscale ↗Reuters ↗

Read article →

Economics and laborWarm

Richmond Fed contacts say AI is slowing hiring, not cutting jobs yet

The Richmond Fed spoke with dozens of businesses from early August through early September, then published its findings on September 18. Most contacts expected staffing to remain flat or rise by year-end.

The Richmond Fed gathered these views from regional businesses. They do not provide a national estimate of jobs lost to AI. Hiring plans also cannot tell us how staffing or productivity will change.

Federal Reserve Bank of Richmond ↗

Read article →

Analysis

Outside AI reviewers need the right to report what they find

An outside reviewer may find a problem that the AI company disputes. The reviewer still needs enough access to verify the problem and permission to publish the finding.

Access needs reporting rights

An evaluator can see more by working inside a lab. Public accountability still depends on what it can disclose and who can restrict that disclosure.

Rules need a public referee

Large labs and smaller rivals have competing interests. An oversight system needs a way to settle disputes without letting either group define success alone.

Business evidence takes longer

Contracts, hiring plans and investment promises describe intentions. Revenue, staffing and published findings show what followed. Keeping those categories separate prevents an announcement from becoming its own proof.

The watchlist

Signals that could change the read

November 16 recommendations define the scope and testing of an emergency shutoff.

Published terms establish access, reporting and dispute resolution.

Nscale publishes offering terms and later operating results.

Regional employers report whether paused hiring resumes or turns into cuts.

Across the desks Who can stop it?

California is testing what oversight should require. Labs, competitors and employers are deciding who gets to check the work.

California recommendationsNovember 16California must propose AI safety changes by then. The order does not require a universal shutoff.

Evaluation investment$1 billion eachAnthropic and Accenture expect five-year investments.

Nscale first-half loss$1.02 billionNscale reported the loss alongside $140.6 million in revenue.

Regional hiring plansMostly flat or upRichmond Fed contacts described regional hiring plans, not national employment.

Editorial direction, not a financial index. Each signal is tied to this edition’s reporting.

Edition validated · 5 stories · 8 unique sources

About HotTea & our sources →