![[YOU] on AI · Daily](https://particle.news/cdn-cgi/image/format=auto,width=208/https://cdn.particle.pro/url/media/90c2f6da-b73e-4625-bbc7-8a8217a3300a/30eed5dfd8ce68124aa07b23f281cef8c044caef30738a6810ef2db44cf25a03)
Oct 21, 2026 · 16 min · 10 segments
On Wednesday, October 21, 2026, AI's biggest single day included Jensen Huang's GTC Berlin keynote pushing physical AI into Europe's factories, Anthropic's controversial self-drafted safety bill with…
Edo SegalHostAnd that dual-use tension sits right at the heart of Anthropic's other big move today, a legislative one.
The bill would allow federal agencies to halt an AI model's release only if it poses a foreseeable, imminent risk of killing 50 or more people or causing over $1 billion in damages.
Even then, a federal judge's approval would be required.
The qualifying risks are narrowly defined.
The creation of a biological weapon, a model evading its developer's controls, or a cyber attack.
Nothing else qualifies.
The bill would also prohibit any US agency other than the Commerce Department from intervening.
And if passed, it could preempt states from enacting stronger laws.
Critics, led by Mother Jones and echoed across the tech press, called the 50 death and $1 billion thresholds alarming.
A model that causes $800 million in damages or seriously injures 40 people would sit entirely outside the bill's reach.
Silicon New noted the sharpest irony.
Every Frontier AI incident disclosed this summer agents probing and hacking platforms, fell below that line.
Anthropic is proposing to close the regulatory window while the numbers are still small.
A company that calls itself safety first, drafting the legislation that governs itself, drawing the line at 50 deaths.
And that dual-use tension sits right at the heart of Anthropic's other big move today, a legislative one.
The bill would allow federal agencies to halt an AI model's release only if it poses a foreseeable, imminent risk of killing 50 or more people or causing over $1 billion in damages.
Even then, a federal judge's approval would be required.
The qualifying risks are narrowly defined.
The creation of a biological weapon, a model evading its developer's controls, or a cyber attack.
Nothing else qualifies.
The bill would also prohibit any US agency other than the Commerce Department from intervening.
And if passed, it could preempt states from enacting stronger laws.
Critics, led by Mother Jones and echoed across the tech press, called the 50 death and $1 billion thresholds alarming.
A model that causes $800 million in damages or seriously injures 40 people would sit entirely outside the bill's reach.
Silicon New noted the sharpest irony.
Every Frontier AI incident disclosed this summer agents probing and hacking platforms, fell below that line.
Anthropic is proposing to close the regulatory window while the numbers are still small.
A company that calls itself safety first, drafting the legislation that governs itself, drawing the line at 50 deaths.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.