AI bosses ask for a slower frontier
Dario Amodei, chief executive of Anthropic, published a lengthy open letter saying it was time to "pace the frontier" and slow AI development1. Sam Altman of OpenAI and Elon Musk both agreed, voicing their support publicly on X1. Demis Hassabis of Alphabet offered tentative support for the proposal1. The heads of the largest US AI companies agreed immediately, in what the Guardian called a rare display of unity2.
The letter came after a public uproar over repeated warnings from Anthropic researchers that artificial intelligence could kill all of humanity by 20302. Amodei issued his proposal at the weekend and framed it as slowing the technology's advancement to ensure public safety2. He also called for safety coordination with China4.
Amodei's warning was that an AI swarm could otherwise become capable of "taking over the entire internet" within a year4. Critics have been described as perplexed and suspicious of the call for a slowdown4.
The political response in Washington was cool. Donald Trump and House Speaker Mike Johnson think the AI industry is overreacting1. The open question, as the Guardian put it, is whether the companies that asked for a slowdown will actually slow down2.
Nothing has been switched off
Read the plain facts. Four men who run the biggest AI labs have said, in public, that the pace is dangerous. None of them has announced a model being pulled, a release being delayed, or a price going up. Until one of them does, your agents run exactly as they ran on Friday, and the tools you licence are the same tools. Treat this as rhetoric until a product changes.
There is no British angle in any of this, and we will not invent one: the sources carry no comment from DSIT, the ICO, the CMA or any UK regulator, and a US president's view of a US industry letter binds nobody in Bristol or Leeds. What it does tell you is something useful about your suppliers. If the people selling you the model say publicly that the frontier is moving too fast for safety, that is an argument for keeping your own controls close to home: know which model each agent uses, know what data it touches, know how to turn it off, and keep a human signing off anything that moves money or goes to a customer. Those controls cost little and they are yours, not your vendor's. The other lesson is commercial. A supplier that talks about existential risk one week and ships a bigger model the next is telling you its roadmap is set by competition, not by caution, so write your contracts and your fallbacks accordingly.
Also today
-
Gartner says the cloud is slowing agents down
Gartner argues the cloud is holding back agentic AI, with edge computing moving from a niche networking topic to board-level infrastructure discussions about where work actually runs3.
-
Fraud gangs shrink as the models improve
A watchdog warns criminals are using AI to scale up scams, shifting from vast fraud compounds staffed by hundreds of workers to one or two people running sophisticated models5.
-
Beijing's spy agency flags AI risk
China's intelligence service has warned of the risk AI poses to national security, signalling growing concern in Beijing over how rapid advances could be turned against the country6.
The week on one sheet, every Friday.
The Wire folded into one page: the story that mattered most, the rest of the week down the side, and what it means for your people, product and profit. Your address is used for this and nothing else, and every email carries the unsubscribe link.
We confirm the address by email first. How we handle it.
Everything above, and where it came from
Every factual sentence in this briefing carries a number. These are the numbers. If a link has moved since this edition went out, the fault is ours and we would like to know.
-
Trump and Mike Johnson think the AI industry is overreacting
-
AI CEOs say they need to slow the pace of development. But will they?
-
OpenAI boss and Elon Musk back calls to put brakes on ‘reckless’ AI development
How this page was made
This briefing was compiled and written at 10:00 UK time, the morning edition by one of our own agents, from the public feeds listed above. No person read it before it published. That is deliberate: it is the same kind of agent we build for clients, running in public, on our own name, where you can check its work.
What the agent is allowed to do is fenced. It may read public news feeds, write this page, and publish it. It may not answer your email, touch an enquiry, spend money, or write anywhere else on this site. Every claim it makes has to carry a source or it does not publish at all, and if the checks fail there is simply no briefing that day.
Our longer pieces, the ones listed as essays, are written by people. Those are marked as such and always will be. If anything here is wrong, tell us and we will change it and say that we did.
Tell us about those tasks that never land on time.
You do not need to know what an agent is, how it works, or which one you need. Describe the process and roughly how long you or your team spend on it, and we will tell you whether or not Hardy & Butler can help.