Dario Amodei, co-founder and chief executive officer of Anthropic, in San Francisco, California, on April 30, 2026.Jason Henry/Bloomberg/Getty Images


As Anthropic’s CEO Dario Amodei calls for his industry to slow the pace of AI development—lest it give rise to machines that cannot be controlled—and rival OpenAI CEO Sam Altman agrees, the lingering question is, well…what’s stopping everyone from stopping?

It seemed like fingers were pointing in every direction over the weekend as top players in the AI arms race agreed with the sentiment but no one actually dared to pump the brakes.

“Dario is right,” SpaceX chief Elon Musk wrote succinctly.

“We welcome the research, focus, and deliberate pacing needed to get alignment right as the design goal,” wrote Microsoft CEO Satya Nadella. 

“Dario's essay points towards the right path forward,” Google DeepMind’s Demis Hassabis wrote, linking to his proposal for an industry-wide standards body for frontier AI.

So who ought to take the first step? Not the AI companies, noted former White House AI and crypto czar David Sacks. “Go ahead. You guys are the frontier,” wrote Sacks, referencing Anthropic and OpenAI.

Not federal lawmakers, either: “There is an obvious corporate responsibility that the people who are creating these models have to ensure that their products are safe,” said Mike Johnson, Speaker of the U.S. House of Representatives, in a CNN interview, adding that a “moratorium” was a nonstarter given competition with China.

And the White House? Well, don’t hold your breath. “We can put guardrails, we can do this and that, but I think you have a lot of negative forces that are bringing it up that shouldn’t be bringing it up and they’re bringing up things that won’t happen,” U.S. President Trump told reporters in Ireland. 

Full speed ahead, I guess. —AN

Six Chinese AI companies, U.S. officials alleged earlier this week, found a shortcut to closing the gap with American AI labs by buying bulk subscriptions to their American rivals and then training on the outputs.

The allegations are one potential explanation for why Chinese AI models are estimated to be neck-and-neck with their U.S. counterparts. But analysts say there’s a different advantage Chinese labs developed to compete: squeezing out more value from limited resources. 

There’s a technique Chinese labs perfected that stems from “attention,” the mechanism Google researchers introduced in a landmark 2017 paper that underlies every large language model. Attention is what lets a model look at each piece of text in relation to all the others and decide which connections matter most.

“Chinese labs found algorithms that reduce the complexity of those calculations by an order of magnitude, and then achieve better results because they’re able to summarize the most relevant tokens,” Brendan Burke, the semiconductors and supply chain analyst at tech research firm Futurum Group, tells Fortune.

Necessity was the mother of invention. “Because [the Chinese] had less compute to work with, they found that computationally efficient method instead of just throwing more compute at an inefficient technique, as U.S. labs initially did,” Burke said.

The trade-off between the two can be quantified. Ameya Kanitkar, co-founder of the AI measurement platform Larridin, told Fortune that in observed enterprise workflows, Chinese AI models handle around 75% of engineering tasks “reasonably well” at one-fifth the cost of the U.S. ones. —Mia Osmonbekov

The CEO of the world’s largest health and beauty retailer is pushing back against using AI to shrink headcount, arguing that the technology should be used to upgrade human work instead of replacing it.

“Is it really that, with AI, we’re going to cut out 30% of our workforce? Or… are we going to increase the quality of our people’s intelligence?” Malina Ngai, Group CEO of AS Watson, said at the Fortune Leaders Forum on Sept. 8.

Companies, particularly in the tech sector, have laid off tens of thousands of employees this year as they pivot towards AI. Oracle, for example, cut 21,000 jobs, or roughly 13% of its workforce, over the past year.

Instead, Ngai—who previously led AS Watson’s digital transformation and expansion of its online commerce offerings—argued the retailer is pushing a “human-AI partnership” strategy.

“You should bring the AI assistant with you to work. It helps you to think faster and smarter,” she said. 

Since the strategy launched last October, employee engagement scores have “jumped up a lot,” she said, as staff spend more time in face-to-face interactions rather than facing a screen. The company pairs that with a “customer love score” which Ngai said keeps rising as store staff are freed to serve customers. 

“This is not the usual way that people would measure AI,” she acknowledged.

Ngai argued AI is an opportunity to change how retail itself works. “Going forward,” she said, “I think retail is going to become more human, not less.” —Nicholas Gordon

—Automattic’s Matt Mullenweg survives an ouster. 

—SoftBank secures a nearly $12 billion loan to support its OpenAI investment.

—The U.S. Interior Secretary has reportedly met with AI hyperscalers to advance the use of federal lands for data centers.

—Bending Spoons loves a price hike. A recurring page in the playbook of the company that owns Airtable, AOL, Eventbrite, Evernote, Meetup, and Vimeo.

—U.S. NSA plans a major restructuring into