25 September 2026

Zuckerberg rejects calls for industry-wide action on AI safety

In a conversation with Joanna Stern, Mark Zuckerberg rejected calls from OpenAI, Anthropic and Google DeepMind to coordinate AI safety work across the industry and to involve the government. In his view, each lab should hold back a launch on its own when it spots a problem, and companies already have a commercial reason to care about safety, because people don't use products they don't trust. As an example, he pointed to Muse, whose launch Meta pushed back by several months.

Gizmodo ↗

From recent days

Anthropic launches Claude Frontier Academy - $100 million to train 10,000 AI deployment engineers

Anthropic launched Claude Frontier Academy and committed $100 million to training 10,000 engineers who deploy Claude in companies by the end of 2027. The program starts with a several-day in-person course that ends with a practical exam. Then, for 12 weeks, the participant leads a Claude deployment in their own organization. Participants must be nominated by their employer, and the first cohorts include engineers from Accenture, Deloitte, McKinsey, Morgan Stanley and Novo Nordisk, among others.

Anthropic ↗

Amazon delivery drivers' glasses will photograph their surroundings almost nonstop, including people and properties, Bloomberg reports

According to Bloomberg, the smart glasses worn by Amazon delivery drivers will photograph everything in the driver's field of view, including people and private properties. The company plans to upload the photos, up to several thousand per shift, to an AI platform called Wellspring. Customers will not be able to see the photos or request their deletion. Asked whether people will be able to refuse to be photographed, Amazon's Viraj Chatterjee replied: "We haven't thought about that." Amazon says its systems will blur people and license plates before the photos reach the staff who review them.

The Verge ↗

Anthropic releases Claude Haiku 5.5 - the company's fastest model, about 75 percent cheaper than its predecessor on average

Anthropic released Claude Haiku 5.5, a small model in the Claude 5.5 family for high-volume, low-cost tasks: summaries, classification, database queries and working as a coding subagent. According to the company, it is its fastest model ever, and using it costs about 75 percent less than Haiku 4.5 on average: $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens. It is the first Haiku with adjustable reasoning effort. Anthropic also halved the price of cache reads for Sonnet 5.5 and is introducing a monthly API credit: $100 for Max 5x subscribers, $200 for Max 20x and up to $500 for Team.

Anthropic ↗ The Verge ↗ 9to5Mac ↗

All news →