AI policy, chip buildouts, and dev workflow pivots

The U.S. has lifted a requirement that Anthropic obtain a license before exporting its Mythos and Fable models abroad, a requirement that effectively cut off public access to what are widely considered the most advanced AI models released to date.

The AI lab said it would begin restoring access to the models on Wednesday, July 1.

On June 12, the U.S. government had added the products to its list of export-restricted technologies, meaning they could no longer be made available to foreign nationals without special approval. Complying with that rule proved impractical at scale, forcing Anthropic to end public access to the models altogether.

Now, after weeks of talks, Secretary of Commerce Howard Lutnick said Anthropic “has agreed to proactively detect and address security risks associated with the models; to work diligently with the U.S. government on protocols and standards and releases for Mythos, Fable and future models; and to inform the US government of any malicious activity.”

Anthropic had already publicly pledged to do much of this voluntarily, months before the export rule existed. That’s part of why cybersecurity experts were skeptical of the restrictions in the first place.

The report says ByteDance is building a $39B data-center complex in a free-trade zone in Ceará, Brazil.

The report emphasizes this as ByteDance’s largest such buildout outside China.

The capacity figure is listed at 1GW.

The item is positioned as a major AI-capacity expansion and regional compute play.

For infrastructure-minded teams, the announcement suggests another serious concentration point for AI workloads in Latin America.

The Hacker News entry points to Google Copybara and describes it as tooling for moving code between repositories.

The external link is Google’s open-source repo at github.com/google/copybara.

The topic is relevant for teams trying to standardize migration behavior across codebases.

Discussion volume is moderate with 209 upvotes and 36 comments.

The linked post describes an ngrok write-up claiming Kubernetes can be used through a browser-based flow.

The main idea is to reduce environment setup friction when interacting with cluster control flows.

That can matter for ephemeral training, troubleshooting, and demos where local toolchain install is a bottleneck.

Engagement sits at 265 upvotes and 80 comments on Hacker News.

Wayve, a U.K.-based self-driving tech startup, is allowing its employees to sell a portion of their vested equity. The $85 million tender offer — essentially a structured opportunity for employees to sell shares back to investors — is being led by the company’s existing and new investors at the company’s latest valuation of $8.5 billion .

That valuation was set in February when the nine-year-old company raised a $1.2 billion Series D led by Eclipse, Balderton, and SoftBank Vision Fund 2, and included participation from Ontario Teachers’ Pension Plan, Baillie Gifford, Microsoft, Nvidia, and Uber.

This is Wayve’s second employee liquidity event. The company previously held a tender offer alongside its $1.05 billion Series C funding round in May 2024.

Wayve’s offering is part of a growing trend of AI startups. Rather than waiting years for an exit, companies are using tender offers as a retention tool, giving employees a reason to stick around rather than jump to a competitor — or start their own shop — the moment their options vest.

Other startups that have recently completed employee tender offers include Decagon , which builds AI agents that handle customer service for enterprises like Duolingo and Hertz; ElevenLabs , the AI voice-generation company behind much of the internet’s synthetic speech and dubbing tools; Linear , a popular project-management platform built for software teams; and Clay , a sales and marketing automation tool that helps companies research and reach prospects. (Clay has run two tenders in the last nine months alone.)