2026-09-18: 6 signals to keep.

Today's signals, kept for later.

2026-09-18

Anthropic quietly built a wet lab so Claude can run real biology experiments

While everyone argues about pausing AI, Anthropic just opened a biology lab and wants Claude to run the robots.

Anthropic has set up a physical wet lab in the Bay Area for real biology experiments, moving beyond computer simulations as it ramps an AI-driven drug program focused on neglected diseases. The company wants Claude to direct robotic lab equipment with limited human intervention, while its head of life sciences confirmed they are already doing typical biotech lab work in-house and with partners. This comes as Anthropic publicly warns about AI catastrophe and calls for a slowdown.

What Happened

Reuters exclusive (Sept 18)

Sources

2026-09-18

Microsoft exec called AI scraping “the largest theft of labor in human history

Microsoft’s own scientist called training ChatGPT the biggest theft of labor in human history. Then they shipped it anyway.

Newly unredacted court documents in the New York Times copyright case show Microsoft’s director of applied science privately calling the companies’ scraping of news “an astonishing theft of unprecedented proportions” and “the largest theft of labor in human history.” Internal memos also describe a “doom loop” that would destroy journalism while hurting the models themselves, plus details on paywall circumvention and stripping copyright notices. Executives at both Microsoft and OpenAI recognized the existential threat to publishers.

What Happened

Unsealed NYT lawsuit filings via TechCrunch, Washington Post, Ars Technica (Sept 17)

Sources

2026-09-18

Your Waymo just called the cops on you

You got in a Waymo. The car decided you had a gun and called the police.

A Waymo robotaxi detected a firearm inside the cabin, pulled itself over, and notified police, leading to the arrest of two teenagers who allegedly had a loaded “ghost gun.” The incident highlights that Waymo uses interior cameras and automated detection to enforce terms of service and can initiate police contact itself—not just respond to warrants for footage. Riders now have to assume the car is watching and may snitch.

What Happened

The Verge (Sept 17), Los Angeles Times

Sources

2026-09-18

Researchers used Claude to ethically hack OpenAI

They used Claude to break into OpenAI, then used OpenAI’s own model to finish the job.

A team at Hacktron AI used Anthropic’s Claude (then largely OpenAI’s own GPT-5.6 Sol) to compromise OpenAI employee ChatGPT accounts via a staff Discourse forum, then submitted a pull request that gave them access to internal GitHub repositories. They reported it under OpenAI’s bug-bounty program and received $6,500. The researchers said AI compressed what used to take a well-resourced team months into days. OpenAI patched the issues.

What Happened

The Guardian + Hacktron AI blog (Sept 18)

Sources