1/ Did @OpenAI just break California’s new AI safety law?

The answer appears to be yes, and OpenAI could owe millions in fines. 🧵 2/ Since December, OpenAI has been racing to reclaim the title of best coding AI.

Yesterday, they launched GPT-5.3-Codex, regaining the lead — but there's a problem: they may have broken California law to do it.

Something strange happened on conservative Twitter on Thursday.

A dozen right-wing influencers suddenly became passionate about semiconductor export policy, posting nearly identical (and often false) attacks over a 27-hour period on a bill most people have never heard of.

🧵 The AI OVERWATCH Act is a Republican bill that would let Congress review AI chip exports to adversaries like China. It's backed by Microsoft and right-leaning think tanks.

But starting January 15, influencers called it pro-China sabotage and a Democrat plot, all in unison.

Today, @TheMidasProj and @Tech_Oversight released The OpenAI Files, a comprehensive investigation into safety issues, integrity concerns, and employee testimonies at @OpenAI as they consider restructuring to a for-profit. Some highlights from the report: (1/16) Dario and Daniela Amodei, who left in 2020 to found Anthropic, described Altman’s tactics as “gaslighting” and “psychological abuse”

For months, we've been calling upon the team at @cognition_labs to make an industry-standard risk evaluation plan. They've consistently refused.

Now, one day after their product launch, a live streamer revealed a massive security vulnerability in front of an audience of 6,000.🧵 What happened?

The tl;dr is that, when sharing your screen while using Devin, confidential access to Devin's workspace is granted via a URL revealed in plaintext with no auth.

In other words, anyone who sees your screen can access your (and Cognition's) sensitive data.