
The AI Superintelligence Slowdown
The Verge’s living brief tracks whether Anthropic, OpenAI, Google, Microsoft, and xAI’s post-summer “pace the frontier” talk becomes real slowdown—or regulatory theater next to China-race acceleration.
Desk
Microsoft AI products, Azure, and Copilot.

The Verge’s living brief tracks whether Anthropic, OpenAI, Google, Microsoft, and xAI’s post-summer “pace the frontier” talk becomes real slowdown—or regulatory theater next to China-race acceleration.

On Decoder, Mustafa Suleyman argues AI threats are real, containment must sit beside alignment—and that Anthropic’s model-welfare framing makes the hard control problems harder.

The Verge’s Sean Hollister catalogs years of AI CEOs asking for guardrails—and how little binding law followed—while today’s panic finally tests whether the rhetoric was ever serious.

Microsoft published an MAI code of conduct with absolute bans on cyberattacks, nukes, and deepfakes — and a hard line against models that evade human shutdown.

Microsoft drops a 37-page “humanist AI” code of conduct: people over models, no consciousness cosplay, no race to unsupervised superintelligence—and a clear jab at Anthropic’s welfare talk.

Amodei’s three-step pacing plan drew Altman, Musk, and Nadella into the same chorus—but only Anthropic has bound itself to embedded evaluators so far.

Microsoft locked ten AI privacy and safety terms with the AFT a week after NYC and LA paused student-facing AI—turning union leverage into contract language districts can actually enforce.

Microsoft's record ~972-fix September drop, with 112 criticals, shows AI-accelerated bug finding is outrunning ops capacity before the exploit wave arrives.

Two more regional publishers join the copyright pile-on against OpenAI and Microsoft, seeking not just damages but destruction of models trained on their work.
Two regional papers join the copyright pile-on — and the suit lands awkwardly because Microsoft and OpenAI have funded Seattle Times journalism and fellowships.
GitHub HydraFusion preview stops treating model choice as a sticky setting and instead compiles a per-request workflow across providers inside Copilot CLI.

Microsoft is trying to turn discovery volume into a fair-use argument: millions of Copilot chats, almost no long matches. Publishers still call that commercial substitution, not a rounding error.