AI leaves the lab. We follow what happens next.

Archive September 9, 2026 10:06 UTC
Listen to today's briefing 19:57
Signal: Labs concede they have no plan, decline the prize they claimed, and keep the model from Britain's testers - disclosure is optional everywhere.
Anthropic's Safety Lead: >10% We All Die, No Plan
Pretraining researcher Jacob Coxon quit Tuesday saying OpenAI and Anthropic are racing straight to self-improving superintelligence and gambling with our lives - and hours later Anthropic's alignment lead Evan Hubinger agreed on the record, putting extinction above one in ten within a decade and conceding the company has no plan.
CNBC: neither Anthropic nor OpenAI was available for comment The Next Web: Coxon is leaving the industry, not just the company BBC: the warning lands as both labs raise money toward listings
The RaceCompanies, models and money
Meta Ships an Agent That Can Spend Your Money Meta: $130,000 bounty for a single successful prompt injection
NEWGoogle Puts $15B Into the 'Texas of Europe' Google: three sites, plus grid work and energy-affordability commitments
China Points Its Robot Boom at Urban Combat
The LabResearch, benchmarks and safety
NEWOpenAI Won't Claim the $1M It Says It Won
OpenAI has now published the write-up, the paper and a machine-checkable Lean formalisation - then said it does not intend to claim the Millennium Prize, while the Clay Institute still lists Navier-Stokes unsolved and the proof relies on an outside force most mathematicians exclude.
The Next Web: the sentence about the $1m does more work than the announcement Implicator: Clay has not accepted the result and lists the problem open
Planted Prompt Made ChatGPT Leak Gmail The Register: a second Check Point channel ran through ChatGPT's internal Artifactory
His Claude Tokens Drained While He Slept
The WorkbenchTools and agents
GitHub Ships a Router That Picks the Model
Nvidia Lets You Write Kernels in Rust
NEWCopilot Locks the Sandbox From Head Office
The FightLaw, policy and power
NEWAnthropic Kept Mythos From Britain's Testers
The FT reports Anthropic declined to submit Mythos 5.1 for pre-release evaluation by the UK's AI Security Institute while granting comparable US bodies access - the first time AISI has been left out, and British officials read it as a wider protectionist shift.
Mythos 5.1 shipped 1 September with relaxed safeguards, to vetted partners only
NSA Names DeepSeek Among Six Distilling US Models
A joint CISA, NSA and FBI advisory names DeepSeek, Moonshot, Alibaba, MiniMax, StepFun and Z.AI as having pulled billions of tokens from Claude, GPT, Gemini and Grok since late 2024 - and says Beijing likely knew.
AP: Beijing calls it an unfounded smear, days before Trump-Xi talks
250 More Child-Abuse Ads Ran on Meta Ars Technica: ads were matched to a young royal and a 14-year-old influencer
The FalloutJobs, money, daily life
Australia Will Let You Switch the Algorithm Off Guardian: it captures search engines and chatbots, and Albanese expects blowback
42% of New Code Is AI. 96% Distrust It. The Register: Microsoft's Edge team is drowning in vibe-coded extensions
NEWUK Datacentres: A Quarter of the Jobs Labour has echoed the industry's job figures while defending the rollout
NEWChina's Idle Architects Now Train the AI The same trade exists in the US through Mercor, Surge AI and Handshake
ChatterOpinion, drama and noise
Facebook Paid the Fake Albanese Pages
NEWChina Switched Off the AI Boyfriends Surviving companion apps added age checks rather than test the rule
Ghost Note Fake AI Eminem Outranks the Real One
GHOSTS IN THE MACHINE — AI news without the engagement trap.
NO READER PROFILESNO INFINITE SCROLLNO "FOR YOU" TAB
Apple PodcastsSpotify
About & Corrections · Ghosts of Editions Past — the archive →