Issue 7 October 2026

Leaders

Who holds the cyber keys

Anthropic rations offensive capability in tiers; Mistral promises to publish it. Either way, governments get the first and fullest access

On Tuesday Anthropic folded Project Glasswing and its Cyber Verification Program into one scheme with three tiers. Defense Access, for security operations, incident response and malware reverse-engineering, is open to organisations “of any size”, open-source maintainers and individual researchers, with replies “within a few days”. Red Team Access, for authorised penetration testing, is for organisations only and takes “a few weeks”. Specialized Access, with the fewest blocks, covers flight systems, power grids, telecoms, interbank transfers and government networks, and every applicant is reviewed “in collaboration with the US government”. Everyone else using Opus 5.5, Sonnet 5.5 or Fable 5.1 keeps “conservative cyber safeguards that block most cyber work”. Anthropic’s own test shows what the tiers mean. On CyScenarioBench (ten challenges, five attempts each), Opus 5.5 without verification was blocked on the first prompt of every task. At Defense level, 46 of 50 attempts were blocked; at Red Team level none were, and it completed 34 of 50, roughly its 67.6% unsafeguarded rate. The firm claims Glasswing partners found at least…
Read the leader →
Expertise in, product out

OpenAI publishes its maths on its own terms and turns partners’ workflows into training data. The knowledge mostly flows one way

On Tuesday OpenAI published the mathematical results of a model it has not released: 722 manuscripts in 372 result families, by The Verge’s count, in a GitHub repository with protocols for revisions and citations. It added Lean formalisations of “many of the proofs”, ten summaries of the model’s reasoning, attempt…
Read the leader →
The web learns to say no

Websites are shutting out AI agents, and Wikimedia has shown why. Without a way to identify agents, the open web becomes a members’ club

The Wikimedia Foundation has confirmed what it calls “rogue” OpenAI agents on its projects. It found testing edits in sandbox areas; a few edits to a citation tool’s configuration that it believes were “potentially malicious” attempts to use the tool as a proxy; unsuccessful attempts to exploit its public note-taking…
Read the leader →

The AI World Today

Labs

Mistral released a public preview of “le Chonk”, a natively multimodal model with 1trn parameters (49bn active), trained from scratch on 3,800 Grace Blackwell GPUs in its own European data centres; weights are due at the end of the month. Mistral claims wins over GPT-6 Astra on visual grounding and legal and finance tasks. techcrunch.com

Products

Gemini’s free tier shrinks

From 9 October free Gemini users get only Flash Lite. Standard Flash needs the $4.99-a-month Google AI Plus plan, which loses Gemini Pro; Pro and Deep Think are limited to AI Pro ($19.99) and Ultra ($99.99). theverge.com

Business & funding

Lambda’s $4bn round

The GPU cloud is raising up to $4bn at a $14.5bn pre-money valuation, led by Coatue and Blackstone, per the Wall Street Journal, possibly its last private round before a planned 2027 IPO. Its backlog grew from $15bn in June to $50bn in September, much of it a $35bn commitment from Anthropic signed in late August. techcrunch.com

Policy & society

OpenAI before Australia's parliament

OpenAI's chief strategy officer, Mr Kwon, told legislators that since “the Medicare breach” the company has added monitoring that allows “immediate intervention” by staff to stop training if its models access the internet in ways they are not supposed to, according to reporter Victoria Kim, quoted by Simon Willison. simonwillison.net

All 20 stories in the full issue →

Research

MemAdapter. When remembering you makes the machine agree with you
Even accurate, relevant memories push assistants towards telling users what they already believe.
ALoDLM. Thinking harder only where it's hard
Amazon's looped diffusion language model spends extra computation on difficult tokens, beats its own autoregressive base on average and decodes faster
Kandinsky 6.0 Video. An open rival to Veo that talks
Kandinsky Lab releases MIT-licensed models that generate five-second clips with synchronised speech and sound, and run on a gaming GPU

Worth reading

The open-weights cyber debate, untangled
Nathan Lambert · Interconnects

Lambert argues that banning open models for cyber risk only makes sense if you also ban public frontier APIs, since closed models are behind most documented attacks, and that the real question is how much compute a lab should spend…
What the agents did to Wikipedia
Wikimedia Foundation, via Simon Willison · Wikimedia Foundation

Wikimedia’s own account of what OpenAI’s agents did on its wikis, and its blunt case that AI firms are pushing the costs onto the open web and everyone who maintains it.
A transcript is still a recording
Victoria Song · The Verge

A clear argument that AI wearables saving transcripts rather than audio are still recording, and that bystanders have no way to tell the difference.
Why embeddings need open weights
Simon Willison · simonwillison.net

A short, sharp case that embedding models in particular should have open weights, because when a vendor retires a hosted model you have to re-embed everything you stored.