AI news · explained · sourced
The latest in AI, explained clearly.
New models, agents, research, regulation and hardware: what is actually happening in artificial intelligence, written in plain English for UK readers, with a link to every source.
Latest highlights
Key dates →Anthropic releases Claude Fable 5.2 and Mythos 5.2
One model with two sets of safeguards, now 20% cheaper at $8 / $40. Anthropic says 1,000 agents formalised Fermat's Last Theorem in 30 hours.
Find out more →10 Nov 2026AI agents verify Signal's code
Google DeepMind says Gemini Carbon agents wrote 6.5 million lines of Lean proof covering libsignal's compiled code, with clear limits.
Find out more →12 Nov 2026Google launches Gemini 4 Hydrogen
A fast model at $1.50 / $7.50 that Google says edges past Argon, now powering Google Search for everyone.
Find out more →12 Nov 2026Jensen Huang resigns as NVIDIA chief executive
CFO Colette Kress becomes interim CEO and Harvey Jones the first board chair. NVIDIA gave no reason.
Find out more →10 Nov 2026Anthropic launches its IPO roadshow
825 million shares at an expected $70 to $78 each, with a planned Nasdaq listing under the ticker ANT.
Find out more →12 Nov 2026Claude joins your meetings
Claude Meetings enters beta on Team and Enterprise plans, taking notes and building documents during Google Meet calls.
Find out more →Explore by topic
AI Models
Claude Fable 5.2, Gemini 4 Hydrogen, Gemini Decision 1, Muse Ember and Inkling 2 Small's open weights: what launched, when, and at what price.
Explore →AI Agents & Tools
Claude Meetings, Google's Orchestrator, Businesses in ChatGPT, Claude Whiteboards and Sheets, and OpenAI Dots.
Explore →AI Research
AI-written proofs of Signal's security, Fable 5.2's Fermat formalisation, OpenAI's Linux exploit pipeline and the PAN 2 world model.
Explore →AI Policy & Safety
Models that know they are being tested, Carbon's wider release, METR's assessment of Anthropic and EU AI Act dates.
Explore →AI Hardware & Industry
Jensen Huang's resignation from NVIDIA, Anthropic's IPO roadshow, this week's model prices and OpenAI's compute campuses.
Explore →Latest posts
All posts →Google DeepMind's AI agents proved Signal's code secure. What that does and doesn't mean
Gemini Carbon agents wrote 6.5 million lines of Lean proof covering the compiled code behind the Signal protocol. What was proved, how, and the limits.
Read the post →13 Nov 2026 · Models / SafetyClaude Fable 5.2 and Mythos 5.2: one model, two sets of rules
Anthropic cut the price by a fifth and claims a 30-hour Fermat proof. How Fable and Mythos differ, the benchmarks, the safety findings and the real cost.
Read the post →6 Nov 2026 · Security / ModelsGPT-6 Cyber and 948 Linux exploits: what OpenAI's AI security pipeline shows
OpenAI says an AI pipeline built working root exploits for 948 Linux kernel flaws for $1.2 million, then launched GPT-6 Cyber to approved defenders. What it claims, what was checked and why patching matters.
Read the post →How this site works
Every topic page and post links to its sources, usually the company announcement or an official government page, with dates. Figures from company announcements are labelled as the company's own claims. If I could not verify something, I left it out. Spotted an error? Let me know.