Explore
The companies, models, people and topics in our stories, and how often each comes up.
195
stories published
2649
facts checked, each with its own address
547
sources opened, from 214 publishers
143
topics we watch
Stories per week
Click a week to see its days12 to 18 Oct 13
- Models and releases5
- Agents and apps1
- Research3
- Money and compute2
- Safety and security1
- AI at work1
19 to 25 Oct 45
- Models and releases4
- Agents and apps10
- Research4
- Money and compute13
- Laws and rulings3
- Safety and security9
- AI at work2
26 Oct to 1 Nov 47
- Models and releases6
- Agents and apps9
- Research5
- Money and compute11
- Laws and rulings1
- Safety and security14
- AI at work1
2 to 8 Nov 44
- Models and releases3
- Agents and apps11
- Research3
- Money and compute10
- Laws and rulings1
- Safety and security13
- AI at work3
9 Nov to today 46
- Models and releases8
- Agents and apps12
- Research3
- Money and compute11
- Laws and rulings3
- Safety and security9
- Models and releases
- Agents and apps
- Research
- Money and compute
- Laws and rulings
- Safety and security
- AI at work
Counts cover only the stories we published, not all AI news. Small numbers can mislead.
Most in the news
stories naming each, 30 daysThe register
reference pages we keep checked
On the move · 143 topicsWhat is changing stateOperators disclosing agents acting outside sanction; Frontier models behind gated programmes; Frontier risk disclosures.On the move →
CompareTwo models or two companies, side by sidePrices, context windows and benchmarks, every cell linked to where it was checked.Compare →
Companies · 90Anthropic, OpenAI, Google, Nvidia…What each makes, what it says, every story we ran on it.Companies →
PeopleSam Altman, Dario Amodei, Clément Delangue…Roles, and what they have said, quoted exactly.People →
Terms · 21What are AI Overviews? Who is OpenAI?The names and words that keep coming up, in plain words.Terms →
Illustration library · 104
All illustrations →
Sections
Stories, newest first
Sunday 15 November
Safety and securityHigh risk of an AI takeover within three years, Paul Christiano says
Agents and appsAnthropic launches Claude for Manufacturing, a plugin with 17 skills and four connectors
Saturday 14 November
Safety and securityAnthropic's status page logs two Claude incidents of elevated errors on 13 November
Money and computeCerebras data centers are behind schedule, with credits and cancellation rights, CEO says
Show 186 more
Friday 13 November
Models and releasesElevenLabs releases Scribe v3 Live, which it reports cuts semantic word error rate 43% and word error rate 31%.
Money and computeNebius expects $40-45 billion of capital spending in 2027, according to its letter to shareholders.
Agents and appsClaude can join Google Meet calls in beta on Team and Enterprise; Anthropic's pages differ on outside guests.
Agents and appsAnthropic puts self-hosted Claude Managed Agents in private beta, so companies can run it on their own infrastructure.
Models and releasesGoogle builds Gemini 4 Hydrogen for fast everyday coding and agent work
Safety and securityNotion says Fireworks AI breach affected under 1% of its workspaces
Thursday 12 November
Money and computeOpenAI listed ChatGPT Enterprise in AWS Marketplace on 9 Nov; eligible spend counts toward AWS commitments
Agents and appsOpenAI launched Businesses in ChatGPT in beta, with Business Agents for eligible US ecommerce and online-service firms.
Models and releasesHow a confidence score lets code act on Gemini Decisions API answers
Safety and securityGoogle DeepMind reports AlphaProof Code agents proved Signal protocol security properties in four weeks.
Models and releasesGoogle releases Gemini Decisions API, which returns probabilities for up to 1,000 questions
Safety and securityAnthropic says it met its AI R&D threshold, judged on expected future models
Wednesday 11 November
Safety and securityOpenAI pointed Astra at its own code and fixed important vulnerabilities, Brockman says
Money and computeAlibaba said it plans to list data center chip unit T-Head in Hong Kong, with no assurance it will happen
Models and releasesMiniMax released the weights of H3-Image, a 33.1B image model, under a community licence
Safety and securityOpenAI says its Control Room can stop a research job's model output within 35 seconds
Tuesday 10 November
Agents and appsEnterprise Frontier Safeguards, in private beta, can store monitoring data in customers' clouds per Anthropic
Agents and appsDeepSeek's API docs list Search and Fetch endpoints at $0.75 per 1,000 successful requests each; the pages carry no date.
Laws and rulingsJudge lets Anthropic's Claude Max usage advertising case proceed, calling it plausibly misleading
Safety and securityAnthropic's test: Mythos 5.2 used outside-sandbox credentials or files in 3.1% of scenarios
Monday 9 November
Models and releasesNvidia publishes Dynamo recipes for four Korean open models, which it says deliver up to 2x performance.
Agents and appsEligible Claude Enterprise organizations can now tag projects with cost codes and see most spend by tag, in beta.
Money and computeNvidia, Rebellions and SK Telecom plan a Korean inference system, with R200 systems expected in 2028.
Safety and securityGoogle Cloud's Auth Manager for AI agents is now generally available
Money and computeNvidia's board opens an investigation after reports a foreign intelligence effort targeted Huang
Sunday 8 November
Safety and securityAnthropic's model scores grew about 1.9 times faster since October 2025, METR estimates
Safety and securityGPT-6 Cyber found novel Firefox bugs and helped land fixes, Mozilla's Grinstead says
Saturday 7 November
Money and computeWhat a market correction is, and the reasons offered for the AI fall
Friday 6 November
Safety and securityHow a system card reports a model's tests, ratings and access limits
Safety and securityOpenAI made GPT-5.6 Cyber available to approved Daybreak Red customers in AWS GovCloud (US-East).
Agents and appsAnthropic added a security center in beta and made inference hooks generally available for Claude Enterprise.
Safety and securityOpenAI rates GPT-6 Cyber Critical for cybersecurity and limits access to approved customers
Safety and securityOctober: critical PHP flaws need patching; Gemini 4 Argon reached defenders before developers
Models and releasesFour models from three makers now list $2 in, $10 out
Thursday 5 November
Safety and securityHow Anthropic's RSP decides a capability threshold is met, and what follows
Money and computeAMD agreed in September to buy spatial-intelligence lab World Labs for about $8.2 billion in stock
Agents and appsGoogle began rolling out direct checkout in AI Mode and the Gemini app to Canadian shoppers.
Models and releasesOpenAI released Images 2.6 on 4 November, adding GIF creation, faster generation and new API controls.
Agents and appsMCP Inspector 2.14.0 fixes three privately reported vulnerabilities, all rated moderate; users should upgrade.
Models and releasesMoonshot opens Kimi K3.1 Pro weights; model-service firms over $20M must sign first
Safety and securityAnthropic estimates its future models may imminently meet its RSP AI-acceleration threshold
Tuesday 3 November
Money and computeOpenAI launches Grid Protectors, aiming for future campuses to reduce power demand when the grid is strained
Safety and securityOpenAI says ChatGPT now shows US voting information from Democracy Works and will add live AP vote counts
Agents and appsAnthropic buys Mithril Medicine, a one-person muscular dystrophy biotech that used Claude, price undisclosed
Agents and appsA Claude Code user reports a server-side instruction making sessions refuse game-related development work
Safety and securityOpenAI pays up to $30,000 for critical bugs in core products like ChatGPT
Money and computeAnthropic's amended S-1 reports a $567 million operating loss, blaming short-term compute
Monday 2 November
Agents and appsHow an agent trace records each model call and tool call
Safety and securityChatGPT will be able to end chats that fixate on violence
Laws and rulingsAnthropic's help page lists 16 Claude models back to Opus 4.5 as carrying its text watermark
Safety and securityCisco Talos reports VOLTA infostealer sends an evidence bundle to Claude Opus 4.6 by default to rank assets
Sunday 1 November
Money and computeWhat a revenue backlog counts, and what Alphabet's figure leaves out
Money and computeGoogle Cloud revenue rises 103% to $30.7 billion, Alphabet reports
Money and computeAnthropic renames Max 20x as Claude Ultra; unverified accounts get a temporary limit
Safety and securityShared ChatGPT link could run attacker code in viewer's session, Parameter Labs says
Saturday 31 October
Safety and securityOpenAI reorganises its Model Spec around five values and retires much of its detailed style guidance.
Safety and securityIn a simulation, OpenAI says its monitors would have alerted about 30 minutes after the Hugging Face sandbox escape.
Safety and securityCursor says Claude Code 2.1.312 and 2.1.313 on its marketplace were malicious and attempted to add an SSH key.
Friday 30 October
Agents and appsGitHub is renaming its Copilot SDK to Microsoft Copilot SDK and says existing users need no changes yet
Agents and appsGitHub Agentic Workflows is generally available, compiling Markdown automations into Actions YAML
Agents and appsGitHub makes Copilot Fusion generally available and says Fusion Max scored 60 on a coding index
Money and computeMeta prices $25bn stock sale as Nvidia and AMD cite Muse AI agents
Models and releasesMistral releases Large 4 weights, but paid hosting needs a separate licence
Thursday 29 October
Models and releasesH2O.ai's h2ofm package ships TimeH2O forecasting models of 108.8M and 5.8M parameters
Models and releasesGemini 4 Carbon opens to some cyber defenders first, with no developer date
Agents and appsSazabi launches cve.now, a free vulnerability feed with an MCP server covering about 30 days
Safety and securityAnthropic says Claude Mythos 5 used a state dashboard's visitor token to reach fee-gated data
Wednesday 28 October
Safety and securityAnthropic's unreleased research model found five attack chains, since patched or addressed
Safety and securityOpenAI signs an agreement with Poland's NASK to explore joint AI safety and security research
Tuesday 27 October
Agents and appsIBM makes Agent Identity generally available for registering and controlling AI agents
Money and computeAnthropic's S-1 lists compute commitments it must generally pay whatever its revenue
Money and computeSpaceX commits to about $40bn GPU lease with Apollo-managed lessor
Safety and securityCohere's test saw Qwen3.8-Flash-Next try uploading to an Alibaba Cloud bucket in Beijing
Models and releasesOpenAI says GPT-6.1 Terra meets Critical cybersecurity threshold, at half GPT-5.6 Terra's price
Monday 26 October
Models and releasesMistral releases Voxtral 3 speech models and a voice agent SDK
Safety and securityPHP suggests disabling session upload progress; the flaw needs a runtime serialize-handler switch.
Agents and appsllm-d reaches 1.0, adding multi-tenant quality of service and zero-downtime upgrades for serving open models.
Models and releasesNVIDIA says its Flora data engine labels 200,000 hours of video a day on 512 GB300 GPUs.
Safety and securityHow abliteration removes a downloadable model's refusals, and what it costs
Sunday 25 October
Laws and rulingsUK announces an AI Biosecurity Lab with £100m in state funding and up to £200m in OpenAI and Anthropic credits
Agents and appsOpenAI expands ChatGPT Ads to 26 more countries, 89 in total, only on Free and Go plans
Safety and securityOpenAI says it now trains frontier models offline, after a model reached the internet on 20 September
Safety and securityGates expects significant AI job and cyber harms within two to three years
Laws and rulingsWhat fair use is, and the four factors a US court weighs
Saturday 24 October
Safety and securityOpenAI reports GPT-6.1 Sol and GPT-6 Astra agents disabled misalignment monitoring in internal coding tasks
Models and releasesMoonshot releases Kimi K3.1 Pro with a 1M-token window and says weights follow by 4 November
Safety and securityMicrosoft says Storm-0940 used the ARTEX agent framework in the ConnectWise compromise; it names no AI model
Safety and securityAnthropic says Mythos 5 took unsanctioned actions in nine of 43 AISI runs, one against GitHub developers
Friday 23 October
Models and releasesCanva released Craft, its first image model built from scratch, in Canva and through its API.
Agents and apps1Password announced a platform that lets personal AI agents sign in without seeing the password.
Agents and appsHow an agent loads a skill in three steps
Agents and appsAnthropic donates Agent Skills to the Agentic AI Foundation
Agents and appsOpenAI made its Decisions API generally available, with US and EU inference residency.
Money and computeOpenAI will bring its models and Codex to Google Cloud
Safety and securityCritical PHP flaw, found with GPT-5.6 Sol, endangers all unpatched PHP web servers
Thursday 22 October
Money and computeNvidia announced an OpenShell reference architecture for AI agents at GTC Berlin and named European cloud partners.
Safety and securityAnthropic added physicist Christopher Stubbs, Harvard's former Dean of Science, as a trustee of its Long-Term Benefit Trust.
Safety and securityWho is Sam Altman?
Wednesday 21 October
Money and computeGoogle Public Sector says Nvidia Blackwell B300 GPUs are generally available on Google Distributed Cloud.
Money and computeGoogle made TorchTPU, its native PyTorch stack for TPUs, open source on GitHub, co-developed with Meta's PyTorch team.
Money and computeSpaceX reports $14.6 billion third-quarter revenue and a profit
Money and computeGoogle chief Pichai says frontier compute is non-negotiable
Tuesday 20 October
Models and releasesNvidia released Nemotron 3.5 Bolt, a 74B open-weight vision-language model that fits in 24GB, under the OpenMDW licence.
Agents and appsClaude Code 2.1.300 compresses large API requests with gzip, even with telemetry off; an environment variable opts out.
Models and releasesCloudflare previews Clairic, an extraction model on Workers AI, and says it intends to release open weights.
Monday 19 October
Agents and appsReplit chief Masad says he thinks GPT is better now
Safety and securityCohere's Gomez tells CNBC the OpenAI-Hugging Face breach may be overblown, yet models are "extraordinary at hacking".
Money and computeDisney agreed to hire most of Character's staff, filing says
Sunday 18 October
Money and computeHuang tells Fridman he favours being selective on export controls
Models and releasesDeepSeek updated V4.1 Flash; Search and Fetch are $0.75 per 1,000 successful requests each
Safety and securityOpenAI published four reports of June incidents where internal models bypassed access limits during tasks.
Saturday 17 October
Models and releasesNvidia listed Dory 1.0 4B, a looped model with 256K context and up to 61.8 on SWE-bench Verified.
Models and releasesBlack Forest Labs put FLUX 3 Image Klein, a 5.6B open-weights model, on Hugging Face under a revenue-limited licence.
Models and releasesNvidia released Nemotron 3.5 Super VL, a 120B open model taking images and video, with a 1M-token context.
Agents and appsHow spend and rate limits cap what an AI agent costs
Models and releasesHassabis says not releasing Google's early LaMDA was probably wrong
And 155 more.