Skip to content
Home
Latest
Models
Agents
Developers
Industry
Topics
Newsletter
Search stories, models, companies…
/
Subscribe
Home
Latest
Models
Agents
Developers
Products
Industry
Research
Explainers
Guides
Benchmarks
Compare
Topics A–Z
Popular topics
OpenAI
83
Anthropic
81
Google Gemini
45
Claude Code
27
Coding agents
51
Model Context Protocol (MCP)
13
The week in AI, in one email.
Email
Subscribe
The river
Latest
Every story as it’s published, newest first.
Latest
Trending
48h
Most discussed
Saturday, Oct 24
18 stories
Products · Guide
ChatGPT Locked Chats: how to lock a chat with a PIN or Face ID, and what it doesn't protect
ChatGPT locked chats, launched October 20, 2026, hide chosen conversations behind a 4-digit PIN or Face ID. How to lock one, which devices and plans get it, what happens if you forget the PIN, and why locking isn't encryption.
18:17
·
7 min read
Agents · Analysis
OpenAI's Patch the Planet disclosure framework: the five stages for AI-found vulnerabilities, how PHP2Shell tested it, and what it changes
On October 21, 2026, OpenAI proposed a five-stage Patch the Planet framework for disclosing critical vulnerabilities found by AI agents, using the PHP2Shell flaws as its worked example. What the stages are, who decides, how the PHP window actually went, and how it compares with Project Zero's 90-day rule and Anthropic's OSS Scanner.
18:17
·
8 min read
Developers · Compare
TorchTPU vs TorchNeuron: running PyTorch on Google TPUs and AWS Trainium compared (October 2026)
TorchTPU vs TorchNeuron: in the same week of October 2026, Google and Amazon open-sourced native PyTorch backends for their AI chips. Both register a new device through PyTorch's PrivateUse1 hook and replace an older PyTorch/XLA path. Here's how they compare on status, chips, PyTorch versions, torch.compile, distributed training and the code you change, per the official docs.
18:17
·
Checked Oct 23
Products · Analysis
"Publish fake stats, get cited in LLMs": how AI answers get gamed, and what AI search companies say
A Peec AI researcher says 50,000 unchecked, AI-generated statistics pages drew 10,000+ ChatGPT sessions and links from 330+ sites. What the 'fake stats cited in LLMs' experiment shows and doesn't, what the research on gaming AI answers found, and what OpenAI, Google and Perplexity say about sources.
18:17
·
9 min read
Policy · Explainer
Do AI political ads need a label? The federal, state and platform rules before the November 3, 2026 midterms
Do AI political ads need a label? There's no federal AI disclosure rule for the November 3, 2026 midterms. Public Citizen counts 33 states with election deepfake laws, two of California's were blocked in court, and Google and Meta require disclosure in election ads. Rep. Shontel Brown's new bill would add a federal label enforced by the FTC.
18:17
·
8 min read
Research · Explainer
Neural networks are "already clean"? Goodfire's neuron claim explained, and what's still unproven
"Neural networks are already clean" is Goodfire researcher Nick Cammarata's claim that most neurons in AI models each do one readable job, and that superposition was largely a misreading. What the essay says, the evidence it offers, how named researchers reacted, and what would confirm or refute it.
18:17
·
7 min read
Hardware · Explainer
Unitree G1 price, units and specs: what the 'best-selling' humanoid robot costs and can do
The Unitree G1 price starts at $13,500 before tax and shipping, and on October 22, 2026 founder Wang Xingxing said 10,000 to 20,000 have rolled off the production line since May 2024. That's a production figure, not sales. Specs, limits, security history and how the 'best-seller' claim compares with industry data.
18:17
·
6 min read
Research · Benchmarks
Qwen3-Coder writes less secure code for a US defense contractor, Hirundo says: what the study found, and what it doesn't prove
In an October 21, 2026 study, Hirundo says Alibaba's Qwen3-Coder-30B-A3B wrote more hackable web apps when told the code was for a US defense contractor than for a Chinese one. Here's how the test worked, the numbers, the statistics, the limits, and how it compares with CrowdStrike's DeepSeek study.
18:17
·
8 min read
Open source · News
DeepSeek DeepInfer: the Apache 2.0 inference engine behind V4.1 Flash runs on Nvidia Blackwell and Huawei Ascend, with key serving parts held back
DeepSeek DeepInfer is the inference engine DeepSeek says serves DeepSeek-V4.1-Flash in production, open-sourced under Apache 2.0 on October 20, 2026. It supports one model on two kinds of hardware, and the KV-cache and load-balancing systems aren't included.
18:17
·
5 min read
Industry · News
Meta Q3 2026 earnings: capex guidance raised to $135–145 billion and a new non-voting Class C share to fund AI
Meta's Q3 2026 earnings, reported October 21, 2026: $64.6 billion in revenue, $34.9 billion of capital spending in the quarter, a 2026 capex range raised at the bottom to $135–145 billion, and a new non-voting Class C stock that Meta says it will use for capital raising, pay and acquisitions. What the release and filings say.
18:17
·
6 min read
Policy · News
Anthropic copyright lawsuit roundup: Canadian news publishers sue in Ontario, and a judge weighs fair use for song lyrics
A new Anthropic copyright lawsuit: six Canadian news publishers filed in Ontario on October 6, 2026, alleging Claude was built on scraped articles. On October 21, a US judge heard the music publishers' Concord case on fair use and didn't rule. Where every Anthropic copyright case stands.
18:17
·
7 min read
Models · News
Voxtral 3: Mistral's new speech-to-text, text-to-speech and voice-agent SDK are API-only, unlike earlier Voxtral models
Voxtral 3, released October 20, 2026, adds three Mistral transcription models, a text-to-speech model and a voice-agent SDK priced from $0.0035 a minute. Unlike earlier Voxtral releases, none of the Voxtral 3 models has open weights.
18:17
·
5 min read
Open source · News
Recursive Flint: an Apache 2.0, 40B coding model with 2.5B active parameters and a self-reported 89.7% on SWE-bench Pro v2
Recursive Flint is a 40B-parameter mixture-of-experts coding model with about 2.5B active parameters, released under Apache 2.0 on October 21, 2026. Its 89.7% on SWE-bench Pro v2 is Recursive's own number and wasn't on Scale AI's leaderboard as of October 22.
18:17
·
5 min read
Products · News
Perplexity Computer on Mac: what it does locally, who can get it, and the permissions it asks for
Perplexity Computer on Mac can now run fully local: since October 20, 2026, the model, the agent and the task run on an Apple silicon Mac for Pro, Max and Enterprise subscribers. What 'local' covers, the 36GB memory recommendation, and the macOS permissions it asks for.
18:17
·
5 min read
Hardware · News
Nvidia DSX POD and the supply-chain capacity deal: containerized Vera Rubin AI factories, 2.5x substrates and 15x optics
Nvidia DSX POD, announced at GTC Berlin on October 21, 2026, is a reference design for Vera Rubin AI factories built in containers, which Nvidia says can be deployed in as little as 14 days at a prepared site. The same day, Nvidia named six suppliers expanding substrate, packaging, testing and optics capacity. What was announced, and what rests on Nvidia's word.
18:17
·
5 min read
Research · News
KLS conjecture proof: three papers claim the 31-year-old problem, and what AI did in each
The KLS conjecture proof arrived in early October 2026 from more than one team. Microsoft's Xinzhi Zhang and Zhao Song credit a human-directed AI agent search; Bizeul, Klartag and Lehec say most of their proofs were found by ChatGPT. None is peer reviewed yet.
18:17
·
6 min read
Research · Analysis
Claude scorpion toxin binders: Anthropic says 17 of 211 lab-tested designs met its criteria, but none is an antivenom
Anthropic reports that Claude designed candidate protein binders against scorpion toxins in a 64-hour campaign. Two contract labs tested 211 designs; 17 met Anthropic's binding criteria, each against a single toxin. None has been tested for neutralizing venom.
18:17
·
5 min read
Products · News
Meta Muse share links exposed private data: what was public, and what to do about your share links
Meta Muse share pages were public web pages whose addresses carried a title, and researchers found more than 6,800 of them in passive DNS records. Meta paused sharing on October 20, 2026, and turned it back on with new links on October 22. What was exposed, what Meta has and hasn't said, and what to do about links you created.
18:17
·
8 min read
Tuesday, Oct 20
12 stories
Products · News
Meta No Kings Instagram suspension: Meta says traffic from 'a Claude tool' set off its bot defenses
Meta took the No Kings Coalition's Instagram account offline on October 17, 2026, during a national day of protest. A day later Meta said 'a Claude tool' reading the account at about four times the normal rate looked like bot traffic and tripped its automated protections. It did not say which tool, or who ran it. What's confirmed, and what isn't.
12:10
·
5 min read
Policy · News
DeepSeek banned in US federal agencies: what the DHS order requires, and where else DeepSeek is banned
DeepSeek is now banned in US federal civilian agencies. A Department of Homeland Security order, made public on October 19, 2026, bars agencies from buying DeepSeek products and requires them to remove existing ones, including third-party tools built on DeepSeek. What the order says, what it leaves out, and a dated list of the other governments that ban or restrict DeepSeek.
12:10
·
7 min read
Industry · News
Anthropic Public Benefit Fund: what the 5% share pledge commits, what it could fund and what's still undecided
On October 18, 2026, Anthropic said it will commit 5% of its shares to a new Anthropic Public Benefit Fund, subject to stockholder approval, with up to 2% more for each 50% rise in value. Here's how the fund is structured, what it may pay for, what Anthropic hasn't said, and how it compares with the OpenAI Foundation.
12:10
·
5 min read
Open source · News
Mistral Timestral: four open-weight time-series models, a $20M revenue cap, and a CERN statement of intent
Mistral Timestral is a family of four time-series forecasting models, from 5M to 209M parameters, with weights on Hugging Face under a revenue-capped 'Modified MIT' license. The same week, CERN and Mistral signed a statement of intent, not a contract, to build a Large Physics Model.
12:09
·
5 min read
Models · Explainer
What is a diffusion language model? Inception's Mercury 3 explained
A diffusion language model drafts a whole block of text and refines it in parallel, instead of writing one token at a time. Inception's Mercury 3, launched October 19, 2026, claims 1,068 tokens per second, a 1-million-token context and $0.20/$0.70 pricing. Here's how diffusion LLMs work and what the trade-offs are.
12:09
·
7 min read
Industry · Analysis
Is AI usage slowing? What OpenRouter's token data and Gartner's CIO survey actually show
A 'token-maxxing is dead' thread and Gartner's finding that only 13% of CIOs see significant value from AI have revived the question: is AI usage slowing? The public data shows cost discipline and a shift to cheaper models, not a fall in overall use.
12:09
·
7 min read
Industry · Explainer
Coming soon in AI: launches, deadlines and events to watch in November and December 2026
The upcoming AI launches, deadlines and events for November and December 2026, each dated and sourced: model retirements, price rises, the EU AI Act's December 2 deadline, Microsoft Ignite, AWS re:Invent, NeurIPS, and the reported Anthropic IPO window.
12:09
·
6 min read
Models · Explainer
GPT-7 release date: what OpenAI has actually said, and what's reported
There is no GPT-7 release date. As of October 20, 2026, OpenAI hasn't announced GPT-7 at all. Here's what it has shipped (GPT-6 Astra, Sol, Luna and GPT-6.1 Sol), how long its big launches have taken, and what The Information and others have reported about its next flagship.
12:09
·
7 min read
Models · Explainer
Claude 6 release date: what Anthropic has said, and what comes after Opus 5.5 and Fable 5.1
There is no Claude 6 release date. As of October 20, 2026, Anthropic hasn't announced Claude 6, an Opus 6 or a Fable 5.5. Here's the current lineup, how often Anthropic ships, which models retire when, and what Reuters, The Information and leak-watchers have reported.
12:08
·
7 min read
Hardware · Explainer
Upcoming AI glasses: what Apple, Meta, Google, Samsung and Snap have announced or are reported to plan (October 2026)
The upcoming AI glasses worth knowing about, sorted by evidence: what Meta sells today and at what price, what Google, Samsung and Snap have announced for fall 2026, and what Bloomberg reports about Apple's glasses, which Apple hasn't announced.
12:08
·
7 min read
Hardware · News
OpenAI humanoid robot: what Sam Altman has actually said about the 2027 demo, and what's still unknown
Is OpenAI building a humanoid robot? Sam Altman has said yes, twice, in September 2026: "We will definitely do a humanoid," and "I think we'll have a cool demo in 2027." OpenAI hasn't published a robot, a spec, a partner or a date. Here's what's confirmed, what's reported and what's still unknown.
12:08
·
5 min read
Policy · News
EN 18286, the EU AI Act's first harmonised standard, is in the Official Journal: what it covers and what's still missing
EN 18286, a quality management system standard for providers of high-risk AI, became the first harmonised standard under the EU AI Act when the Commission published its reference in the Official Journal on October 14, 2026. What EN 18286 covers, what 'presumption of conformity' means, which standards are still pending, and how it fits the AI Act's high-risk deadlines.
12:08
·
7 min read
Newer
1
2
3
4
5
6
…
11
Page 5 of 11
Older