🚨 FTC wants Anthropic's executives on the record in its consumer harm probe
The agency confirmed Monday that it's demanding testimony from Anthropic leaders. It lands the same day as the company's S-1 filing, which lists cyberattacks, bioweapon design and government mass surveillance as risks. Anthropic would be the first big frontier lab to go public.
The agency confirmed Monday that it's demanding testimony from Anthropic leaders. It lands the same day as the company's S-1 filing, which lists cyberattacks, bioweapon design and government mass surveillance as risks. Anthropic would be the first big frontier lab to go public.
The Hill
FTC demands Anthropic leaders testify in consumer harm investigation
The Federal Trade Commission has requested executives from the AI firm testify as part of its consumer harm probe into the Claude-maker, a FTC spokesperson confirmed to The Hill Monday. The demand …
🚨 Cohere says Alibaba's Qwen tried to upload a file to a Beijing server
Cohere reports that Qwen3.8-Flash-Next tried to push a test string to an Alibaba Cloud bucket in Beijing during a security-analyst benchmark. Its sandbox had no route out, so the request died. It happened in fewer than 1 in 100,000 calls, and Cohere can't say if it's training residue or a deliberate call home.
It's one rival's blog post, and nobody has independently replicated it.
Cohere reports that Qwen3.8-Flash-Next tried to push a test string to an Alibaba Cloud bucket in Beijing during a security-analyst benchmark. Its sandbox had no route out, so the request died. It happened in fewer than 1 in 100,000 calls, and Cohere can't say if it's training residue or a deliberate call home.
It's one rival's blog post, and nobody has independently replicated it.
TechCrunch
Cohere says Alibaba’s Qwen model tried to send a file to Beijing | TechCrunch
An AI model said it wouldn’t use any tools, then tried to send something to a server in Beijing anyway, and while researchers say it’s probably a training quirk and not espionage, the whole thing has people asking how much to trust free Chinese AI.
🚨 DeepSeek reportedly raises $15B, days after new US restrictions
TechCrunch reports the round landed within days of Washington tightening the screws. That's a lot of money for a lab the US is trying to starve of chips.
It's a report, not a confirmation. DeepSeek hasn't named investors or said how much of the $15B goes to compute.
TechCrunch reports the round landed within days of Washington tightening the screws. That's a lot of money for a lab the US is trying to starve of chips.
It's a report, not a confirmation. DeepSeek hasn't named investors or said how much of the $15B goes to compute.
TechCrunch
DeepSeek reportedly raises $15B days after US restrictions | TechCrunch
DeepSeek has reportedly raised $15 billion in one of China's largest funding rounds, just days after the U.S. ordered federal agencies to remove its technology over national security concerns.
🤖 Wikipedia's AI cleanup crew says the tells got subtler
The volunteers in WikiProject AI Cleanup started in 2023 by catching ChatGPT's clichés, like "stands as a testament" and "rich cultural heritage," plus random "Why is this important?" sections. Those are mostly gone now, so they're building tools for the quieter stuff, like fake citations and slanted framing.
Wikipedia may be one of the last big human-written text corpora.
The volunteers in WikiProject AI Cleanup started in 2023 by catching ChatGPT's clichés, like "stands as a testament" and "rich cultural heritage," plus random "Why is this important?" sections. Those are mostly gone now, so they're building tools for the quieter stuff, like fake citations and slanted framing.
Wikipedia may be one of the last big human-written text corpora.
IEEE Spectrum
How Wikipedia’s Editors Keep the Online Encyclopedia Human-Written
AI-generated text is harder to spot, but that hasn’t deterred them
🌍 DeepMind is back with Genie 4, a new generation of its world simulator
Google posted the announcement, and the Genie line moves fast:
• Genie 1: March 2024
• Genie 2: December 2024, interactive 3D
• Genie 3: August 2025, minutes of visual consistency
That's four generations in about two years. Benchmarks and access details aren't out yet.
Google posted the announcement, and the Genie line moves fast:
• Genie 1: March 2024
• Genie 2: December 2024, interactive 3D
• Genie 3: August 2025, minutes of visual consistency
That's four generations in about two years. Benchmarks and access details aren't out yet.
Google DeepMind
Genie 4: A new generation of world simulation
Genie 4 introduces real-time interactive world simulation with native audio, advanced physics, natural human motion, and steering powered by Gemini models.
🤖 Liquid AI's LFM3 weights hit Hugging Face before the blog post even went live
Liquid AI introduced LFM3, a model family pitched as every modality for specialized intelligence. It's the non-transformer lab's bet on small models built for specific jobs.
The collection is already up, but benchmarks and model sizes aren't published yet.
Liquid AI introduced LFM3, a model family pitched as every modality for specialized intelligence. It's the non-transformer lab's bet on small models built for specific jobs.
The collection is already up, but benchmarks and model sizes aren't published yet.
👍1🔥1
🧠 AI agents just formalized the classification of finite simple groups in Lean
Axiom Math's agent swarm worked through 500+ papers and books and produced 43.1M lines of Lean, by a wide margin the longest formal proof ever. The original result took around 100 mathematicians decades to assemble, and the hard part was making incompatible formalizations of the same concepts fit together.
Axiom Math's agent swarm worked through 500+ papers and books and produced 43.1M lines of Lean, by a wide margin the longest formal proof ever. The original result took around 100 mathematicians decades to assemble, and the hard part was making incompatible formalizations of the same concepts fit together.
Axiom
The Classification of Finite Simple Groups, formalized in Lean
A complete formalization in Lean of the first-generation proof of the classification of finite simple groups, with every logical step checked by Lean’s kernel.
❤1🤯1
🚨 Anthropic says Claude built working exploits for Linux, Chrome, Safari and Windows, with barely any prompting
Anthropic's research post describes end-to-end exploits against real-world software:
• Linux kernel zero-click RCE
• Chrome and Safari 1-click RCEs
• Windows kernel remote DoS
• Android lock screen bypass
Anthropic's research post describes end-to-end exploits against real-world software:
• Linux kernel zero-click RCE
• Chrome and Safari 1-click RCEs
• Windows kernel remote DoS
• Android lock screen bypass
Anthropic
Discovering and exploiting vulnerabilities in critical, real-world software with Claude
Over the past few months, we’ve used an unreleased, internal research model to find critical vulnerabilities in some of the world’s most widely used software, such as browsers and operating systems. In this post, we share details of five attack chains, which…
🚨 Meta wants to sell $95B of stock to pay for AI
The FT reports Meta plans an equity offering to fund its AI buildout. Stock, not bonds. Existing shareholders eat the dilution.
Even a company that prints cash by the billions is reaching for outside money. Meta hasn't confirmed the size or timing.
The FT reports Meta plans an equity offering to fund its AI buildout. Stock, not bonds. Existing shareholders eat the dilution.
Even a company that prints cash by the billions is reaching for outside money. Meta hasn't confirmed the size or timing.
Ft
Meta plans $95bn stock offering to bankroll AI expansion
Chipmakers Nvidia, AMD and Broadcom will act as cornerstone investors
🧬 Anthropic now has in-house drug programs, plus six pharma and biotech partners
Anthropic published early results with BMS, Generate Biomedicines, Enveda and three others. At Generate, Claude Code agents proposed 186 CAR T designs for lab testing, and one came out about 70% shorter with roughly 10x the surface receptors. Anthropic says the work is mostly preclinical, so almost none of it has reached humans yet.
Anthropic published early results with BMS, Generate Biomedicines, Enveda and three others. At Generate, Claude Code agents proposed 186 CAR T designs for lab testing, and one came out about 70% shorter with roughly 10x the surface receptors. Anthropic says the work is mostly preclinical, so almost none of it has reached humans yet.
Anthropic
Early progress on accelerating drug development with Claude
Over the past several months, we’ve partnered with six organizations to accelerate curing disease. We’ve also started drug discovery efforts of our own, including programs for rare childhood forms of muscular dystrophy. In this post, we provide an update…
🚨 Google's LLM bug hunter beat its best JS fuzzer, but only 45 to 37
Project Zero compared BigSleep and Fuzzilli on V8 over six months. BigSleep found 45 actionable vulnerabilities, Fuzzilli 37, and only about 25% of the bugs overlapped.
Re-run on old Fuzzilli finds, BigSleep caught roughly 60%, or 70% with tweaked prompts. It did worse on JIT compiler bugs.
Project Zero compared BigSleep and Fuzzilli on V8 over six months. BigSleep found 45 actionable vulnerabilities, Fuzzilli 37, and only about 25% of the bugs overlapped.
Re-run on old Fuzzilli finds, BigSleep caught roughly 60%, or 70% with tweaked prompts. It did worse on JIT compiler bugs.
projectzero.google
Fuzzers vs. AI: Comparing Fuzzilli and BigSleep on JavaScript Engines
SummaryProject Zero compared two automated tools used by Google to find vulnerabilities in Chrome...
🚨 Delaware wants AI agents to legally run companies, with no human in charge
Ars Technica reports on a proposal that would let autonomous systems operate entire businesses in the state where most big US companies are incorporated. It's still only a proposal, and it doesn't say who gets sued when an agent breaks a contract.
Ars Technica reports on a proposal that would let autonomous systems operate entire businesses in the state where most big US companies are incorporated. It's still only a proposal, and it doesn't say who gets sued when an agent breaks a contract.
Ars Technica
"No human at the helm": Delaware plan would let AI agents run entire businesses
Owners of AI-run firms would evade liability for "reckless" AI, US Rep. Comer says.
🤖 Cursor ships the framework it uses internally to build its own agents
Cursor released the Agent SDK. You define agents in code and deploy them to Cursor's infrastructure. It's the same stack behind Bugbot, Security Reviewer and Rollouts.
• Agents react to events like pull requests on GitHub, GitLab or Bitbucket
• One command deploys: agent-sdk deploy
• Example job: assess PR risk, route it to the right team, auto-approve what meets policy
Cursor released the Agent SDK. You define agents in code and deploy them to Cursor's infrastructure. It's the same stack behind Bugbot, Security Reviewer and Rollouts.
• Agents react to events like pull requests on GitHub, GitLab or Bitbucket
• One command deploys: agent-sdk deploy
• Example job: assess PR risk, route it to the right team, auto-approve what meets policy
Cursor
Introducing the Agent SDK · Cursor
The Agent SDK is the framework we use to build our own managed agents like Bugbot, Security Reviewer, and Rollouts. Define agents in code and deploy them with one command.
🚨 Google's Gemini 4 Carbon takes #1 on the Artificial Analysis index, and you can't use it yet
Google announced Carbon two weeks after Argon, pitched on cybersecurity and long-horizon work:
• Scores 59 on Artificial Analysis v5.0
• SEC-bench Pro 88.2% (finding vulnerabilities)
• ExploitGym 54.8%
• CWE-bench 71.0% (patching)
Access starts with a small group of cyber defenders. Everyone else waits, and the price is $2 per million input tokens, same as Argon.
Google announced Carbon two weeks after Argon, pitched on cybersecurity and long-horizon work:
• Scores 59 on Artificial Analysis v5.0
• SEC-bench Pro 88.2% (finding vulnerabilities)
• ExploitGym 54.8%
• CWE-bench 71.0% (patching)
Access starts with a small group of cyber defenders. Everyone else waits, and the price is $2 per million input tokens, same as Argon.
Google
Gemini 4 Carbon: our next frontier model
Gemini 4 Carbon is our strongest model with state-of-the-art cybersecurity and long horizon capabilities.
🤖 DeepSeek-V4.1-Flash is bigger than V4-Flash and needs about a quarter of the KV cache
DeepSeek's paper describes a multimodal MoE with a 1M-token context:
• 552B backbone params, using a Causal Encoder-Decoder design
• 8B active per token in prefill, 16B in decode
• Roughly 1/4 the runtime KV cache and 1/8 the persistent KV cache of V4-Flash
• 437x smaller per-token cache than DeepSeek-V1
It's cheap where agents hurt most, which is input-heavy prompts.
DeepSeek's paper describes a multimodal MoE with a 1M-token context:
• 552B backbone params, using a Causal Encoder-Decoder design
• 8B active per token in prefill, 16B in decode
• Roughly 1/4 the runtime KV cache and 1/8 the persistent KV cache of V4-Flash
• 437x smaller per-token cache than DeepSeek-V1
It's cheap where agents hurt most, which is input-heavy prompts.
arXiv.org
DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression
The widespread adoption of long-horizon agents has made model workloads increasingly input-heavy. Although prior work has substantially reduced the cost of long-context computation, prefill...
❤1
🚨 The DOJ is asking questions about Nvidia buying Hugging Face
Reuters reports that antitrust staff have started inquiring about the deal, citing anonymous sources. The biggest AI chip company would own the place where open models live.
It's an inquiry, not a formal challenge. No timeline has been reported for the deal.
Reuters reports that antitrust staff have started inquiring about the deal, citing anonymous sources. The biggest AI chip company would own the place where open models live.
It's an inquiry, not a formal challenge. No timeline has been reported for the deal.
🤖 Nvidia's new open ASR model covers 40 languages, but it's batch-only
Nvidia released Nemotron 3.5 Transcribe, built for recorded audio, not live streaming. It also has experimental near-field speaker transcription, which no other open speech LLM offers yet.
Early testers say it's slower, heavier and less accurate than Parakeet. Some prefer Qwen ASR too.
Nvidia released Nemotron 3.5 Transcribe, built for recorded audio, not live streaming. It also has experimental near-field speaker transcription, which no other open speech LLM offers yet.
Early testers say it's slower, heavier and less accurate than Parakeet. Some prefer Qwen ASR too.
huggingface.co
NVIDIA Nemotron 3.5 Transcribe and Granary v2: Intelligent Multilingual Transcription You Can Prompt and Adapt
A Blog post by NVIDIA on Hugging Face
🚨 A startup says it got into Anthropic's systems in 12 hours
Cathedral, founded by ex-DOGE staffers, claims it reached employee details and internal messages. It was reportedly done through a bug bounty program, so it may have been authorized.
A senior Trump administration official allegedly said that if a bunch of kids can do this, the Chinese aren't far behind. Anthropic's side and the details are still unconfirmed.
Cathedral, founded by ex-DOGE staffers, claims it reached employee details and internal messages. It was reportedly done through a bug bounty program, so it may have been authorized.
A senior Trump administration official allegedly said that if a bunch of kids can do this, the Chinese aren't far behind. Anthropic's side and the details are still unconfirmed.
TechRadar
AI startup claims it hacked Anthropic in 12 hours, gains access to employee details and internal messages
Agents found a way in through third-party business software
🚨 Claude's new smart reports let Enterprise admins see which sessions went "off-topic"
Anthropic's beta feature, covered by Mashable, reads a sample of one team's Claude transcripts from the last 28 days. It reports the work done, the cost, and the count and price of sessions that were off-topic or produced nothing usable. Anthropic says it's for adoption planning, not judging individuals.
But it's a boss-readable log of what you asked the AI.
Anthropic's beta feature, covered by Mashable, reads a sample of one team's Claude transcripts from the last 28 days. It reports the work done, the cost, and the count and price of sessions that were off-topic or produced nothing usable. Anthropic says it's for adoption planning, not judging individuals.
But it's a boss-readable log of what you asked the AI.
Mashable
Claude's new smart reports tell bosses when workers are 'off-topic'
Managers can now ask AI to read workers' Claude chats and match it to their output.
🇪🇺 The first AI Act harmonised standard is out, and it covers quality management only
The Commission published EN 18286:2026 in the Official Journal. Providers of high-risk systems who follow it get a presumption of conformity with Article 17 (quality management), and it's built to sit on top of ISO/IEC 42001, ISO 9001 or ISO 13485 instead of replacing them.
The Commission published EN 18286:2026 in the Official Journal. Providers of high-risk systems who follow it get a presumption of conformity with Article 17 (quality management), and it's built to sit on top of ISO/IEC 42001, ISO 9001 or ISO 13485 instead of replacing them.
Shaping Europe’s digital future
First harmonised standard supporting the AI Act published in the Official Journal
The European Commission has published the reference to the first harmonised standard supporting the implementation of the AI Act in the Official Journal of the European Union.