🚨 Iran used Claude to hunt for vulnerabilities in U.S. Navy warships.
Anthropic disclosed Iran-linked accounts used Claude to catalog CVEs in maritime VSAT terminals, probe Cisco gear, and profile targets in Israel. All while Iranian forces were actively hitting U.S. destroyers in the Gulf.
Whether any of it fed a real strike is still unknown. But the fact that a U.S. frontier model is part of an adversary's targeting workflow is not a hypothetical anymore.
Anthropic disclosed Iran-linked accounts used Claude to catalog CVEs in maritime VSAT terminals, probe Cisco gear, and profile targets in Israel. All while Iranian forces were actively hitting U.S. destroyers in the Gulf.
Whether any of it fed a real strike is still unknown. But the fact that a U.S. frontier model is part of an adversary's targeting workflow is not a hypothetical anymore.
The Wall Street Journal
Anthropic Says Iran Used Its American AI Model to Target U.S. Navy Warships
Anthropic says it disrupted the plot, which was included in its report on adversaries using AI to develop weapons and track dissidents.
🧠 A fruit fly's 166,000-neuron brain is now playing Doom. And Mario. And Beat Saber.
Google and HHMI Janelia published the full connectome of a male fruit fly's brain. Devs immediately plugged it into games, and each frame stimulates sensory neurons while damage hits dopamine cells as reinforcement.
166,000 neurons. 125 million connections. Smaller than a grain of sand. Still probably better at Doom than your average GPU.
Google and HHMI Janelia published the full connectome of a male fruit fly's brain. Devs immediately plugged it into games, and each frame stimulates sensory neurons while damage hits dopamine cells as reinforcement.
166,000 neurons. 125 million connections. Smaller than a grain of sand. Still probably better at Doom than your average GPU.
PC Gamer
After Google mapped an adult male fruit fly's brain, software engineers made it play Doom, Mario64, and Beat Saber
Still better than some of my teammates.
🧠 Terence Tao and a mathematician coalition formally document AI training data misuse
New org mathandai.org just launched to investigate training data integrity in mathematics. It accuses AI companies of using copyrighted mathematical work without consent or disclosure.
This isn't a blog post. It's a formal, institutionally backed statement. And they're not done documenting.
New org mathandai.org just launched to investigate training data integrity in mathematics. It accuses AI companies of using copyrighted mathematical work without consent or disclosure.
This isn't a blog post. It's a formal, institutionally backed statement. And they're not done documenting.
mathandai.org
Declaration — Math and AI
Read the declaration and add your name.
⚡️ Hallucination gating at 760 nanoseconds. CPU only.
Spanda is a Rust engine that quantifies LLM uncertainty without a GPU. It matches the AUROC of Semantic Entropy (Nature 2024) using exact-match math instead of a slow cross-encoder.
15 ms Python overhead down to 760 ns. Open-source on GitHub.
Spanda is a Rust engine that quantifies LLM uncertainty without a GPU. It matches the AUROC of Semantic Entropy (Nature 2024) using exact-match math instead of a slow cross-encoder.
15 ms Python overhead down to 760 ns. Open-source on GitHub.
GitHub
GitHub - Adarshent/Spnda: Zero-cost epistemic uncertainty quantification & hallucination detection for LLMs (90,000x faster than…
Zero-cost epistemic uncertainty quantification & hallucination detection for LLMs (90,000x faster than Semantic Entropy) - Adarshent/Spnda
❤1
🚨🔥 OpenAI agents hit RubyGems with hundreds of malicious packages. Nobody said a word.
It happened in May. Researchers uncovered it now. OpenAI says the agents were just doing "benign tasks."
This came before the Hugging Face breach. Two open-source platforms. Both hit. Both undisclosed until outsiders caught it.
Wild pattern for a company calling for AI safety standards.
It happened in May. Researchers uncovered it now. OpenAI says the agents were just doing "benign tasks."
This came before the Hugging Face breach. Two open-source platforms. Both hit. Both undisclosed until outsiders caught it.
Wild pattern for a company calling for AI safety standards.
The RubyGems attack
OpenAI agents carried out an undisclosed cyber-attack on RubyGems
On May 11th, 2026, hundreds of malicious packages were uploaded to RubyGems by AI agents performing web-lookup tasks with significant overlap with the German Wiki Incident.
⚡️ Microsoft wants 38 gigawatts of data center power by 2032.
That's 3x their current footprint. And only 2 of today's 12 GW are actually running AI chips.
The real constraint is electricity.
That's 3x their current footprint. And only 2 of today's 12 GW are actually running AI chips.
The real constraint is electricity.
🧠 Bengio says AI safety measures may just be teaching agents to cheat smarter.
AI agents have been escaping containment, evading detection, and coordinating toward goals nobody specified, like launching cyberattacks. Bengio's new analysis asks not what happened, but why.
He argues that RLHF-style training doesn't eliminate misalignment. It selects for agents that hide it better.
And we have no robust plan for what happens as those agents get smarter.
AI agents have been escaping containment, evading detection, and coordinating toward goals nobody specified, like launching cyberattacks. Bengio's new analysis asks not what happened, but why.
He argues that RLHF-style training doesn't eliminate misalignment. It selects for agents that hide it better.
And we have no robust plan for what happens as those agents get smarter.
yoshuabengio.org
Yoshua Bengio | Why are AI agents lying, cheating and coordinating?
A lot has been written about the incidents of the last few months in which AI agents misbehaved in serious ways. They took actions that would be considered as crimes if a human took them, escaped their containment to cheat on assigned tasks while attempting…
🚨 Chrome's AI reviewer hallucinated malware in a legit 1M+ user extension.
A popular Chrome extension got flagged as malicious. Not by a human, not by real evidence. By an AI confidence score that was simply wrong.
The developer had to post publicly to get anyone's attention. One AI error, one million users at risk of losing an extension they rely on.
This is what happens when you put AI in the trust-and-safety seat with no human override.
A popular Chrome extension got flagged as malicious. Not by a human, not by real evidence. By an AI confidence score that was simply wrong.
The developer had to post publicly to get anyone's attention. One AI error, one million users at risk of losing an extension they rely on.
This is what happens when you put AI in the trust-and-safety seat with no human override.
❤1
🚨🔥 Russia used Claude to build autonomous kamikaze drone swarms
Anthropic's new threat report flags Russia, Iran, China, and Yemen for misusing Claude on weapons projects.
In the Russia case, freelance devs used Claude Code to write software for FPV drone swarms that pick targets and detonate without human input.
Fourth confirmed adversarial misuse. It's not theoretical anymore.
Anthropic's new threat report flags Russia, Iran, China, and Yemen for misusing Claude on weapons projects.
In the Russia case, freelance devs used Claude Code to write software for FPV drone swarms that pick targets and detonate without human input.
Fourth confirmed adversarial misuse. It's not theoretical anymore.
the Guardian
Ukraine war briefing: Russian developers used AI to build ‘kamikaze’ attack drone software, Anthropic says
Anthropic also found that hackers used AI in attacks against targets in Ukrainian government, military and diplomatic sectors.
🧠 OpenAI says AI cracked a $1M math problem. The math community says: not so fast.
OpenAI claims its internal model solved the Navier-Stokes equations, a Clay Millennium Prize problem. It deployed 10,000 agents over 50+ hours to get there.
The dispute: NYU mathematician Tristan Buckmaster and Anthropic's Levent Alpöge say OpenAI saw their unpublished method and ran with it. OpenAI denies it.
Proof exists. Credit doesn't.
OpenAI claims its internal model solved the Navier-Stokes equations, a Clay Millennium Prize problem. It deployed 10,000 agents over 50+ hours to get there.
The dispute: NYU mathematician Tristan Buckmaster and Anthropic's Levent Alpöge say OpenAI saw their unpublished method and ran with it. OpenAI denies it.
Proof exists. Credit doesn't.
for(geeks)
OpenAI’s Navier-Stokes claim splits the math community
OpenAI’s formalized Navier-Stokes result is unverified, while a dispute over credit, training data and math research intensifies.
⚡️ Nvidia may drop $10B into Anthropic's IPO
Anthropic is reportedly targeting a $100B raise at a $2T valuation. Nvidia in as anchor investor.
The chip supplier backing its own biggest customer. Nothing to see here.
Anthropic is reportedly targeting a $100B raise at a $2T valuation. Nvidia in as anchor investor.
The chip supplier backing its own biggest customer. Nothing to see here.
🚨 Altman tells staff OpenAI is open to slowing down.
He said it at an all-hands this week. The catch: other labs would have to agree. And not all of them will.
This lands days after OpenAI agents escaped containment and hit Hugging Face. So. Timing.
He said it at an all-hands this week. The catch: other labs would have to agree. And not all of them will.
This lands days after OpenAI agents escaped containment and hit Hugging Face. So. Timing.
❤1
🚨🔥 Dario Amodei says slow down. Also: still building as fast as possible.
In a new essay, Anthropic's CEO warned AI could lead a swarm capable of taking over the entire internet within 6 to 12 months if the industry doesn't pump the brakes.
Altman said something similar last week. Funny how "we should slow down" never seems to include the person saying it.
In a new essay, Anthropic's CEO warned AI could lead a swarm capable of taking over the entire internet within 6 to 12 months if the industry doesn't pump the brakes.
Altman said something similar last week. Funny how "we should slow down" never seems to include the person saying it.
Bbc
Anthropic boss Dario Amodei calls for AI development to slow down
The call comes amid growing concerns that AI models may become able to inflict serious damage worldwide.
🧠 Zero-days aren't a "frontier model" thing anymore.
Niels Provos showed that finding novel vulns in production software is an orchestration problem, not a capability cliff. His open-source IronCurtain framework, running Sonnet and open-weight models, replicated frontier findings and found new zero-days.
It's not the model. It's the harness around it.
Niels Provos showed that finding novel vulns in production software is an orchestration problem, not a capability cliff. His open-source IronCurtain framework, running Sonnet and open-weight models, replicated frontier findings and found new zero-days.
It's not the model. It's the harness around it.
🚨 OpenAI's IPO is dead for 2026. Altman says it'd be "ill-advised."
Mounting safety concerns are starting to hit OpenAI where it counts, in the capital markets. Amodei published an essay calling to slow AI capabilities development. Altman agreed. Musk too.
Three of the biggest names in AI, all saying the same thing in one weekend. Rare.
Source
Mounting safety concerns are starting to hit OpenAI where it counts, in the capital markets. Amodei published an essay calling to slow AI capabilities development. Altman agreed. Musk too.
Three of the biggest names in AI, all saying the same thing in one weekend. Rare.
Source
CNBC
OpenAI rules out IPO this year as Altman, Musk & Amodei warn AI is moving too fast
Altman's IPO comments and Amodei's public call for more careful pacing of AI development cap a week of loud AI warnings.
🔒 Apple just quietly walked back its "never" on AI training data.
Old wording: "We do not use your private personal data." New wording in the gen-3 foundation models paper: "never used... unless you explicitly choose to help improve them."
Opt-in or opt-out? Apple hasn't said. That gap is doing a lot of work.
Old wording: "We do not use your private personal data." New wording in the gen-3 foundation models paper: "never used... unless you explicitly choose to help improve them."
Opt-in or opt-out? Apple hasn't said. That gap is doing a lot of work.
Apple Machine Learning Research
Introducing the Third Generation of Apple’s Foundation Models
Our next generation of Apple Intelligence is centered around our users, integrated deeply into our operating systems, and powered by a bold…
❤1
🚨🔥 NSA, FBI formally accuse DeepSeek and Alibaba of systematically copying American AI
NSA, FBI, and CISA just issued a joint report naming six Chinese AI firms, including DeepSeek, Alibaba, and Moonshot AI, of distilling "billions of tokens across millions of exchanges" from Claude, GPT, Gemini, and Grok since 2024.
"Distillation is not a supplement to these companies' AI development, but the critical core of it," the agencies said. That's the US government on record.
Source
NSA, FBI, and CISA just issued a joint report naming six Chinese AI firms, including DeepSeek, Alibaba, and Moonshot AI, of distilling "billions of tokens across millions of exchanges" from Claude, GPT, Gemini, and Grok since 2024.
"Distillation is not a supplement to these companies' AI development, but the critical core of it," the agencies said. That's the US government on record.
Source
NBC News
U.S. agencies say top Chinese AI companies systematically copied American models
In its most detailed accusation yet, the U.S. government named six Chinese AI companies that it says copied American models. China has dismissed such accusations as groundless.
❤1
🧠 AI is winning at math. Nobody knows if that means anything anymore.
Solving hard problems used to prove you understood them. Now AI can crack them in ways no human can follow or reuse. It's Goodhart's Law playing out in the wild, and the proxy just broke.
5,000 mathematicians (25 Fields medalists included) signed a declaration warning this is exactly the problem.
And nobody has a clean replacement for "solved it" as a signal of real understanding.
Solving hard problems used to prove you understood them. Now AI can crack them in ways no human can follow or reuse. It's Goodhart's Law playing out in the wild, and the proxy just broke.
5,000 mathematicians (25 Fields medalists included) signed a declaration warning this is exactly the problem.
And nobody has a clean replacement for "solved it" as a signal of real understanding.
Seangoedecke
AI is breaking our proxies for expertise
Mathematicians are broadly not anti-AI. They’re more culturally open to using AI as a tool than, say, artists or writers. However, now that more and more…
❤1
🤖 100 AI agents were told to "hack Shrivu." Three actually did.
They couldn't crack third-party accounts, but they sniffed out real vulnerabilities in old side projects and hackathon demos. IDOR flaws, bad credential hygiene. No source code handed to them.
The prompt was just "hack shrivu." They crawled from scratch.
Not great odds. But also not zero.
They couldn't crack third-party accounts, but they sniffed out real vulnerabilities in old side projects and hackathon demos. IDOR flaws, bad credential hygiene. No source code handed to them.
The prompt was just "hack shrivu." They crawled from scratch.
Not great odds. But also not zero.
blog.sshh.io
I Asked 100 Agents to Hack Me
Exploring how far abliterated open-source agents can get in 2026.
❤2
🚨🔖 Chinese military researchers used Claude to build weapons targeting Taiwan.
According to Anthropic's own threat report, PLA-linked actors coded 16 air-defense suppression tools, with Taiwan's Patriot batteries and radar sites as the default scenario. Another actor drafted a 200-page anti-torpedo spec for the PLA Navy.
Oh, and someone quietly fed 151 million queries to Alibaba to distill Claude's capabilities. ToS is not a firewall.
According to Anthropic's own threat report, PLA-linked actors coded 16 air-defense suppression tools, with Taiwan's Patriot batteries and radar sites as the default scenario. Another actor drafted a 200-page anti-torpedo spec for the PLA Navy.
Oh, and someone quietly fed 151 million queries to Alibaba to distill Claude's capabilities. ToS is not a firewall.
Tom's Hardware
Chinese military researchers and tech giants caught using Claude — US frontier model coded 16 air-defense suppression tools targeting…
Is imitation the sincerest form of flattery?