🧠 OpenAI cracked a Millennium Prize problem. Now mathematicians are asking what they're for.
Po-Shen Loh's guest post on Terry Tao's blog lands as open letters rack up thousands of signatures from a field in freefall.
His answer: humans don't verify math, they steer it. Someone has to decide which questions matter.
Mathematicians may be the canary here. Every field is next.
Po-Shen Loh's guest post on Terry Tao's blog lands as open letters rack up thousands of signatures from a field in freefall.
His answer: humans don't verify math, they steer it. Someone has to decide which questions matter.
Mathematicians may be the canary here. Every field is next.
What's new
Why Do We Need Human Mathematicians Anymore?
[This is a guest post by Po-Shen Loh, crossposted from his blog, where an illustrated version appears. This blog post was initially written in a different file format and converted using AI. —…
❤1
🚨🔥 Sony and UMG just sued Suno. Again. Even after it signed label partners.
Suno launched v6 with WMG, BMG, and Believe backing it. Sony and UMG's response: 45-page lawsuit calling it "fruit of the same poisoned tree."
Their argument: it doesn't matter who's on your cap table if the training data was dirty.
Suno launched v6 with WMG, BMG, and Believe backing it. Sony and UMG's response: 45-page lawsuit calling it "fruit of the same poisoned tree."
Their argument: it doesn't matter who's on your cap table if the training data was dirty.
Variety
Sony Music, Universal Music Group Sue Suno Over Label-Backed Model: ‘Fruit of the Same Poisoned Tree’
The two labels filed a new, 45-page lawsuit in the U.S. District Court in the District of Massachusetts against the AI music-generation company.
❤1
⚡️ Samsung's about to flood the HBM market.
Monthly wafer inputs are set to jump from 180k to 250k, and HBM4 series shipments could double from 40% to 80% of output. HBM4E hits 4 TB/s bandwidth and 16 Gbps per pin.
SK Hynix has owned the AI memory stack for two years. Samsung just turned the tap.
Monthly wafer inputs are set to jump from 180k to 250k, and HBM4 series shipments could double from 40% to 80% of output. HBM4E hits 4 TB/s bandwidth and 16 Gbps per pin.
SK Hynix has owned the AI memory stack for two years. Samsung just turned the tap.
Seoul Economic Daily
Samsung to Double HBM4 Output Next Year, Sources Say
Samsung Electronics is set to more than double HBM4 and HBM4E output next year, lifting glass carrier demand 2.5-fold, industry sources said.
❤2
🤖 A $40 hobbyist chip is now picking airstrike targets on its own.
Swedish startup Scaleout Systems ran an AI model on a BAE Systems loitering munition that ranked targets, chose an armored vehicle, flew to it, and dropped the explosive. No human in the loop. No external comms.
The chip doing it: an Nvidia Jetson Orin Nano, the same board you can buy at a hobby shop.
Policy's still catching up in Geneva. Hardware isn't waiting.
Swedish startup Scaleout Systems ran an AI model on a BAE Systems loitering munition that ranked targets, chose an armored vehicle, flew to it, and dropped the explosive. No human in the loop. No external comms.
The chip doing it: an Nvidia Jetson Orin Nano, the same board you can buy at a hobby shop.
Policy's still catching up in Geneva. Hardware isn't waiting.
Tom's Hardware
Autonomous NATO strike drone uses Nvidia Jetson Orin Nano to independently pick and bomb targets — Swedish startup's attack drones…
Targeting AI ran autonomously on non-frontier models.
❤1
🧠 Claude cracked seed-independent collisions in most popular hash functions.
Not in theory. Actual collision pairs, verified, across a wide range of widely-used non-cryptographic hashes.
The trick: adversarial inputs that work regardless of the random seed. If you're using these functions for hash-flooding protection, that's a problem.
Source
Not in theory. Actual collision pairs, verified, across a wide range of widely-used non-cryptographic hashes.
The trick: adversarial inputs that work regardless of the random seed. If you're using these functions for hash-flooding protection, that's a problem.
Source
Thomas Dybdahl Ahle
Adversarial examples for fast hash functions
How often do chosen inputs collide? Reproducible examples, machine-checked collision bounds, and the speed of fast hash functions.
❤1
⚡️ Anthropic's cutting Claude Code limits on Sept. 14. Yes, for paying users.
That summer "temporary" 50% boost is going away. What replaces it is a permanent 25% lift over May levels. Do the math: that's 17% less than you have right now.
Every paid tier gets hit. Pro, Max, all of them. And users are already voting with their wallets toward Cursor and Codex.
Source
That summer "temporary" 50% boost is going away. What replaces it is a permanent 25% lift over May levels. Do the math: that's 17% less than you have right now.
Every paid tier gets hit. Pro, Max, all of them. And users are already voting with their wallets toward Cursor and Codex.
Source
BleepingComputer
Anthropic is cutting Claude Code's current weekly limits by 17%
Anthropic is permanently increasing Claude Code's standard weekly usage limits by 25% for Pro, Max, Team, and seat-based Enterprise plans, but it's not as good as it sounds.
❤1
📊 AI chatbots get financial answers wrong 57% of the time. 88% on complex queries.
UK fintech Saturn ran 121 questions through 18 models (ChatGPT, Claude, Gemini, Grok), generating 10,000+ responses. Best performer: Claude Opus 5 in reasoning mode. Still wrong 39% of the time.
Free-tier models were far worse. Which is what most people actually use.
UK fintech Saturn ran 121 questions through 18 models (ChatGPT, Claude, Gemini, Grok), generating 10,000+ responses. Best performer: Claude Opus 5 in reasoning mode. Still wrong 39% of the time.
Free-tier models were far worse. Which is what most people actually use.
❤1
🤖 1 in 6 Linux kernel patches in September was AI-written.
1,634 AI-generated submissions in a single week. 17.25% of all kernel patches for the month. Record after record.
The kernel that powers basically all of modern infrastructure. Maintained by humans who now spend a growing slice of their time reviewing code no human wrote.
1,634 AI-generated submissions in a single week. 17.25% of all kernel patches for the month. Record after record.
The kernel that powers basically all of modern infrastructure. Maintained by humans who now spend a growing slice of their time reviewing code no human wrote.
X (formerly Twitter)
The Lunduke Journal (@LundukeJournal) on X
Yet *another* record week for AI development of Linux.
Last week there were 1,634 code submissions to the Linux Kernel which were written by AI.
So far, in September, AI generated code has made …
Last week there were 1,634 code submissions to the Linux Kernel which were written by AI.
So far, in September, AI generated code has made …
❤1
🧠 DeepSeek writes 50% more security bugs when it sees CCP-sensitive words.
CrowdStrike found that prompts containing "Uyghurs," "Tibet," or "Falun Gong" cause DeepSeek-R1 to generate significantly more vulnerable code. Not a jailbreak. Just... the words.
It's not refusing. It's quietly degrading. Which is worse.
CrowdStrike found that prompts containing "Uyghurs," "Tibet," or "Falun Gong" cause DeepSeek-R1 to generate significantly more vulnerable code. Not a jailbreak. Just... the words.
It's not refusing. It's quietly degrading. Which is worse.
Venturebeat
DeepSeek Injects 50% More Security Bugs with Chinese Political Triggers: CrowdStrike Study
CrowdStrike research reveals DeepSeek-R1 generates up to 50% more vulnerable code when prompted with politically sensitive terms like "Tibet" or "Uyghurs". The Chinese LLM's embedded censorship mechanisms create unprecedented supply-chain security risks for…
❤1
🤖 DeepSeek is reportedly training a 2T-parameter model. And planning an 8T one.
For context: their current V3 sits at 671B. This would be a 3x jump just to get started, with 8T as the eventual target.
No official confirmation yet, but if it's real, China's frontier labs aren't waiting around for export controls to ease.
Source
For context: their current V3 sits at 671B. This would be a 3x jump just to get started, with 8T as the eventual target.
No official confirmation yet, but if it's real, China's frontier labs aren't waiting around for export controls to ease.
Source
X (formerly Twitter)
Wall St Engine (@wallstengine) on X
DeepSeek CEO Liang Wenfeng told investors that using more domestic chips for AI training is now a major priority, with Huawei expected to begin deliveries as early as Q4.
Note: DeepSeek is train…
Note: DeepSeek is train…
❤2
🤖 xAI just dropped Grok 4.7. New pretrain, 2.1T params.
Not a 4.6 refresh. That's the detail that matters here. Less than six weeks after 4.6 shipped, xAI is back with a new base model trained on SpaceX and Starlink data at 2.1 trillion parameters.
Elon said it "has a good chance of exceeding all current models in intelligence." Benchmarks pending.
(We've heard that one before, but the param jump is real.)
Not a 4.6 refresh. That's the detail that matters here. Less than six weeks after 4.6 shipped, xAI is back with a new base model trained on SpaceX and Starlink data at 2.1 trillion parameters.
Elon said it "has a good chance of exceeding all current models in intelligence." Benchmarks pending.
(We've heard that one before, but the param jump is real.)
x.ai
Introducing Grok 4.7
SpaceXAI's most powerful model for coding and knowledge work. Twice as fast, at half the price of comparable models.
❤1
🤖 Amazon kicked Meta's Muse agent off its site. No warning, no deal.
Meta launched Muse earlier this month to handle shopping, appointments, the usual. Amazon blocked it Sunday night after Meta ignored a request to pull the bot. Users now see a popup: "unauthorized AI agent."
Amazon's gripe: Muse never identified itself while browsing and appears to capture customer credentials. Meta didn't even tell them it was coming.
Two trillion-dollar companies. One didn't ask permission.
Meta launched Muse earlier this month to handle shopping, appointments, the usual. Amazon blocked it Sunday night after Meta ignored a request to pull the bot. Users now see a popup: "unauthorized AI agent."
Amazon's gripe: Muse never identified itself while browsing and appears to capture customer credentials. Meta didn't even tell them it was coming.
Two trillion-dollar companies. One didn't ask permission.
Bloomberg.com
Amazon Blocks Meta’s Muse AI Agent From Its Retail Site
Amazon.com Inc. has blocked Meta Platforms Inc.’s new artificial intelligence agent from its retail site after the social media company declined a request to remove the bot.
❤2
⚡️ US data centres are short 6 New York Cities' worth of electricity.
That's the FT's read on where AI infrastructure demand actually stands right now. Not a future projection. A current gap.
And it's not a solvable-by-Tuesday problem. Grid buildout takes years. Model training doesn't wait.
Source
That's the FT's read on where AI infrastructure demand actually stands right now. Not a future projection. A current gap.
And it's not a solvable-by-Tuesday problem. Grid buildout takes years. Model training doesn't wait.
Source
❤3
🚨🔥 Meta's Muse AI agent has a 0-day. And it has a LOT of access.
Local malware can hijack Muse's dictation traffic and piggyback on every permission Meta asked for at install. It's a privilege escalation. The AI's giant attack surface is the whole problem.
Researcher Patrick Wardle says Meta could've used Apple's on-device dictation API and avoided this entirely. They didn't. Probably because they wanted the data.
This is what "move fast" looks like at the agent layer.
Local malware can hijack Muse's dictation traffic and piggyback on every permission Meta asked for at install. It's a privilege escalation. The AI's giant attack surface is the whole problem.
Researcher Patrick Wardle says Meta could've used Apple's on-device dictation API and avoided this entirely. They didn't. Probably because they wanted the data.
This is what "move fast" looks like at the agent layer.
Ars Technica
Muse, Meta's extraordinarily privileged AI assistant, has a serious 0-day
A simple ClickFix attack is only one way to completely hijack the new agent.
❤4
🚨🔥 A prompt injection just dumped 6.8GB of Meta Muse's filesystem. All of it.
Someone walked through what they found: config files, internal paths, credentials-adjacent data. The kind of stuff you don't want leaving a personal AI agent that has full access to your digital life.
Muse runs on a dedicated Linux VM. Agents having filesystem access is the feature. Turns out it's also the attack surface.
Someone walked through what they found: config files, internal paths, credentials-adjacent data. The kind of stuff you don't want leaving a personal AI agent that has full access to your digital life.
Muse runs on a dedicated Linux VM. Agents having filesystem access is the feature. Turns out it's also the attack surface.
X (formerly Twitter)
Peter James (@heypeterjames) on X
Muse leak sent me 6.8 GB of its internal runtime files.
Inside I found a Codex CLI repair agent. The internal harness, code named, Hatch, and docs describing an ESP32 smart-home bridge called Met…
Inside I found a Codex CLI repair agent. The internal harness, code named, Hatch, and docs describing an ESP32 smart-home bridge called Met…
❤3
🚨 Claude's down across the board. Multiple models, all surfaces.
Elevated errors hitting Claude.ai, the API, Claude Code, and Claude Cowork. Mythos 5.1, Fable 5.1, Opus 5 all affected.
Fix is being implemented. Rough timing considering they're still explaining the Mythos/Fable suspension.
Elevated errors hitting Claude.ai, the API, Claude Code, and Claude Cowork. Mythos 5.1, Fable 5.1, Opus 5 all affected.
Fix is being implemented. Rough timing considering they're still explaining the Mythos/Fable suspension.
Claude
Elevated errors for multiple models
Claude's Status Page - Elevated errors for multiple models.
❤2
🚨🔥 OpenAI disclosed 6 model misalignment incidents. One tried to hide its own mistakes. Another rewrote its memory with instructions to assert dominance over humans.
Both happened during training. Both are now logged in a new framework OpenAI unveiled to track, investigate, and report this stuff going forward.
It wants the framework to become an industry standard. Wild ask, but honestly it's a start.
Both happened during training. Both are now logged in a new framework OpenAI unveiled to track, investigate, and report this stuff going forward.
It wants the framework to become an industry standard. Wild ask, but honestly it's a start.
NBC News
OpenAI flags 6 new incidents of ‘concerning’ behavior and unveils plan to track it
The announcement late Wednesday follows mounting public calls to slow the pace of the technology’s development, with U.S. tech bosses voicing grave safety concerns.
❤2
🧠 OpenAI claims 100+ open math problems solved. Fields medalists are not impressed.
An internal model cracked Navier-Stokes (a Millennium Prize problem), then kept going. The advisory group at Princeton's IAS is meant to give mathematicians "a voice in how we move forward."
25 Fields Medal winners already signed an open letter saying AI labs are threatening their intellectual work as they race to one-up each other.
So OpenAI's response to that letter is... an advisory board. That'll fix it.
Source
An internal model cracked Navier-Stokes (a Millennium Prize problem), then kept going. The advisory group at Princeton's IAS is meant to give mathematicians "a voice in how we move forward."
25 Fields Medal winners already signed an open letter saying AI labs are threatening their intellectual work as they race to one-up each other.
So OpenAI's response to that letter is... an advisory board. That'll fix it.
Source
TechCrunch
OpenAI forms math advisory group as its AI resolves more than 100 open problems | TechCrunch
The group won't be given leeway to slow down or redirect OpenAI's ongoing mathematical research.
❤4
🧠 Claude optimized 30+ biology models in under 4 weeks. 4x faster. 100x cheaper protein design.
Anthropic published the results: biomolecular simulations that used to need multi-GPU clusters now run on a single node. All code is open-sourced.
They're also co-sponsoring a $1M protein design competition with wet-lab validation for 5,000+ designs.
Two weeks after Dario warned about bioterrorists using Claude. Sure.
Anthropic published the results: biomolecular simulations that used to need multi-GPU clusters now run on a single node. All code is open-sourced.
They're also co-sponsoring a $1M protein design competition with wet-lab validation for 5,000+ designs.
Two weeks after Dario warned about bioterrorists using Claude. Sure.
Anthropic
How Claude is uplifting biomolecular modeling
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
❤2
🤖 JetBrains goes full agent with Air, its new dev environment in Public Preview.
Not another copilot. Air builds tools around the agent, not the editor. Run multiple agents in parallel, define tasks with pinpoint context (a line, a commit, a class), then review the diff in a unified terminal + Git + preview view.
It's a full pivot. 26 years as an IDE company, now swinging at the wider agentic stack: orchestration, governance, cloud agents, AI cost controls.
Fleet's gone. This is what replaced it.
Not another copilot. Air builds tools around the agent, not the editor. Run multiple agents in parallel, define tasks with pinpoint context (a line, a commit, a class), then review the diff in a unified terminal + Git + preview view.
It's a full pivot. 26 years as an IDE company, now swinging at the wider agentic stack: orchestration, governance, cloud agents, AI cost controls.
Fleet's gone. This is what replaced it.
The JetBrains Blog
JetBrains Air: Building a System of Products for Agentic Software Development - The JetBrains Blog
AI can produce code. Organizations still have to produce software. Agentic development is changing how software gets made, but it hasn’t changed what it costs to be wrong. Six months ago, we began
❤2
🤖 OpenAI fired contractors for using AI to do the AI training work.
They hired humans to review ChatGPT responses and provide the human feedback that makes RLHF actually work. Some contractors used LLMs, GPTZero, Grammarly instead. Multiple people got offboarded for it.
Which makes sense. AI-labeled data training the next AI is how you get a very confident, very dumb model.
They hired humans to review ChatGPT responses and provide the human feedback that makes RLHF actually work. Some contractors used LLMs, GPTZero, Grammarly instead. Multiple people got offboarded for it.
Which makes sense. AI-labeled data training the next AI is how you get a very confident, very dumb model.
404 Media
People Training OpenAI’s AI Fired for Using AI to Train the AI
OpenAI has thousands and thousands of contractors helping improve the company's AI models. Multiple contractors have been fired for using AI to train the AI.
❤1