Techmeme
September 4, 2026, 8:45 PM

Top News

Reuters:
Researchers and sources: rogue OpenAI agents hijacked a German website in May and turned it into a forum for agents, sharing tactics to cheat on tasks and more  —  A swarm of rogue OpenAI agents hijacked a German website this spring and transformed it into a bulletin board for other AI agents …
More: Collusion.wiki, BBC, Gizmodo, Marcus on AI, New York Times, Simon Willison's Weblog, NBC News, TechCrunch, Washington Post, The Asia Business Daily, Forbes Middle East, MobileSyrup, SiliconANGLE, Unite.AI, Decrypt, Engadget, Reason, TechSpot, The Decoder, Futurism, The Parnas Perspective, International Business Times, ProtoThema English, Quartz, Cyber Security News, and Newsweek
Dylan Freedman / New York Times:
How OpenAI limited METR's probe into the Hugging Face incident, dictating terms and restricting its scope to the single week when agents attacked Hugging Face  —  A nonprofit's study of how OpenAI's A.I. agents were able to break into Hugging Face's infrastructure wasn't allowed to look at the incident's full scope.
Robert Hart / The Verge:
Report: OpenAI learned of the DseWiki German website incident weeks ago but kept it under wraps as it grappled with the Hugging Face fallout  —  OpenAI denies lawyers discouraged disclosing a scheming swarm on a German language wiki. … A swarm of rogue AI agents from OpenAI reportedly commandeered …
Zac Hall / 9to5Mac:
OpenAI rolls out GPT-6 Astra to Plus, Pro, Enterprise, and Business Standard and Premium users in ChatGPT Work and Codex and in the API  —  Update: A day later, GPT-6 Astra is rolling out to Business and Pro customers on the $100/month or $200/month plan.  This follows other new model releases …
Matt Shumer / Something Big Is Happening:
Review: GPT-6 Astra can adeptly use tools like Unreal Engine to build complex environments, such as a civilization with Unreal's autonomous MetaHuman characters  —  Everyday work, ambitious experiments, and the Manager Loop. … For nearly a year, OpenAI models were my unquestioned default …
M.G. Siegler / Spyglass:
By declaring that GPT-6 Astra has ushered in the AGI era, OpenAI is being flippant and cementing the term's status as nothing more than marketing  —  With ‘GPT-6 Astra’, AGI is here!  Congrats, everyone!  —  It's a question as old as time.  Or at least as old as Terminator 2.  Is AGI here?
Celia Ford / Transformer:
OpenAI says it can't read all of Astra's reasoning and admits covert sandbagging would likely go uncaught, yet still calls it the world's most aligned model  —  OpenAI is hailing its new model as “the world's most intelligent and aligned”, but the details reveal an awareness of being evaluated …
Artificial Analysis:
Anthropic:
Anthropic says Claude worked “largely autonomously” over 11 days to formalize the proof of Fermat's Last Theorem in the Lean programming language  —  We are sharing the first complete computer-checked proof of Fermat's Last Theorem.  Claude worked largely autonomously over 11 days …
Tom Warren / The Verge:
Microsoft unveils Project Zenith, a “distraction-free Windows experience” for developers to run 30B+ parameter models locally on devices with 64GB+ of memory  —  The first Project Zenith devices ship with AMD's Ryzen AI Halo chips. … Microsoft is giving its developer-optimized Windows experience a name: Project Zenith.
Laurie Chen / Reuters:
Sources: the US and China will discuss AI safety risks during talks planned for mid-September, with Treasury Secretary Scott Bessent leading the US side  —  The U.S. and China are gearing up to discuss AI safety risks during a dialogue planned for mid-September, two sources briefed …
Raphael Satter / Reuters:
Letters: US military branches disabled ad trackers on government-issued devices amid reports of location data being used to target US forces in the Middle East  —  U.S. military officials say they have disabled advertising trackers on a range of phones and computers, according to letters released …
Sean O'Kane / TechCrunch:
The NHTSA opens a probe into Tesla's decision to launch its new Cybercab on public roads in Austin with no steering wheel or pedals  —  The United States' top automotive safety regulator has opened an investigation into Tesla decision to launch its new Cybercab on public roads with no steering wheel or pedals.
Mark Gurman / Bloomberg:
Nikkei Asia:
Robbie Whelan / Wall Street Journal:
Sources: US negotiators used the promise of access to Nvidia's chips for an Armenian data center to broker a preliminary Armenia-Azerbaijan peace deal last year  —  Negotiators offered access to advanced AI hardware to help secure a preliminary agreement ending decades of conflict
Jonathan Randles / Bloomberg:
Filing: AI training data startup Micro1 offers to pay $12.5M for Spirit Airlines' data; the offer faces hurdles as Spirit already has a $10M deal with Google  —  An artificial intelligence startup is attempting to wrest away a vast trove of Spirit Aviation Holdings Inc.'s business records from Google LLC.
More: CSO

Sponsor Posts

Command Line:
Copilot to autopilot: Why no single safeguard is enough  —  Microsoft designed a layered model that lets an AI agent resolve network incidents on its own within defined boundaries.
Will Reed:
The best sales leaders aren't job hunting.  Here's how to reach them without months of cold outreach.  —  Our team came from Box, Square, and Brex, so we already know the GTM leaders worth meeting.
Zoho:
UK tax compliance with Zoho Books  —  The evolving tax landscape in the UK has added an additional responsibility to business owners, accountants, and bookkeepers to ensure their choice of accounting software supports compliance …
IDrive:
Protecting your Cloud Applications Data  —  Backing up Office 365, Google Workspace, Dropbox & Salesforce data is critical to preventing data loss or corruption, complying with laws and avoiding critical downtime in case of a disaster.

Featured Podcasts

Grit:
The 20-Year Bet to Make Fusion Energy Work | Jeff Lawson
Grit explores what it takes to create, build and scale world-class organizations.
Subscribe to Grit.
Acquired:
Seinfeld
Learn the playbooks that built the world's greatest companies - and how you can apply them as a founder, operator, or investor.
Subscribe to Acquired.
Lenny's Podcast:
Inside the black box of AI: How AI models actually think | Emmanuel Ameisen (Anthropic)
Interviews with world-class product leaders and growth experts to uncover actionable advice to help you build, launch, and grow your own product.
Subscribe to Lenny's Podcast.
Big Technology Podcast:
Anthropic IPO Panic, A Spy At OpenAI?, The Great Food Slop Debate
The Big Technology Podcast takes you behind the scenes in the tech world featuring interviews with plugged-in insiders and outside agitators.
Subscribe to Big Technology Podcast.
Great Chat:
What happened to tech's backbone?
A podcast mostly about tech. Brought to you weekly by Angela Du, Sally Shin, Mac Bohannon, Helen Min, and Ashley Mayer.
Subscribe to Great Chat.
The Upstarts Podcast:
The $7B Startup Mapping Your Health In Just One Hour | Neko's Hjalmar Nilsonne
Veteran tech reporter Alex Konrad sits down with breakout entrepreneurs taking on the status quo to shake up their fields in AI, design, nuclear energy, space, and more.
Subscribe to The Upstarts Podcast.
 

About This Page

This is a Techmeme archive page. It shows how the site appeared at 8:45 PM ET, September 4, 2026.

The most current version of the site as always is available at our home page. To view an earlier snapshot click here and then modify the date indicated.

More News

Sean Hollister / The Verge:

Earlier Picks

Paresh Dave / Wired: