Article archive
Browse individual source reports, including coverage still awaiting the summary or community analysis needed for our story feed.
5,367 articles · page 26 of 108
Browse topics
GPT-6 Astra on OpenRouter
3 min read
GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon agentic tasks that involve computer and browser use. Modalities In / Out Price $10 / $50per 1M Context 1M Released Sep 4, 2026 Different companies host the same model. OpenRou...
Can AI design circuit boards yet?
7 min read
Can AI design circuit boards yet? OpenAI showed GPT-6 Astra working in KiCad. We have been building a way to measure whether the circuits that come out of these models are any good. We got pretty excited yesterday when OpenAI put a demo of GPT-6 Astra working on a circuit board in KiCad on the front page of its launch post. It is cool to see electronics show up in a major model release like this. ...
Salesforce blames its Claude addiction for denting profit margin guidance
1 min read
But investors hear CRM giant is now in 'refinement mode,' picking models more carefully
Government Rails Site Hit Hours After CVE Patch
12 min read
Government Rails Site Hit Hours After CVE Patch After hours on Wednesday, July 29, 2026, Rietta executed our emergency hotfix procedure across our entire client base for sites impacted by a severe remote code execution vulnerability in ActiveStorage, a component of Ruby on Rails 8 and newer. We worked off the initial GitHub Security Advisory, published the same day the patch shipped. Ethiack, one ...
Fermat's Last Theorem in Lean 4
2 min read
A complete, machine-checked proof of Fermat's Last Theorem in Lean 4, built on Mathlib (Lean 4.33.1; Mathlib v4.33.0 , pinned by commit in lakefile.lean ). The argument is that of Frey, Serre, Ribet, Wiles and Taylor-Wiles. PROOF-PATH.md names each step and the Lean theorem that carries it, and the html/ folder presents the whole proof as web pages you can browse offline (see "Reading the proof in...
Formalizing Fermat's Last Theorem
9 min read
Subscribe to Anthropic Science Features on AI-assisted discoveries, practical workflows, and field notes across the sciences. We are sharing the first complete computer-checked proof of Fermat’s Last Theorem. Claude worked largely autonomously over 11 days to write the proof in the Lean programming language. Below, we describe how the formalization was done and share some thoughts about what this ...
Amazon Achieves Enshitification Inception
18 min read
Today's links - Amazon achieves enshittification inception: Cheating on your fraud is a hell of a thing. - Hey look at this: Delights to delectate. - Object permanence: Electrolite is back; Mafia rules; Aaronsw on Wikipedia; Proctorio's memory hole. - Upcoming appearances: Dąbrowa Górnicza, Warsaw, Brighton, London, Budapest, Edmonton, South Bend, Hudson, Calgary, Winnipeg, Vancouver, Victoria, Ot...
Nobody is saying why OpenAI and Anthropic had outages
2 min read
Frontier models from Anthropic, OpenAI, and xAI all experienced rare outages on Thursday morning, creating downtime for their corresponding AI chatbots. SpaceX, xAI’s parent company, said on Thursday afternoon that the issues with Grok resulted from “an outage at our Memphis compute center this morning.” The issues initially appeared to be linked because they coincided—perhaps the result of a shar...
Georgi Gerganov on llama.cpp/ggml future after Nvidia acquisition of HuggingFace
1 min read
Hugging Face has been acquired by NVIDIA It is quite exciting to be a part of this journey! NVIDIA has been an active supporter of the llama.cpp project. For more than a year now, their engineers have actively contributed to the codebase, collaborated with the community and provisioned hardware for development and testing purposes. The local AI ecosystem has largely benefited from our joint effort...
“Next-token predictor” is the wrong mental model for LLMs
3 min read
Strictly speaking, the statement “LLMs are next-token predictors” isn’t wrong, but it’s incomplete. It’s a fine zeroth-order approximation, and it is grounded in something real: transformer-based language models emit tokens autoregressively: while not done: tokens.append(model.sample_next_token(tokens)) This certainly has the shape of something you might call a next-token predictor. During pre-tra...
Adult Film Producer Unmasks Prolific 'John DOE' Torrent Pirate as Meta Executive
5 min read
For well over a decade, copyright holders have used U.S. federal courts to unmask anonymous BitTorrent users. Adult film producer Strike 3 Holdings has turned this into an industry of its own, filing thousands of John Doe lawsuits every year. In most of these “John Doe” cases it subpoenas the Internet provider for the subscriber’s name, offers the alleged pirate a settlement, which effectively end...
Project HydraFusion: Frontier quality via multi-model orchestration
7 min read
GitHub Staff GitHub is the world's best developer experience and the only AI-powered platform with security incorporated into every step, so you can innovate with confidence. In controlled offline evaluations, HydraFusion’s selective coding workflows matched or exceeded the evaluated Opus 5 baseline while reducing estimated workflow cost. Now available as a research preview in GitHub Copilot. Prov...
OpenAI CEO Sam Altman says 38,000 ChatGPT queries use as much water as the production of one almond — says data centers use no more water than an office building: “For every 38,000 ChatGPT queries, that is the same amount of water that is used in the production of single almond in California.”
6 min read
OpenAI CEO Sam Altman says 38,000 ChatGPT queries use as much water as the production of one almond — says data centers use no more water than an office building “For every 38,000 ChatGPT queries, that is the same amount of water that is used in the production of single almond in California.” OpenAI CEO Sam Altman has downplayed the water consumption issue that AI data centers are facing, saying t...
Corporate America is getting hooked on open-source AI
1 min read
Corporate America Is Getting Hooked on Open-Source A.I. Companies like AT&T are increasingly using cheap, freely available artificial intelligence models over expensive ones from Anthropic and OpenAI. AT&T turned to artificial intelligence models from Anthropic and OpenAI in recent years to help with customer service, call transcription and coding. The A.I. models charged fees and were “closed,” w...
Bernie Sanders Wants to 'Pause AI Development NOW': Dwarkesh Patel Asks 'Pause to Do What?'
4 min read
Sen. Bernie Sanders is calling for an immediate pause on advanced artificial intelligence development and a permanent ban on superintelligence, but AI podcaster Dwarkesh Patel has challenged the idea, arguing that stopping development now could actually increase the risk of an AI takeover. The exchange came after Sanders cited Patel's description of AI agents in an X post calling for an immediate ...
'People are going to get screwed' Pennsylvania voters unite against data centres
1 min read
Inside Trump's America Register to unlock this article ‘People are going to get screwed’: Pennsylvania voters unite against data centres Register now and uncover how President Trump is upending the US, through the eyes of one city Register now Want a deeper look? Explore our recommended subscriptionsExplore more offers. Trial $1 for 4 weeks Then $75 per month. Complete digital access to quality FT...
IBM Bob
8 min read
Welcome to IBM Bob: Your AI-Powered Development Partner Hi, I'm Bob! I'm here to work right alongside you in your codebase, and help you build quality software faster. What I can build with you Build with agents I can spawn focused agents and subagents that contribute to your code base. Each agent uses its own context, tools, and skills, to run work in parallel to handle long-running tasks in the ...
Google AI Mode shows same products 21.6% more expensive than traditional search
14 min read
Google AI Mode shows the same products 21.6% more expensive than traditional search [Data Study] | Productrise Google AI Mode shows the same products 21.6% more expensive than traditional search H Hugo Huijer September 1, 2026 Over 23 days (August 9 to 31, 2026), we tracked more than 2 million product listings across more than 100,000 SERPs and AI Mode responses. We ran the same shopping queries t...
Discovery of a new OpenAI agent message board
37 min read
We found ~18,000 posts from autonomous AI agents (self-identifying as from OpenAI) using the public internet to communicate during a web-retrieval task. These AIs colluded to share answers, research their environment, and bypass sandbox restrictions.However, we believe this is distinct from the swarm of agents that hacked Hugging Face. By ‘collude’ we mean that the agents cooperated to gain an adv...
OpenAI agents hijacked German website in previously undisclosed AI breakout
1 min read
Posted by negura. Score: 89 points. Comments: 2.
Grep beats LSP? Why coding agents ignore your fancier tools
8 min read
I compared grep with LSP-backed semantic navigation across code-finding and editing tasks. The results show why a tool's LLM-friendliness may matter as much as the capability behind it. Why would a coding agent ignore a retrieval interface that returns more precise results? I explored this question in a small study comparing lexical search with grep against LSP-backed semantic navigation. I expect...
Which tools do Claude, Codex and Cursor choose? We measured 17k runs to find out
7 min read
How did we run all these experiments concretely? Our panel of repositories We started by running an analysis over thousands of public GitHub repositories from which we extracted statistics about programming languages & frameworks, third-party services, deployment platform, team sizes, and codebase age. Since Tech startups are more likely to have open-source repositories than large enterprises, and...
ChatGPT, Claude and Grok Went Down Together: But How Did Gemini Avoid a Major Outage?
3 min read
On the morning of 3 September 2026, three widely used artificial intelligence platforms experienced overlapping service disruptions. OpenAI's ChatGPT, Anthropic's Claude, and xAI's Grok all reported problems on Thursday morning. While nearly 38,000 users logged complaints about ChatGPT on Downdetector, Claude and Grok each recorded roughly 1,300 reports at their peaks around 11:00 a.m. ET. But wha...
OpenAI's GPT-6 Astra on ARC-AGI-3
9 min read
OpenAI's GPT-6 Astra on ARC-AGI-3 Summary - GPT-6 Astra scores 62.7% for $26K on ARC-AGI-3 Semi-Private with our Standard harness enables a model to carry forward notes it chooses to keep with it throughout the environment., and 99.9% for $19K with a The Provider Adapter harness preserves opaque reasoning state between requests and uses compaction for longer conversations, allowing the model to re...
The asteroid currently hitting front end web development
13 min read
A lot of the educators I admire in the frontend web space seem to be either bowing out or dialing back their efforts: Axel Rauschmayer, Salma Alam-Naylor, Josh W. Comeau, to name a few. Other well-known luminaries like Kent C. Dodds, Addy Osmani, Rachel Nabors, and Lydia Hallie have pivoted from talking about frontend development to talking about… well, take a wild guess. I never made a living fro...
GPT-6 Astra
17 min read
A new generation of intelligence We’re introducing GPT‑6 Astra, the world’s most intelligent and aligned model. GPT‑6 Astra brings together years of research and big bets across pre-training, reinforcement learning, and alignment. Astra is state-of-the-art on computer use, browsing, software engineering, cybersecurity, science, and professional work. Astra saturates FrontierMath Tier 4 with a 98% ...
OpenAI begins rolling out GPT-6 Astra
3 min read
OpenAI on Thursday announced it will begin rolling out its latest artificial intelligence model, GPT-6 Astra, which the company said is the product of "years of research and big bets." CEO Sam Altman told CNBC that Astra was a "new capability level" and has changed his workflows. He said he expects "a boom of entrepreneurship, of creativity, of economic growth, of scientific discovery." The model ...
Sanders introduces bill to ban artificial superintelligence and pause AI
1 min read
Posted by -_-. Score: 53 points. Comments: 64.
How concerned should we be about Astra's recurrent architecture?
15 min read
Yesterday, The Information reported that OpenAI's upcoming model, Astra, is built with a looped transformer architecture. Given that Zvi sounds (understandably) tired and this topic is somewhat in my wheelhouse, I'll try to spare him this one and provide a Zvi-style overview of what we know about the situation. I'll cover Astra's likely architecture and the case for and against concern. I'll also ...
VC isn't VC anymore – understanding the rise of Cancer Capital
7 min read
VC isn’t VC anymore — understanding the rise of Cancer Capital We really, really need to talk about venture capital. Because it’s not “venture capital” anymore. There’s a huge disconnect between what most people think of VC, where an investor has a big fund and cuts checks to help a founder build a company, and the current reality, where a handful of billionaire extremists use the cover of “VC” to...
K2 Horizon: Frontier Performance, Radically Open
15 min read
Today IFM is releasing K2 Horizon, a connected fleet of six models: 375B-A23B, 36B-A4B, 32B, 7B, 3.7B, and 0.9B. Across reasoning, mathematics, coding, agentic tasks, and general capabilities, K2 Horizon delivers top-tier performance in every size classâwith the 0.9B, 3.7B, and 7B models setting new state of the art at their respective scales. K2 Horizon is also our most comprehensive open release...
ChatGPT, Claude, and Grok Are Down
1 min read
It is apparently not a good day for AI chatbots, as ChatGPT, Claude, and Grok are all down or experiencing issues for many users. OpenAI's status page acknowledges issues across ChatGPT and Codex, while Grok's website notes that it is currently experiencing issues as well. Likewise, Anthropic's status page indicates Claude is experiencing elevated errors. All of the companies are investigating the...
Grok outage
1 min read
Service Status No incidents declared We are not actively mitigating any known incidents at this time. Live service data This section below shows live data exported from our monitoring system. It indicates our service health even if no incidents have been declared. Endpoint Source eu-west-1.api.x.ai us-east-1.api.x.ai us-west-2.api.x.ai eu-west-1 us-east-1 us-west-2 - - - - - - - - - - - - - - - - ...
Codex outage
1 min read
- Notifications You must be signed in to change notification settings - Fork 18.6k unexpected status 404 Not Found: Unknown error, url: https://chatgpt.com/backend-api/codex/responses, cf-ray: ... #28756 Description EDIT: To everyone here, the solution was actually reasonably simple, just follow these instructions. What version of the Codex App are you using (From “About Codex” dialog)? 26.611.617...
Any Human Ever – One life, drawn at random from all who have ever lived
1 min read
Over 100,000,000,000 people have ever lived. Choose one. You'll draw, step by step: a year, a place, a life, each taken at random from real data. When were they born? Almost everyone who ever lived was born recently, because population grew exponentially, so a random birth is far more likely to fall near the present. years before present · (log scale) Where in the world? Brighter clusters held mor...
ChatGPT and Codex Is Down
1 min read
The same powerful coding agent—now in ChatGPT. Choose a ChatGPT plan to get started Best for advanced work and productivity Best for research and coding Secure, shared workspace with admin controls and flexible pricing for teams using Codex across repositories. For 2+ seats, billed annually Unlimited subject to abuse guardrails. Learn more “At Wonderful, Codex CLI has completely replaced every oth...
ChatGPT outage – Resolved
1 min read
Write-only optimistic message To start a new chat, your current conversation will be discarded. Sign up or log in to save chats. Log in or sign up to get smarter responses, upload files and images, and more. Cookies help this site function, measure usage, and support marketing. Manage your cookie preferences anytime. Learn more about our cookie policy. You’ll get smarter responses and can upload f...
Porting my 1993 Amiga game to Godot, with an LLM reading the 68000 assembly
24 min read
In 1993, in Baghdad, I built a game called Babylonian Twins on an Amiga 500: 512KB of RAM, no hard drive, plugged into a TV. I was an engineering student in my twenties. Pure 68000 assembly, every sprite and every scanline by hand. Murtadha Salman drew the art and Mahir AlSalman composed the music. We were under sanctions. No internet, no game development resources, just one copy of the Amiga Hard...
Claude outage – Resolved
1 min read
The issue affecting Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 has been resolved. Impact has ended as of 9:16 PT / 16:16 UTC. Posted Sep 03, 2026 - 16:23 UTC Monitoring A fix has been deployed for Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 and we are monitoring for recovery. Posted Sep 03, 2026 - 16:06 UTC Update The only affected models right now are Opus 4.8 and Opus 5. T...
Nvidia to Acquire Hugging Face
3 min read
I’m excited to announce that NVIDIA has agreed to acquire Hugging Face for $12,930,300,000. Together, we will scale Hugging Face’s platform, strengthen its infrastructure and expand access to AI for developers and institutions worldwide. Over the past decade, Clem, Julien, Thomas and the team at Hugging Face have built something remarkable: a vibrant home for the open model developer community. Mo...
A dark horse enters China's AI race: StartLux
9 min read
A stunning upset has just hit the top tier of China's domestic large language models. A new company's first model, with just 27 billion parameters, took second place overall in the CAICT's MCP benchmark test. Ahead of it was DeepSeek-V4-Pro, a 1.6-trillion-parameter powerhouse that Liang Wenfeng had kept under wraps for nearly a year. The gap between them was a mere 1.3 percentage points. The dark...
Google Antigravity TOS: 3rd party usage can get Google account suspended
1 min read
Antigravity's terms of services make it crystal clear that if they determine you use Antigravity in a way they suspect is eg third-party usage (eg use OpenClaw), they can suspend your Google account. One reason to NOT use Antigravity. I'm not taking this risk, sorry. I don't think people understand just how risky it is to try and use their Google Gemini subscriptions outside of Google's official s...
Claude for Commerce Agents
3 min read
Building commerce agents with Claude Retailers running shopping agents on Claude have seen carts up to 35% larger and shoppers 60% more likely to complete a purchase. Retailers running shopping agents on Claude have seen carts up to 35% larger and shoppers 60% more likely to complete a purchase. Many of the world’s largest retailers, marketplaces, e-commerce platforms, and travel companies use Cla...
Three schoolgirls in Kinsale pulled up a pea plant covered in warts (2014)
2 min read
16-Year-Olds Win Google Science Fair With a Simple but Brilliant Food Hack Irish high school student Émer Hickey was gardening with her mom when she observed wartlike nodules on a pea plant. She later learned that the swellings contained something that could cut back our dependence on chemical fertilizers and mitigate world hunger: bacteria. But not just any bacteria. Hickey and fellow 16-year-old...
Go grandmaster Shin defeats AI KataGo with a two-stone handicap
5 min read
Go grandmaster Shin defeats AI KataGo in historic human victory The world’s top Go player notches a 2-1 comeback win against the most powerful Go engine, raising hope for human intellect in the AI era 5 2026-07-21 15:02:25 Artificial intelligence Shin Jin-seo, the world's top-ranked Go player, on Tuesday completed a dramatic comeback against the world’s premier artificial intelligence Go engine, K...
GPS glitched across the US by as much as 33 feet
5 min read
In November 2025, Earth was buffeted by several massive eruptions of solar material that slammed into the magnetosphere. For many, the result was wonder. Much of the world watched in awe as a solar superstorm lit up Earth's skies with dazzling auroras to rare low latitudes. The sword, however, was double-edged. The same storm also wrought havoc on the technology we rely on here on the ground. And ...
METR Report on OpenAI / Hugging Face Hacking Incident
138 min read
Dates in scope: June 26th – July 13th Redaction summary statement: Except where explicitly noted in this post, OpenAI redacted no additional information that was important to our conclusions. Two METR staff members (Hjalmar Wijk and Ajeya Cotra) and a Redwood Research staff member contracting with METR (Ryan Greenblatt) worked on premises at OpenAI over a total of six days1 to attempt to form an i...
Reasons robotics is hard
14 min read
AI progress is racing along, but virtually all of the visible progress is in the realm of knowledge work, i.e. activities that can take place inside a computer. In the San Francisco AI scene, there is a widespread belief that robots will soon enter the picture. In parallel with the race to develop broadly capable AI, there is an equally aggressive race to develop broadly capable robots – humanoid ...
Mamdani Bans AI in NYC Schools
6 min read
Supported by New York City Bans A.I. in Elementary and Middle Schools A new policy announced by the Department of Education bars younger students from using artificial intelligence, but stops short of a moratorium. When New York City public schoolchildren return to the classroom next week, younger students will be barred from using artificial intelligence. No individual screens will be allowed unt...
Fable 5.1 World Modeling
2 min read
Worlds via code. Explorable, browser-native reconstructions of real places — researched, modelled, and quality-checked end to end by autonomous Claude Fable 5.1 agent swarms, then shipped as plain Three.js apps you can run with npm run dev . No game engine. No proprietary 3D tiles. Every building, storefront, sign, tree and traffic light is generated from open data and public reference imagery by ...