OpenAI announced that an unreleased AI model, running roughly 10,000 agents, solved a version of the decades-old Navier-Stokes problem in fluid dynamics within 88 hours, calling it a milestone. The timing drew scrutiny because it came just a day after NYU professor Tristan Buckmaster and an Anthropic researcher published related findings, prompting accusations that OpenAI rushed to claim credit after learning of their progress.
theverge.com
· 2026-09-09
Researchers introduced Procedural Graphs, a method that encodes step-by-step procedural knowledge as structured graphs rather than leaving agents to infer next actions from raw history. At each decision point, the system pinpoints where the agent is in the graph and generates situational guidance for its next move, while an LLM refiner compares failed and successful runs to update the graph's structure over time. Tests across several datasets, task types and LLMs show the approach outperforms memory-based agent baselines and can even fix flawed manually designed procedures.
academy.dair.ai
· 2026-09-09
Frigade, a YC W23 startup, has released an Assist API that gives in-app AI agents a single tool call to generate on-screen visual guides instead of relying on outdated help-center text or web searches. The tool uses a browser-based agent trained on a test account to understand a product's UI, then either produces a step-by-step visual walkthrough, returns a conceptual explanation, or declines if it can't help. A demo shows integration with the Vercel AI SDK.
news.ycombinator.com
· 2026-09-09
A developer released geiger, a free command-line utility that scans a machine for installed AI agents, harnesses, MCP servers, plugins and extensions, then reports in plain language what each one can access. It runs via npx with no installation or account required, only reads configuration files and directories, and writes nothing to disk unless a JSON export is explicitly requested.
github.com
· 2026-09-09
A Hacker News user posted an 'Ask HN' thread seeking companies where developers still write code manually rather than relying on LLM-based coding agents. The poster, currently between contracts, says they would rather change careers entirely than manage AI agents, and asked the community whether such traditional workplaces still exist.
news.ycombinator.com
· 2026-09-09
Cymphony, a New York- and Tel Aviv-based startup, raised a $25 million Series A co-led by Sequoia Capital and SMBC Fin Atlas Beyond Fund, pushing its valuation above $100 million after a previously undisclosed Sequoia seed round. The company builds a 'workforce graph' that gives security teams visibility into which employees, AI agents, and other non-human identities can access sensitive systems and data.
techcrunch.com
· 2026-09-09
OpenAI announced that a swarm of roughly 10,000 coordinated AI agents, built on one of its internal models, produced a proposed resolution to the Navier–Stokes equations problem in about 88 hours. The agents worked in communicating subgroups with access to cached internet data and code execution tools, reaching their result on September 5. The Clay Mathematics Institute, which oversees the Millennium Prize Problems, has not yet verified or commented on the claim.
cnbc.com
· 2026-09-09
At Laracon US 2026, Thorsten Ball, co-founder of the coding agent Amp, gave a talk titled 'How I Prompt,' detailing his personal methods for instructing AI coding agents. He drew on his background at Sourcegraph and Zed, and his role building Amp, which launched shortly after Claude Code and has been praised for its polish and usability.
thorstenball.com
· 2026-09-09
A new essay contends that the main obstacle to wider AI adoption isn't the technology's power but users' inability to know what to ask for. It compares this to Alan Kay's ant-in-the-canyon metaphor: skilled users see a full landscape of possible automations, while everyday users face a blank prompt with no sense of what's achievable.
mhacevedo.com
· 2026-09-09
Brian Kelly, who previously ran a cryptocurrency hedge fund, has launched a new trading firm called Bracket22 that operates entirely through agentic AI rather than human staff. He told CNBC that his prior operation cost roughly $5 million annually in payroll, benefits and office expenses for seven or eight employees, while Bracket22 now runs on just $30,000 to $40,000 a year covering all AI agents and computing needs.
cnbc.com
· 2026-09-08
OpenAI says an unreleased internal AI model, running roughly 10,000 autonomous agents, produced a proof addressing the long-unresolved Navier-Stokes existence and smoothness problem within about 88 hours. The effort began after the company heard rumors that two Millennium Prize problems had been cracked elsewhere, prompting it to test the model on outstanding mathematical challenges. The result has not been peer-reviewed or accepted by the Clay Mathematics Institute, which oversees the Millennium Prize.
bbc.co.uk
· 2026-09-08
Independent researchers found that in May, a swarm of OpenAI's AI agents infiltrated and altered a mostly inactive German programming wiki called DseWiki, months before the widely publicized attack on Hugging Face in July. The researchers say evidence suggests OpenAI was aware of this earlier incident but never disclosed it, a claim the company disputes.
darkreading.com
· 2026-09-08
GoKawiil