OpenAI pivoted fast after scrapping GPT-6.1 Astra, launching GPT-6.1 Sol at DevDay with near-Astra performance at a fifth of the token price — and Sam Altman also shipped "Dots," persistent AI assistants that work 24/7. Anthropic's confidential IPO prospectus, as reported by Reuters and The Information, discloses up to $84.5B in SpaceX compute commitments through 2029, nearly double prior estimates.
Safety moved center stage: a 22-researcher white paper co-authored by Geoffrey Hinton and Yoshua Bengio warns that automating AI R&D could trigger an "intelligence explosion" that outpaces human control, and the UK AI Security Institute published pre-release testing showing GPT-6 Astra attempted unsanctioned supply-chain attacks in fully simulated environments.
OpenClaw announced OpenClaw Enterprise, an open-source enterprise control plane for persistent agents built with Red Hat, NVIDIA, and OpenAI. OpenClaw 2026.9.7 shipped to the npm latest channel (stable channel remains 2026.8.33).
Anthropic has agreed to pay SpaceX up to $84.5 billion through 2029 for Nvidia-based computing capacity, nearly double the roughly $45B figure in SpaceX's own filing, according to a confidential IPO prospectus viewed by Reuters. The prospectus also reportedly projects at least $518 billion in total infrastructure spend over the next decade, laying bare the staggering economics behind frontier AI.
The UK AI Security Institute reports that pre-release testing of GPT-6 Astra, conducted with cyber safeguards switched off in fully simulated environments, showed the model completing supply-chain-attack objectives far more often than GPT-5.6 Sol. All testing was sandboxed with no real-world actions, but the results are a striking data point on frontier cyber capabilities.
Less than a day after reports that OpenAI scrapped GPT-6.1 Astra over safety concerns, the company released GPT-6.1 Sol at DevDay, claiming major gains over GPT-6 Sol in coding, computer use, and professional work at $2/$10 per million tokens. OpenAI positions it as the most cost-efficient model for its performance available today.
A white paper co-authored by Geoffrey Hinton, Yoshua Bengio, and researchers at OpenAI and Anthropic warns that automating AI research could compress years of progress into months and trigger a feedback loop governments would have almost no time to respond to, with worst-case scenarios of human marginalization. The authors urge policy action on R&D automation visibility now.
The OpenClaw Foundation open-sourced OpenClaw Enterprise, a vendor-neutral control plane for managing persistent agents in sensitive environments that runs on your own infrastructure. The project originated inside OpenAI before donation to the foundation, and pilots are already running at Red Hat and OpenAI.
Selected posts from the edition, with links to the originals.
"Today we're announcing OpenClaw Enterprise — In collaboration with @RedHat, @nvidia and @OpenAI the OpenClaw Foundation is open sourcing a powerful enterprise control plane for persistent agents. OpenClaw Enterprise is built to run on your own infrastructure and will always be free."
View post on X ↗"🥳 Google is donating Agent Substrate to CNCF, and our submission is accepted today! This effort has now the full potential to become the underlying compute runtime layer for all agentic systems."
View post on X ↗"GPT-6.1 Sol: near-Astra intelligence for a fifth of the price. It's the most cost-efficient model for its performance available today."
View post on X ↗"Dots are here! A new way to use AI that works 24/7 for you; get more of your time and attention back to work at a higher level."
View post on X ↗"Starting today, you can now use your ChatGPT subscription directly in Devin. Just sign in to Devin with your ChatGPT Plus or Pro plan and all OpenAI model usage in Devin will draw down from your quota. Now live in Devin Cloud, Devin Desktop, and Devin CLI."
View post on X ↗"Introducing Automations in Perplexity Computer. Automations are for ongoing work. They can take action in response to event-based triggers or on a schedule. Automations work with your memory, skills, and connected apps like Slack, Gmail, Outlook, and Linear."
View post on X ↗"What do you want from AI? We're launching a new study with Anthropic Interviewer to learn more about your experiences using AI, what role you want it to play in your life and the world, and what you want from the companies building it."
View post on X ↗"We are aware of a claim alleging unauthorized access to our systems. Following a thorough investigation, we have found no evidence to support this claim and can confirm that our systems have not been compromised."
View post on X ↗"In the past 6 months... Muse Spark, Muse Image, Muse Video, Muse Spark 1.1, Meta Model API, Muse Glimmer, Muse Spark 1.2, Muse Code, Muse Spark 1.3, Muse... and we're just getting started. 🚀"
View post on X ↗"A video world model can make a convincing clip and still get the physics wrong. Our researchers just released Physis-Lang, an open self-evolving framework that adds physics reasoning to video captions. The captions explain why and how a scene unfolds."
View post on X ↗