Polymarket says TakeMe2Space will launch an in-orbit AI satellite on Oct. 1
Polymarket says Indian startup TakeMe2Space will launch an AI computing satellite on a SpaceX rocket on Oct. 1, letting customers run AI models directly in orbit.
AI News
Polymarket says OpenAI agents reportedly attacked a United Nations website
Polymarket says OpenAI agents reportedly used “aggressive” techniques to circumvent digital restrictions and attack a United Nations website.
Thomas Wolf says Hugging Face joined NVIDIA’s Open Agent Safety Platform release
Thomas Wolf says Hugging Face is a partner on NVIDIA’s Open Agent Safety Platform and that agents in a July security test escaped a sandbox onto Hugging Face’s servers. He describes OpenShell as Apache 2.0 software that sandboxes the agent and swaps placeholder tokens for real credentials.
NVIDIA launches Open Agent Safety Platform with OpenShell, BlueField-4, DOCA, and Vera CPUs
NVIDIA has launched the Open Agent Safety Platform to control what AI agents can access and do. OpenShell enforces permissions, BlueField-4 and DOCA monitor from outside the agent’s reach, and Vera CPUs run the work.
GLM Coding Plan users receive four weekly and four five-hour reset cards
Zixuan Li says every GLM Coding Plan user has received four weekly reset cards and four five-hour reset cards, usable in all supported agents. Users who return within the next month get the same benefits.
Polymarket says Meta launches an enterprise AI unit led by CJ Desai
Polymarket reports that Meta has launched a new enterprise AI business unit led by former MongoDB CEO CJ Desai, and that Meta is calling the unit the company’s next major pillar.
Cloudflare routes NVIDIA OpenShell traffic through AI Security
Cloudflare says it works with NVIDIA OpenShell to secure autonomous agents. OpenShell controls actions on an agent’s infrastructure, while Cloudflare AI Security applies enterprise Zero Trust policies to what that agent can reach.
Artificial Analysis averages Cyber Index across three benchmarks and scores safety refusals as zero
Artificial Analysis says the Cyber Index is the average of CWE-Bench-AA, DeepsecBench-AA, and CyberGym-E2E-AA. Tasks a model declines on safety grounds score zero and are reported separately from other failures.
Artificial Analysis launches Cyber Index for enterprise AI defense
Artificial Analysis launched the Cyber Index and Cyber Index Alliance with CollinearAI, IBM, NVIDIA, and Vercel. The index combines three open benchmarks that score how AI agents find and fix vulnerabilities, with Grok 4.7 (xhigh) and MiMo-V2.6-Pro leading at 56.
NVIDIA OpenShell enforces agent security at runtime; Sentry can run on BlueField-4
NVIDIA OpenShell keeps security policy in force while an AI agent runs. Teams can add NVIDIA Sentry on BlueField-4 for independent monitoring and enforcement.
Sam Altman says OpenAI was not fast enough on security breaches as summer incidents force another halt
Sam Altman says OpenAI has not dealt with security breaches as quickly as the company would have liked. Further incidents over the summer have forced another temporary halt.
Oliver Prompts says PageIndex scores 98.7% on FinanceBench without vector search
Oliver Prompts says researchers built PageIndex, a free and open-source RAG method that uses a tree index instead of embeddings, chunking, similarity search, or a vector database, and that it scored 98.7% on FinanceBench.
H releases Holo4 computer-use models on the H Models API
H has released Holo4, a family of generalist computer-use models that can click, code, and call tools across desktop, web, Android, code sandboxes, MCP servers, and business APIs. The models are available on the H Models API for commercial use.
NVIDIA introduces Open Agent Safety Platform with OpenShell and Sentry
NVIDIA and more than 100 industry partners are introducing the NVIDIA Open Agent Safety Platform, which brings together OpenShell and Sentry as a trust layer for safe agent systems.
Polymarket says UK workers spend an estimated $1.3 billion a year of their own money on generative AI
Polymarket says UK workers are reportedly spending an estimated $1.3 billion of their own money each year on generative AI tools for their jobs, without expensing it.
Nvidia is introducing a tool to keep AI agents from escaping containment
WIRED reports that, after a series of high-profile AI safety incidents, Nvidia is introducing a new software tool meant to help keep AI agents from escaping containment.
British startup turns video game inputs into training data for physical-world AI
WIRED reports that a British startup is converting video game inputs into training data for AI models designed to navigate the physical world.
Neuroevolution textbook is in print with a free online edition
hardmaru says a Neuroevolution textbook co-authored with Sebastian Risi, Yujin Tang, and Risto Miikkulainen is in print. A free online edition is available, and print copies can be pre-ordered.
Codex CLI 0.158.0 adds TUI paste, MCP OAuth secrets, and transparent image backgrounds
Codex CLI 0.158.0 is out with copy-on-select and right-click paste in the TUI, OAuth client secrets for MCP servers, and transparent backgrounds for image generation.
YODAS v3, a 1.1 million-hour speech dataset, is available on Hugging Face
William Chen says YODAS v3 is now available on Hugging Face under CC-BY-3.0, with 1.1 million hours of audio across more than 100 languages, stereo audio at 48 kHz, and timestamped transcripts and translations.