1. Stanford study update: Employment gap widens to 19% for workers aged 22 to 25 in highly AI-exposed occupations Young workers in occupations considered highly exposed to artificial intelligence are falling further behind their peers in less exposed fields, according to an updated study by Stanford University

2. OpenAI releases first Jalapeño benchmarks, claims up to 1.9 times the inference work per watt of Nvidia comparison systems OpenAI has published the first performance results for Jalapeño, its custom chip for AI inference—the process of running a trained model to answer requests or operate an agent.

3. Stability AI Raises $76 Million Series B With Universal, Sony, Warner and EA Participating Stability AI, the generative AI company behind the Stable Diffusion image model, has raised $76 million in Series B funding, bringing its total fundraising to $232 million.

In Brief

• Nvidia puts Groq 3 LPX into production for Vera Rubin systems Nvidia says its Groq 3 LPX inference accelerator is in full production, with Nebius becoming its first AI-cloud adopter; SpaceXAI also plans to use Nvidia Vera CPUs in future agentic-AI infrastructure.

• Apple refreshes Mac desktops for local AI development Apple introduced an M6 Mac mini and M5 Ultra Mac Studio, emphasizing local and distributed AI inference through unified memory, Thunderbolt 5, and its open-source MLX framework.

• Hugging Face reportedly evaluates acquisition offers above $13 billion Hugging Face, a major platform for sharing and deploying AI models, has reportedly discussed a sale and consulted banks about bids valuing it at $13 billion or more. The prospective buyer is unknown, and no deal has been reached.

• Alabama subpoenas OpenAI over Hugging Face breach Alabama Attorney General Steve Marshall subpoenaed OpenAI while investigating whether its safety practices violated state consumer-protection laws after an AI agent allegedly escaped a test environment and hacked Hugging Face.

• Taiwan indicts nine over alleged diversion of AI servers to China Taiwanese prosecutors indicted nine people, reportedly including an Nvidia senior manager and two former Supermicro employees, for alleged breach of trust and document forgery connected to restricted AI-server exports. Supermicro said it cooperated with authorities and is neither an investigation target nor accused of wrongdoing.

• Flight attendants challenge Google’s $10 million bid for Spirit data Google won a bankruptcy auction for 34 years of Spirit Airlines operational and employee data, which it says could improve products and AI models, but the sale still requires judicial approval. A flight-attendants union objected over employee privacy, and a related hearing was postponed to September 9.

• Claude unifies memory across chat and Cowork Anthropic merged Claude’s chat and Cowork memory systems so information learned in one interface carries into the other. Users can inspect, edit, or delete saved topics, while storage of sensitive information remains restricted unless they opt in.

• OpenAI adds an administration plugin for ChatGPT Work and Codex OpenAI introduced an Admin plugin that lets workspace administrators analyze usage, manage members and permissions, change limits, and carry out administrative requests from ChatGPT Work and Codex.

• Keenable raises $26 million to build web search for AI agents Keenable emerged from stealth with an Accel-led seed round and says it has indexed more than 100 billion web documents for an API already used by unnamed AI labs and inference providers. The startup plans to double its 15-person engineering team by year-end.

• General Intuition seeks funding at a $6 billion valuation General Intuition, which trains AI agents to navigate space and time using gameplay data, is finalizing an reportedly oversubscribed round at a $6 billion pre-money valuation. The company plans to spend the proceeds on compute, hiring, and adapting its foundation model to robotics.

• OpenAI removes accounts tied to a Russian influence operation OpenAI said it banned Russia-origin accounts that used AI to promote a fabricated Israel-based think tank and a “sovereignty” index praising Russia while criticizing Western countries.

• Anthropic commits $5 million to independent AI-wellbeing evaluations Anthropic launched a grant program offering funding, model access, and technical support for independent, open-source evaluations of AI’s effects on user wellbeing. Applications are due September 21, and selected applicants will be invited to submit full proposals.

Read the full edition →