It’s Tuesday
Here’s what’s worth your time today.
Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, has signed a multiyear partnership with Nvidia to deploy at least one gigawatt of compute powered by its Vera Rubin systems starting next year. The scale of the deal, alongside a significant investment from Nvidia, signals Murati’s ambition to compete at the frontier of AI rather than build on top of other labs’ models. The partnership will also include technical collaboration to optimize Thinking Machines’ systems for Nvidia chips as the company prepares to develop and train its own large-scale models.
Supported by Descope
Every company has an AI agent or MCP server project, but identity woes often become a blocker. Following the MCP auth spec, debating on DCR vs CIMD (or both), issuing short-lived credentials for AI agents…the list goes on.
The Descope Agentic Identity Hub is a dedicated identity provider for AI agents that lets developers easily add auth and consent, granular authorization, credential management, and policy controls to their AI systems.
WisdomAI got MCP auth up and running in 1.5 days
Daylight Security deployed a customer-facing MCP server with auth-built in
Cequence Security uses it under-the-hood in their AI Gateway
A new survey from KPMG finds that fewer than 1 in 10 CEOs at large U.S. companies plan to cut jobs because of AI in 2026. Instead, 55% say AI investments will lead to increased hiring, while 36% expect no change to headcount. Executives remain optimistic about AI’s long-term impact, but many say the near-term returns have been limited as companies struggle to integrate AI into existing workflows and systems. The broader takeaway is that while AI disruption is widely expected, most companies are still early in the transition from experimentation to real operational change.
Anthropic has filed lawsuits against the Pentagon after being designated a “supply chain risk,” a label typically reserved for foreign adversaries. The designation requires companies working with the Defense Department to stop using Anthropic’s Claude model in related operations. Anthropic argues the move violates its First Amendment rights and unlawfully punishes the company for advocating limits on AI use in areas like mass surveillance and autonomous weapons. The Pentagon says the issue is about maintaining operational control over military technology, not free speech. The lawsuits seek to overturn the designation while negotiations between the two sides continue.
Jensen Huang says the global AI expansion is still in its early stages, arguing the technology will require trillions of dollars in new infrastructure over the coming years. In a rare blog post, Huang described AI as essential infrastructure rather than just software, powered by a stack that runs from energy and chips to models and applications. He also suggested the shift will create new jobs as companies expand what they can build and serve. The broader message is clear: despite record spending on chips and data centers, the largest phase of the AI buildout hasn’t happened yet.
OpenAI has postponed the launch of ChatGPT’s planned “adult mode,” saying it will focus instead on improvements to intelligence, personalization, and the overall user experience. The feature, first teased by Sam Altman last October, would allow erotica and other adult content once the company had reliable age-verification systems in place. OpenAI says it still plans to release the mode but needs more time to improve age detection and safeguards for younger users before rolling it out.