1. The jury needed two hours to reject every Musk claim against Altman After roughly two hours of deliberation, the jury in Musk v. Altman returned a unanimous verdict against the plaintiff. Elon Musk lost on every count.

2. Anthropic now owns the tool OpenAI uses to ship its SDKs Anthropic has acquired Stainless, the four-year-old New York startup that automates SDK production for OpenAI, Google, and Cloudflare. Terms were not disclosed.

3. Two HN front-page essays this week gutted the "buy AI, ship faster" pitch from opposite sides Two essays attacking the same enterprise AI sales pitch from opposite sides hit Hacker News's front page within days this week.

In Brief

• NVIDIA hand-delivered first Vera CPUs to Anthropic, OpenAI, and xAI NVIDIA VP Ian Buck personally delivered the first Vera CPUs to Anthropic in San Francisco, OpenAI in Mission Bay, and SpaceXAI in Palo Alto on Friday. Oracle Cloud Infrastructure received its unit in Santa Clara on Monday.

• OpenAI and Dell push Codex into on-premise enterprise data centers OpenAI and Dell announced a partnership to deploy Codex coding agents in hybrid and on-premise environments. The deal targets regulated enterprises that cannot send code or data to OpenAI's cloud.

• Jensen Huang calls AI demand "utterly parabolic" at Dell Technologies World Huang said Vera Rubin NVL72 runs agentic inference at one-tenth the per-token cost and that Vera CPUs run agent sandboxes 50% faster than traditional CPUs. He cited 5,000 enterprises on Dell AI Factories, including Lilly, Samsung, and Honeywell.

• Anduril and Meta detail eye-tracking drone strikes for military AR headset Anduril shared details of the AR headset it is prototyping with Meta for the US military. Operators could call in drone strikes via eye-tracking and voice commands, said Anduril VP Quay Barnett, a former Army Special Operations officer.

• Google heads into I/O this week as a distant third in the foundation model race Google's annual developer conference opens Tuesday with Gemini trailing OpenAI and Anthropic across major model benchmarks. MIT Technology Review notes Google was still considered a contender a year ago.

• SandboxAQ ships drug discovery models inside Claude SandboxAQ made its molecular-simulation models callable from Claude, removing the need to write computational chemistry pipelines. Rivals Chai Discovery and Isomorphic Labs are chasing better models; SandboxAQ is betting access matters more than accuracy at the margin.

• Amazon Alexa Plus generates AI podcasts on any topic Amazon updated Alexa Plus to produce podcasts hosted by AI voices from any user-supplied topic. Users can preview the outline and steer the conversation before generation.

• Korean startup LetinAR pitches thumbnail-sized lens as AI glasses standard South Korean firm LetinAR makes a thumbnail-sized optical element it is positioning as the backbone for consumer AI glasses. The pitch lands as Meta, Anduril, and others race to ship displays for wearable assistants.

• Bug bounty programs drowning in AI-generated junk reports Corporate hacking reward programs report a flood of low-quality AI-generated submissions that triage teams cannot keep up with. The volume of unusable reports is straining the economics of paid vulnerability disclosure.

• Court tosses revenge lawsuit built on AI-hallucinated citations A man suing Facebook users who called him a bad date in an "Are We Dating the Same Guy" group filed an AI-assisted complaint stuffed with fake case citations. The court dismissed the case over the fabricated references.

Read the full edition →