Dropbox founder Drew Houston promoted Ashraf Alkarmi to co‑CEO and will transition to executive chairman as the company leans into AI and product work.

The Supreme Court declined to block Vermont’s consumer‑protection suit accusing Instagram of engineering features that addict teens, allowing the case to proceed. The decision follows recent state court setbacks for Meta.

After a failed flu trial, Marianne De Backer closed labs, cut staff and struck licensing deals that refocused San Francisco’s Vir on hepatitis and cancer. The result is a leaner company with big‑pharma partners.

Modelers have removed the extreme RCP8.5/SSP5‑8.5 marker from the CMIP7 design; the revision narrows the upper‑end of projections but still leaves large warming and financial risk.

UC Berkeley Law will bar most student use of generative AI in coursework and exams this summer, citing fabricated citations and faulty legal analysis. The policy keeps narrow exceptions for AI‑fluency training.

Theo Baker’s memoir shows how secret seminars, early VC money and campus rituals funnel Stanford freshmen into Silicon Valley power networks.

Kawasaki and Nvidia are reportedly lining up a San Jose robotics R&D hub to test medical and mobility robots with partners including Microsoft and Fujitsu.

Parag Agrawal’s Parallel launched Index, a dashboard that tracks AI agent use of web content and routes payments to publishers and creators.

Google's I/O keynote shifted Gemini into an "agentic" era, debuting Gemini Spark — a 24/7 AI agent — alongside faster models and a new AI video tool. Rollouts start with paid tiers and testers.

A May 12 federal complaint says Hisense and VIDAA snapshot screens every half‑second, build cross‑device profiles and share them with ad‑tech and Chinese affiliates.

Top AI researchers say limited access to TPUs and server time at Google pushed them to found startups that can buy or control compute. The shift is reshaping Bay Area AI.

OpenAI quietly acquired voice‑cloning startup Weights.gg, took its Replay app offline and removed a public catalog of celebrity and political voice models. The tiny team has been folded into OpenAI’s voice work.

A UC Berkeley working paper ties a post‑ChatGPT surge in A grades to students’ use of generative AI. Campuses are redesigning exams and reinstating proctoring to protect degree value.