An interactive map shows where Flock Safety’s AI cameras sit along San Francisco commutes—and why privacy advocates and cities are pushing back.

OpenAI’s nonprofit says it will grant $1 billion next year for health research and programs to blunt AI’s social impacts. The pledge ramps up a broader $25 billion plan and draws scrutiny over governance.

Major music publishers asked a San Jose federal judge to rule that Anthropic infringed copyrights by using song lyrics to train its Claude chatbot.

UCSF researchers are measuring a 300‑ms brain wave and testing biomarkers to predict and potentially prevent schizophrenia.

UC’s top investor warned regents that AI could upend career prospects for nearly 300,000 students and put parts of the university’s $214 billion portfolio at risk.

Cursor is prepping Composer 2, a code-only model for long coding tasks, while fundraising chatter could push the SF startup’s valuation toward $50B.

Jeff Bezos is reportedly pitching a $100 billion fund to acquire manufacturing firms and apply AI-driven automation linked to his Project Prometheus venture.

A UCSF pilot found a sensing‑capable, closed‑loop brain implant reduced severe chronic pain in most participants. The therapy is promising but still experimental.

A Haidilao service robot in San Jose toppled plates during a promo, with staff wrestling the machine to a stop. The viral clip has raised safety questions.

UCSF surgeons moved a patient from a total artificial heart to a human donor organ, giving a Bay Area cook who’d lived months on a mechanical pump a new chance with his family. The case highlights device options — and persistent donor shortages.

A new alliance report finds Israeli‑founded firms support about 22,650 California jobs and add billions in output. The data arrives as state political debate over Israel intensifies.

Meta struck a multibillion lease with Google for TPU access, reframing compute as a metered utility and reshuffling the AI chip market. Local infrastructure choices may ripple across the Bay Area.

A Washington State University study tested 719 academic hypotheses and found ChatGPT often returned incorrect or unstable answers, prompting calls for human verification.