AI News Today Evening: AI Jobs Shock, Regulation Money, and Google's Talent Drain

RAISE US launches a $500M response to AI job disruption

A new bipartisan nonprofit, RAISE US, launched today with more than $500 million to help American workers adapt to artificial intelligence. The effort is led by former Commerce Secretary Gina Raimondo and former Indiana Governor Eric Holcomb, with anchor partners including Anthropic, the OpenAI Foundation, Microsoft, Amazon, Bank of America, IBM, AMD, Cisco, GM, UPS, Eli Lilly, and Mastercard.

The first pilots will run in Arkansas, Connecticut, Maryland, and Utah, testing ideas like wage insurance, short-term credentials, employer retraining incentives, and AI-powered career coaching.

Why it matters: AI business trends are getting less theoretical. If generative AI can automate work at scale, workforce strategy becomes part of AI adoption, not a separate HR cleanup project.

AI regulation turns into a campaign-finance fight

New York's 12th Congressional District primary became an unusually visible AI regulation proxy fight. The Verge reported that super PACs tied to competing AI policy camps spent $27.41 million around Assemblyman Alex Bores, a coauthor of New York's RAISE Act. Bores lost narrowly to Micah Lasher, 39.1% to 35%, but both candidates supported tougher AI safeguards.

Business Insider reported that Lasher used his win to push back against major AI companies and their allies, arguing that he would not take policy cues from them on jobs, children, data centers, or public safety.

Why it matters: AI regulation is no longer a white paper issue. It is becoming a state-by-state political fight, and companies building or deploying advanced AI should expect policy risk to move faster than federal consensus.

Anthropic alleges Alibaba extracted Claude capabilities

Reuters, via The Economic Times, reported that Anthropic accused Alibaba-affiliated operators of illicitly extracting Claude model capabilities through more than 28.8 million exchanges across nearly 25,000 fraudulent accounts between April 22 and June 5. Alibaba did not immediately respond to the outlet's request for comment, and the claim remains an allegation from Anthropic's Senate letter.

Why it matters: Model outputs are becoming strategic assets. For enterprise AI teams, this is a reminder to treat prompts, logs, model access, vendor accounts, and data retention as security controls, not convenience settings.

Google's Gemini delay now sits beside a talent drain story

This morning's Google story was the reported delay of Gemini 3.5 Pro into July, with Business Insider saying Google is gathering more early-user feedback for long-horizon tasks and agent workflows. Tonight's added context is talent: Business Insider separately framed Google's recent AI departures as a pre-IPO equity story, while Times of India reported that two more senior Google researchers, Jonas Adler and Alexander Pritzel, are expected to join Anthropic.

Why it matters: The frontier AI race is not just model cards and benchmark charts. It is compensation, compute allocation, research focus, and whether elite teams believe their next breakthrough is more valuable inside a public giant or a frontier startup.

Mirendil wants scientists building their own AI systems

The Wall Street Journal reported that Mirendil, founded by former Google and Anthropic researchers Behnam Neyshabur and Harsh Mehta, raised $200 million in seed funding at a $1 billion valuation from investors including Andreessen Horowitz, Kleiner Perkins, and Nvidia. The startup's pitch is AI that helps scientists and open-source researchers develop specialized AI models for fields like medicine and materials.

Why it matters: This is a sharp enterprise AI signal. The next phase of AI automation may be less about one general assistant for everyone and more about domain teams building narrow, high-value model systems with their own research loops.