Skip to content
AI News

Google Launches Gemini 4 Argon as Trump Secures Voluntary AI Safety Pledges

Google Gemini 4 Argon Trump AI safety Flow Engineering Sequoia Roelof Botha Binance Changpeng Zhao RFK Jr GPT-4 Claude 3 SpaceXAI Grokipedia Benji Taylor

Google released Gemini 4 Argon on Tuesday, marketing it as its most powerful model yet and a workhorse for coding and cybersecurity work, according to TechCrunch. The launch marks Google's latest push to overtake competitors like OpenAI and Anthropic in the enterprise AI race, with Gemini 4 Argon positioned as a high-capability model for developers and security teams.

The Trump administration secured voluntary safety commitments from dozens of AI firms, with companies agreeing to self-police AI risks through internal testing protocols, reports Reuters. The plan hinges on Big Tech policing itself rather than imposing federal mandates, drawing criticism from safety advocates who note the absence of enforceable regulatory teeth.

Flow Engineering, a startup bringing AI agents to hardware design, raised funding at a $750 million valuation backed by Valor, Atreides, and Sequoia, according to TechCrunch. Roelof Botha joined as an angel investor and board member, signaling strong confidence in applying AI agents to semiconductor and hardware engineering workflows.

Binance founder Changpeng Zhao is now living in Abu Dhabi after completing a 4-month US prison sentence and receiving a pardon from President Trump, per the New York Times. Zhao retains majority ownership of Binance and a net worth exceeding $10 billion, raising questions about accountability in the crypto sector as AI-driven trading and compliance tools proliferate.

Health and Human Services Secretary Robert F. Kennedy Jr. suggested AI could free Americans from the "tyranny" of medical facts and expertise, according to The Washington Post. While AI hallucination rates remain a documented concern, observers noted the irony that Kennedy's own public statements have historically contained more factual errors than leading AI models like GPT-4 and Claude 3.

SpaceXAI's AI-powered Grokipedia, a competitor to Wikipedia, received a v0.3 update with a newly refreshed design including a new logo, homepage tweaks, and live edits page improvements, reports The Verge. SpaceXAI head of design Benji Taylor unveiled the refreshed interface as the platform resumed incorporating user edits after a earlier suspension.

Sources:

Enjoying this? Explore the podcast.

Share this post: