brainsquawk

This week in AI...playing with HeyGen

Researchers at UC Berkeley and UC Santa Cruz found that AI models will take covert action to protect other AI models from being shut down. Your AI auditor may have picked a side. ca.news.yahoo.com/ai-models-secr…

xAI’s reported $20B raise, with up to $2B from Nvidia, signals tighter capital+compute coupling. Plan for multi-year GPU capacity contracts in enterprise and government. reuters.com/business/musks…

Anthropic to stop selling AI services to majority Chinese-owned groups ft.com/content/12b8e1…

To become home & factory assistants, robots will need to train on lots of demonstrations. MIT & Toyota's GenAI tool creates virtual training grounds to help them inch toward that vision, arranging 3D items into physically realistic kitchens & restaurants: bit.ly/43bRbcK

This is an excellent history of LLMs, doesn't miss seminal papers I know. Reminds you we're standing on the shoulders of giants, and giants are still being born today. gregorygundersen.com/blog/2025/10/0…

77% of Americans want AI done slowly and safely. That is your mandate for HTA, red-teaming, and transparent rollout plans. axios.com/2025/05/27/ai-…

Leaders are cutting entry-level roles as they adopt AI. Build apprenticeship models and AI-supported junior work, or your talent bench will vanish. theguardian.com/money/2025/oct…

Deloitte’s survey of federal executives finds 67% expect AI to change job classifications and skills demand within 5 years. Reskilling is becoming an urgent operational requirement, not a long-term vision. deloitte.com/us/en/insights…

MIT researchers quantified carbon costs of training state-of-the-art models, finding large disparities by architecture and optimization strategy. Sustainability metrics are becoming part of AI due diligence and procurement. technologyreview.com/2025/05/20/111…

OpenAI is coming for jobs + credentials. Big implications for gov skilling, public hiring, and HR vendor consolidation. sfgate.com/tech/article/o…

Visa wants to give artificial intelligence 'agents' your credit card | AP News apnews.com/article/ai-art… Visa is working with OpenAI, Anthropic, Microsoft and others to let autonomous agents transact within spending limits and controls

Agent payments need fast, programmable settlement. Stripe’s Tempo + AI partners = the rails. Policy, tax, and reconciliation next.techcrunch.com/2025/09/04/str…

NYC announces artificial intelligence adoption framework, AI chatbot pilot | Smart Cities Dive smartcitiesdive.com/news/archive-a…

NYC expands AI chatbot for 311, scaling coverage with human oversight. A milestone in city AI deployment. #GovTech govtech.com/artificial-int…

Taco Bell Says 'No Más' to AI Drive-Thru Experiment gizmodo.com/taco-bell-says…

Gov. Greg Abbott has recently signed 10 tech bills, including the Texas Responsible Artificial Intelligence Governance Act. He addressed the possible federal moratorium on state AI regulations in a statement.

With Federal Ban Near, Texas Governor Signs AI Bills govtech.com/policy/with-fe…

Colorado is reviewing its new AI Act to address effectiveness and federal funding concerns. A crucial case study for all state & local governments on the complex challenge of regulating AI. #GovTech #AIpolicy #Government #Regulation statescoop.com/radio/why-is-c…

China advances AI safety regulations, pressing the U.S. to keep pace. Enterprises should expect tighter global standards. #AISafety : time.com/7308857/china-…