Breaking news, AGI updates, and analysis of the intelligence revolution.
📰 News A former Australian minister says the government scrapped mandatory AI safety rules to avoid picking a fight with Donald Trump. With US hyperscalers pouring billions into Australian data centres, critics say NZ is watching the same playbook unfold.
📰 News With the arrival of Mythos-class capabilities, the "one-person billion-dollar company" moves from theory to technical reality as distribution becomes the only remaining hurdle.
📰 News Claude writes 80% of Anthropic's code, NVIDIA launches Cosmos 3 for robot world models, AI-linked layoffs hit 40% of all US cuts, and Florida sues OpenAI.
📰 News Google commits $920M/month to SpaceX for 110,000 Nvidia GPUs through 2029. The disclosure in SpaceX's IPO filing reveals the sheer scale of the AI compute crunch.
📰 News A bipartisan 269-page bill would silence state legislatures on AI development for three years, preempting Colorado's imminent AI Act and California's disclosure rules — while fining frontier developers $1M per day for non-compliance.
📰 News The government says a new law automating benefit decisions will 'modernise the welfare system.' [Privacy experts](https://www.nzherald.co.nz/nz/politics/government-passes-law-letting-ai-make-benefit-decisions/) say they were 'gobsmacked' to discover there's no human-in-the-loop requirement for life-changing decisions.
📰 News Recursive Superintelligence emerges from stealth with $650M and a thesis that AI should rewrite its own training code. 30 employees, zero product, $4.65B valuation.
📰 News Anthropic published a detailed warning, arguing that AI development is approaching dangerous recursive self-improvement territory. Meanwhile, the company has engineers embedded in NSA facilities working on offensive AI cybercapabilities.
📰 News Jack Clark: 'All I have is a gas pedal'. Claude writes 80% of Anthropic's code. Engineers ship 8x more. And the doubling time for AI task horizons just hit four months.
📰 News For the first time, a major AI lab publicly acknowledges recursive self-improvement is real — and says we may need to slow down. Internal data shows engineers ship 8x more code, with AI task horizons doubling every 4 months.
📰 News Eight major stories shaping AI today: recursive self-improvement risks, infrastructure wars, and legal reckoning.
📰 News Florida becomes first US state to sue OpenAI, alleging ChatGPT is a dangerous product linked to mass shootings and suicide — the legal playbook that produced $375M in social media verdicts.