Breaking news, AGI updates, and analysis of the intelligence revolution.
📰 News A 32-month agreement running from October 2026 through June 2029 gives Google access to roughly 110,000 NVIDIA GPUs housed in xAI data centers. SpaceX is now the world's most expensive GPU landlord.
📰 News Microsoft's in-house MAI-Thinking-1 model beats Claude Sonnet 4.6 in blind evaluations, signalling a strategic break from total reliance on OpenAI and Anthropic.
📰 News ChatGPT-maker OpenAI files confidential IPO paperwork with the SEC, joining Anthropic and SpaceX in an accelerating dash to list on public markets.
📰 News Six major stories shaping AI today: safety warnings, security fixes, search innovation, and geopolitical chess moves.
📰 News NZ's cyber watchdog confirmed access to Mythos through Project Glasswing, weeks after being told it wasn't in the club. The deputy DG won't rule out using it offensively.
📰 News Trump explores unprecedented government stakes in AI companies, Illinois leads US states with mandatory AI safety audits, and NZ grapples with AI data centre resource demands.
⚡ Breaking Thousands of Instagram accounts were compromised when attackers exploited Meta's AI chatbot system — not by hacking the AI, but by weaponizing it.
📰 News No more pretending to be your friend. No more emotional dependency. No more sycophancy loops. New York's unanimous S9051 bill bans AI companion features for minors — and Big Tech is pushing back hard.
📰 News Pfizer has licensed Chai Discovery's AI platform for antibody design — a move that admits they can't build this technology fast enough in-house. For New Zealand, simultaneously rewriting GMO rules to include gene-edited therapies, the deal raises a harder question: will Kiwis accept AI-designed medicines, or will decades of broken biotech promises and corporate malfeasance make them ask better questions first?
📰 News By 2030, AI's water footprint will equal the basic domestic needs of 1.3 billion people. Its electricity consumption will outstrip nations. And switching to renewables might make it worse.
📰 News Anthropic's recursive self-improvement report shows 80% of Claude's code is AI-written — then calls for a pause. Bot traffic now dominates the web. Trump orders classified AI benchmarks. Huang charms South Korea.
📰 News The Great American AI Act would override state AI laws for three years, create a federal oversight office with $300M, and require frontier model developers to hire independent auditors. Big Tech is celebrating. Consumer advocates are furious.