News & analysis, rated.
Every story from the daily brief, each rated so you know what matters. Hand-picked by the desk, not an automated firehose.
Browse the news archive →IMF puts a number on AI as Asia's demographic offset
IMF managing director Kristalina Georgieva told the IMF-World Bank meetings in Bangkok Monday that AI productivity gains could offset Asia's demographic cliff , with specific numbers attached for the first time.
OpenAI's $50B run rate undershoots the number investors had priced in
OpenAI's annualized revenue sits near $50 billion at the end of September — roughly $20B below the figure some outlets had been citing for weeks.
Nvidia-backed Firmus scraps $5B Sydney IPO as AI-infra demand cracks
Firmus abandoned a planned $5 billion Australian IPO Friday as demand dropped amid valuation, debt and execution risk concerns — would have been Australia's largest tech listing in nearly 30 years.
UK ICO opens the first regulator inquiry scoped to AI agents
The ICO launched a call for evidence on agentic AI risks and confirmed enquiries into recent deployments — the first regulator to scope agents as a distinct compliance surface.
Mistral ships a 1T-parameter flagship as Europe's sovereign-AI bet
Mistral launched Mistral Large 4 in public preview Monday — a 1.05T-parameter MoE with 49B active, trained in European datacenters, open weights promised by month-end.
Sophos says OpenAI's Daybreak cut MDR investigation time 96%
OpenAI published a customer story Friday saying Sophos cut average response time from 38 minutes to 89 seconds using Daybreak agents, with 52% of MDR cases resolved end-to-end by AI.
Anthropic drops Haiku 5.5 to $0.10/M, matching GPT-6 Luna on the eve of its IPO
Anthropic shipped Claude Haiku 5.5 Wednesday with a two-tier price that lands exactly where OpenAI's cheapest model sits — days before Anthropic's IPO.
UK ICO extracts agent-data commitments from ten labs and opens a formal inquiry
The UK Information Commissioner's Office said Thursday it has secured data-protection changes from ten of the world's largest foundation-model developers and opened a six-week call for evidence on agentic AI.
OpenAI publicly rolls out GPT-6 Sol and Luna with an Intelligent UI surface
OpenAI brought the GPT-6 family out of staged access, with two named variants — Sol and Luna — and a new "Intelligent UI" deployment surface.
FDA opens a comment window on autonomous clinical AI, closing October 19
The FDA is accepting public comment on evaluating autonomous AI in direct clinical roles — a category the agency has not previously ruled on. The window closes October 19.
An OpenAI model is credited with "solving" 300+ open math problems — on one source
The Washington Post reported that an OpenAI model produced solutions to more than 300 open research problems in mathematics, outpacing human researchers over the test window.
OpenAI turns ChatGPT into a visual ad surface for 1.2B weekly users
OpenAI opened a native visual-ad slot inside ChatGPT image generation and wired a full measurement stack behind it, with US testing starting later this month.