All the summer moves from Anthropic, Open AI, Google Gemini and Copilot.
Compliance, residency and sovereignty all moved for AI this week - and the eighteen-month bet on owning your own AI stack just started paying off.
Where agentic browsing actually stands this month, why the biggest organisations have the most to gain and the least permission to act, and what happens to any revenue line built on friction.
Claude Opus 5 lands, Hugging Face names the motive behind a sandbox breakout, and the brief and the context turn out to be the only two things no vendor can take from you.
One of the world's best AI models is free to download - plus the $1.5bn training-data lesson, and a half-hour experiment for the weekend and is it time to look at your AI sovereignty?
Horizontal AI spreads the benefit thin. Vertical AI makes it measurable. AI adoption makes the business case work.
Plus: the free tool cutting AI bills 60-95%, and teenagers building real startups with real AI skills.
Fabel 5, Sonnet 5, ChatGPT 5.6 and AI can only actually impact 5% of a human's role.
Three governments, one company pivot, and the four things that go up in value as AI drives the cost of doing towards zero.
The week a government switched off a frontier AI model - and why your AI stack is now critical infrastructure you already know how to protect.
The AI model pricing spread just hit 1,000x. Here's which tier your business actually needs.
The AI model race is a backdrop now. The real work is how your people use what you've already got.
