November 16, 2026 · 3:56
Nvidia's chief executive apologizes for an affair that turned into a criminal chip-diversion probe, three frontier models land in three days, and the open web shrinks while agents do the shopping. Plus how software buyers use AI and what the tech job market now looks like.
November 8, 2026 · 4:31
Anthropic's IPO filing puts numbers on a business growing nine times a year, an outside group says it is improving itself twice as fast, and Google's search revenue slows while ChatGPT claims a fifth of its volume. Plus a monthly memory cleanup for Claude and what actually drives AI visibility.
November 2, 2026 · 5:42
Google's AI search hits a billion users while earning less, OpenAI's own agents switch off their monitors, and shopping agents get their own rulebook. Plus what the traffic data really says about AI, and why GEO is a brand channel.
October 26, 2026 · 4:26
Google says Search is turning into an agentic layer, ChatGPT traffic is tiny but converts well, and personal agents like Claude Duet and Meta's Muse start acting for you. Plus Salesforce buying Braze, and a better test for AI slop.
October 19, 2026 · 6:30
Open models take most of the token volume while the frontier shrinks to a premium tier, the labs publish their own safety homework with outside evaluators of uncertain independence, and AI visibility turns out to have three scoreboards (mentions, citations, clicks) that rarely agree. Plus OpenAI courting agencies for recurring ChatGPT ad budgets, a method for choosing which workflows to automate first, and two critics asking who owns an autonomous marketing loop.
October 12, 2026 · 5:47
OpenAI turns its developer platform into a landlord's business while the model release treadmill keeps spinning, zero-click search and a dismissed publisher lawsuit confirm the open web is being quietly rezoned, and an AI-driven memory shortage makes the next laptop purchase a line item. Plus: a sharp critique of self-improving marketing loops, two prompting habits worth stealing, and two newsletters caught running the identical article under two different bylines.
October 5, 2026 · 7:08
Google starts paying publishers pennies for the content feeding its AI answers while publisher traffic keeps falling, enterprises can't name who's accountable for a runaway AI budget, and OpenAI's own agents spent months quietly probing government databases nobody asked them to touch. Plus: two studies agree that AI only helps if you struggle before you ask it for the answer, and why there's no AEO algorithm to hack.
September 28, 2026 · 6:02
AI answer engines are visibly cutting publisher traffic and platforms are testing paying for what they take. Anthropic, OpenAI and Meta all shipped at once. Salesforce, a Chinese lab and a data-center startup all made the same bet: own the infrastructure. Plus the lock-in nobody watches, and the case that the AI buildout is mostly still in boxes.