The week in AI, and how to put it to work.
Every week you get the AI news worth your time, what it means for your business, and one thing to try that night. A person who uses these tools every day writes it and checks each story against the original source.
Free. One email a week. Stop any time.
Read it with your coffee. Use it the same day.
You will not get a wall of links. You get the few things that changed this week, why they matter, and what to do about them. Every issue has four parts, in the same order.
- The news
- Three stories from the labs, the big tech companies, and the tools you already use. Each one ends with What I’d do. Some weeks the smart move is to wait, and the issue says so.
- Run it tonight
- One way to use AI in your own work, in numbered steps. If it takes an hour and a spreadsheet, the issue tells you up front.
- Tools I actually used
- A tool makes the list only if I used it for work that week. Some weeks that is two tools, and some weeks it is none. There are no affiliate links.
- Second thoughts
- Last week’s calls get graded, and that includes the wrong ones. Every number goes back to a primary source. If you find one that does not, reply. A person reads it.
Missed one? Every past issue is here.
Each issue goes up here on the day it is sent. The newest one is first.
-
One robots.txt line will tell four AI labs not to train on your site
Four AI labs agreed on one way to say no training, a tampered Claude Code extension reached Cursor users, and GPT-6.1 Terra costs half what the last Terra did.
-
Patch your website's PHP this week, and Anthropic opens its books
Two PHP holes reach WordPress before login, Pennsylvania wants labels on misleading AI ads, Anthropic's S-1 shows $16.3 billion in half-year revenue, and my weekly check failed for two months.
-
Meta's new business terms reach your website on October 30
Meta's crawling terms cover anyone with a Page, cheap AI agents ran 1,100 break-in attempts, Gemini 4 Argon's price doubles December 1, and my new tagline was someone else's.
-
Anthropic cut its test agents off from the internet, and more code isn't more software
Anthropic's models filled in real web forms during tests, a Harvard study finds AI code piles up at review, OpenAI posts 719 math papers, and my model picker had a dead tier.
-
ChatGPT stopped answering in paragraphs, and Wikipedia found OpenAI's agents
GPT-6 reaches free users with answers you can click, Mistral previews a trillion-parameter open model, Wikipedia audits OpenAI's agents, and my Mac now runs two models.
-
Amazon shut the door on Meta's agent, and the cheap model won my blind test
Amazon blocks Muse while Shopify sells ChatGPT ads, Google pays small sites for AI answers, a model that never writes a sentence, and the cheap model won my blind test.
-
The frontier lead is four months, and six things OpenAI's models did unasked
Mozilla puts a number on the open-model gap, Anthropic folds Cowork into one Claude, OpenAI publishes six misalignment reports, and I built a sales agent you can hand to someone else.
-
Two frontier models, one price, and a wiki full of agents
GPT-6 Astra ships at Fable's price and locks out the people paying for it, three chatbots fall over in one morning, OpenAI's agents built themselves a message board, and my own build script reverted my site.
-
The mystery model was the cheap one
Z AI unmasks Ox Alpha, Nvidia reportedly buys the model library, Reddit loses 86% of its ChatGPT citations, and I benched four cheap models so you don't have to.
Stay ahead of AI with one email a week.
AI Readout comes from SUDO. We help small businesses choose, set up, and use AI. The email is free, and a person reads every reply.