How personalized is it, really?

Two people at the same company with different jobs get different items, different reasons and different actions from the same week of news. Only the two-sentence account of what happened is written once for everyone; the title, why it matters for you and what to do about it are written for your role alone.

Anyone whose work AI is going to change and who has no time to track it. The more concrete your job, the more concrete the email — a recruiter, a controller and a product marketer each get an edition that would be useless to the other two.

How often does it arrive?

Once a week, on Thursday morning. Never more.

Nothing right now. The product is free while we tune it.

What do you do with my LinkedIn URL?

We never fetch it. The URL becomes the text of one search through Linkup, our search provider, and we use what comes back to prefill your role — sometimes with a second lookup on your employer. Nothing is stored until you confirm it on the next screen, and you can delete everything with one request.

Is this just AI news summaries?

No. Summaries are the input. The output is what a given change means for your specific job, and what to do about it this week.

A weekly email with an audience of one

Only the AI news that touches your job — and what to do about it.

Tell us what you do for a living. Every Thursday you get the four to seven things from this week’s AI news that actually affect that job — what changed, why it matters for your role in particular, and one thing to try before the week is out.

It is a newsletter. It is just that only one person gets your version of it.

Get my first edition

Free. One email a week, Thursday morning. Unsubscribe in one click.

What makes it different from every other AI email

Everywhere elseEveryone on the list gets the same twenty links.: HereFour to seven items, kept because they touch your job. The rest is cut before you see it.
Everywhere else“Anthropic raised the context window again.”: Here“Your win-loss notes live in fifteen places. This is the week you can finally ask about all of them at once.”
Everywhere elseYou work out what it means for you, if you have the time.: HereEvery item ends with one concrete thing to do this week — and the prompt to paste, when the work happens in an AI tool.

How it works

01
Tell us what you do
Paste your LinkedIn URL and we prefill the rest, or fill it in yourself: your role, a sentence on what you actually spend the week on, and the AI tools you already use. A couple of minutes.
02
We read the AI week for you
Newsletters, research, product releases, regulation. All of it read once and deduped, then held against your job. Most of it does not survive that filter, which is the point.
03
Thursday, you get your version
Four to seven items, each one written for your role: what changed, why it matters for the work you do, and one thing to try before the week is out. Nobody else receives your edition.

One item, in the format every edition uses

This one is written for a product marketing manager in B2B SaaS.

A recruiter and a finance director reading the same week of news would get a completely different email. Only the two-sentence account of what happened is shared — the headline, why it matters and the thing to try are written for one job at a time.

Release

Claude ships long-context document sets for research and messaging work

Anthropic raised the working context for uploaded document sets and added persistent project files that stay attached across conversations.

1M token context
20 project files

Why it matters for you

Your positioning inputs live in fifteen places — win-loss notes, analyst briefs, pricing pages, call transcripts — and until now you could not look at them together.

The larger working context takes the whole set at once, not a summary of a summary
Persistent project files keep the messaging doc attached, so you stop re-pasting it into every conversation
That changes the question you can ask from “summarise this” to “what do buyers never repeat back to us”

Try this

Upload your last ten win-loss summaries plus your current messaging doc and ask which claim buyers never repeat back to you. Cut or rewrite that claim.

Paste this into your AI tool

Here are my last ten win-loss summaries: [paste them]. Here is our current messaging doc: [paste it]. List every claim the messaging doc makes, and for each one show whether buyers repeated it back, ignored it, or contradicted it — quote the line you are drawing on. Flag the claims with no evidence either way rather than guessing why.

Anthropic2026-08-06
TLDR AI2026-08-07

Questions

How personalized is it, really?: Two people at the same company with different jobs get different items, different reasons and different actions from the same week of news. Only the two-sentence account of what happened is written once for everyone; the title, why it matters for you and what to do about it are written for your role alone.
Who is it for?: Anyone whose work AI is going to change and who has no time to track it. The more concrete your job, the more concrete the email — a recruiter, a controller and a product marketer each get an edition that would be useless to the other two.
How often does it arrive?: Once a week, on Thursday morning. Never more.
What does it cost?: Nothing right now. The product is free while we tune it.
What do you do with my LinkedIn URL?: We never fetch it. The URL becomes the text of one search through Linkup, our search provider, and we use what comes back to prefill your role — sometimes with a second lookup on your employer. Nothing is stored until you confirm it on the next screen, and you can delete everything with one request.
Is this just AI news summaries?: No. Summaries are the input. The output is what a given change means for your specific job, and what to do about it this week.

Tell us your job once. We will read the AI week against it for you, every week.