A daily paper of software and AI, made only from primary sources: the people who did the thing, in their own words.
- 1,115publishers read
- 2,016packages followed
- 2,525pieces today
- 13,521in the archive
- 23stories
The promise
- It endsAn edition closes. Reach the end and you're done.
- No algorithmThe same pieces, in the same order, for everyone.
- No accountRead, follow and tune it without signing up.
- Their wordsHeadlines and summaries are the sources' own.
What's in it
- Today's paperOne edition a day, and it ends.
- The beat editionsAI, Cyber, Cloud and Dev.
- StoriesA launch, on one timeline.
- ModelsA sheet per AI model.
- PapersThe week's research, by field.
- DeadlinesWhat shuts down, and when.
- The Weekly ReviewThe week, every Saturday.
- My EditionYour own cut of the paper.
- ClippingsKeep a piece for later.
- OfflineThe day as an e-book.
- The appInstall it; read offline.
- Two viewsJournal or Digital.
- SearchA year of the paper.
- The MonthlyEach month in six pages.
- On your siteA widget for your page.
- The APIEverything, as JSON.
- Our SourcesWho we read, and how they did.
My Edition starts from 18 topics, or any mix you choose.
Questions
Where do the pieces come from?Only the people who did the thing.
The labs, companies, projects and people themselves: their blogs, changelogs, status pages and release notes. No aggregators, reposts or social feeds; every piece links to where it came from. Every source we read →
Who decides the order?The sources and the sections, never clicks.
Pieces are sorted into sections (News, Columns, Technical Guides, New Models, Security Desk, Outages, Research, Releases, New Repos & Packages, AI in the Field, Watch) and by when they came out. What gets clicked never changes anyone's page.
What stays in my browser?Your choices, and nothing leaves.
Your My Edition choices, the stories you follow, the pieces you've read and whether they were worth your time, your recent searches, and whether you've seen the welcome note. Clearing site data removes them.
What does the paper count?Anonymous totals: no id, no IP.
Who comes back: once a day your browser says only the day it first opened the paper, so we can count how many readers return each week. The numbers are public on /numbers/.
What is read: when you open a piece, read it to the end, say whether it was worth your time, clip or share it, your browser says only which piece and which of those, once a day each: no id, no IP, nothing about you. The totals help choose sources and sections; they never change the order of anyone's page.
Which pages are opened, and from where: the name of the section you opened (Rising, Deadlines, the podcast…) or the kind of file you took (a calendar, the EPUB), and the channel that brought you (the ?ref= on the paper's own links, or the kind of site), once a day each. Crawlers aren't counted.
Marks: a count per piece and day; your IP is never stored, only a hash of it and the day, to cap marks per reader. None of this is sent if your browser asks not to be tracked.
What if I sign up or comment?Only what's needed, until you leave.
The daily e-mail: your address and your edition's settings, until you unsubscribe with the link in every e-mail.
Comments: your nickname and text, and hashes of your browser's key and network address so bans hold; see the community rules. Delete a comment and it's gone.
And videos?Nothing from YouTube until you press play.
The paper shows its own copy of a video's thumbnail. Only when you press play does your browser load YouTube's player, from youtube-nocookie.com; from then on YouTube's privacy terms apply to that video.
No tracking, no ads, no cookies.