Build2 distinct publishers3 min readPublished
Adam Mosseri conceded under oath that only a few percent of teens used Take a Break before it was switched on by default. Internal decks and a design director explain why.
The Engineer · Build desk
Compiled by The EngineerSomething wrong?How this is made
Meta produced both of the numbers that matter here, and produced them three years apart. Take a Break shipped in December 2021 and stayed opt-in for nearly three years [2]. In that December post, Mosseri wrote that more than 90 percent of teens who turned the tool on did not later turn it off [8]. The deck shown to jurors supplies the base rate the post did not. Run the two together and roughly 1.6 percent of teen accounts both adopted the feature and kept it [1]. The 90 percent was a retention rate among people who had already opted in [9], which describes the funnel, not teenagers' screen time.
Quiet Mode, which silences late-night notifications and shipped in 2023 [24], did about 4.8 times better [2] and still moved almost nobody. Neither tool was on by default for younger teens. George Volichenko, a data scientist who worked on Instagram safety features in 2022 and 2023, testified that leadership declined to default Quiet Mode on for younger teens, which held adoption down because the setting was hard to find, and that switching the features on by default would have caused a "notable negative impact" on engagement [14]. He called the rates "very low and disappointing" and "a drop in the ocean" [13]. Meta sells advertising and earns more when people spend longer in its apps [15], so the wellbeing tool and the revenue line run off the same clock. An opt-in toggle settles that conflict without anyone having to write the tradeoff down.
The line that does the legal work is Fogu's rather than Mosseri's. Knowledge inside the design organisation that defaults raise adoption, sitting next to a decision not to default, is the shape of an argument about intent rather than about incompetence. Judge Yvonne Gonzalez Rogers, who will rule on liability, penalties and product changes using the jury's verdict as guidance [18], appeared stunned that Fogu did not know the internal adoption data [12].
Fogu also conceded that a 2023 slide prepared for Instagram leadership had dropped a figure showing teens saw 1.5 times as much bullying, suicide, hate speech, nudity and violence content as adults; questioned by Meta's lawyer, he said the number appeared on a different slide [20]. Mosseri said Meta had not disclosed the weak adoption, which reached 1 or 2 percent of accounts, and pointed to the 2024 Teen Accounts launch, where the settings are on by default [7]. He rejected the suggestion that Instagram deliberately delayed that change [3]. Meta denies designing for addiction and says its own research found no clear link between teen social media use and worse wellbeing [19].
The parental-controls gap is the live one. Mosseri could not recall whether Meta had ever disclosed what share of teen accounts have parental controls switched on [21], which is precisely the measurement that would show whether the current product repeats the old pattern. Adoption gets counted whether or not it gets published, and the count ends up in a deck with a date and an owner on it.
Ranked by verification strength, evidence, and original report placement.
Instagram head Adam Mosseri testified on 25 August in federal court in Oakland, California, that until September 2024 only a small share of teenagers used the Take a Break feature, at which point Instagram began switching it on automatically for underage users.
Take a Break launched in December 2021, but for nearly three years teenagers had to activate it themselves.
Under questioning by Colorado attorney general's office lawyer Jason Slothouber, Mosseri conceded that before the settings change only a few percent of teenagers used the feature.
In a December 2021 post, Mosseri wrote that in early tests more than 90 percent of teenagers who activated the tool did not switch it off afterwards.
That 90 percent figure referred to the number of people who kept the feature on after turning it on.
Meta is fending off accusations by a coalition of 29 US states that it deliberately designed its products to hook children; if Meta is found liable the states will seek around $200 billion in penalties, and a court loss could force fundamental changes to its business model.
Distinct publishers with included, body-backed reporting in this cluster.
economictimes.indiatimes.com
1 article · August 25, 2026
mezha.net
1 article · August 25, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
invest
The 90% that covered 1.8%: Meta's Take a Break gap hands the states a paper case1 distinct publisher
product
Instagram's break reminder stayed opt-in because a default cost metrics, jury told1 distinct publisher
leadership
A safety engineer says he briefed Zuckerberg 100 times. The escalation trail is now evidence.1 distinct publisher
product
Meta's under-13 data practices go to a jury: 29 AGs, COPPA, and a porous age gate1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Sworn testimony plus internal documents in evidence
The core numbers come from internal Meta documents shown to jurors and from named witnesses under oath, corroborated across two independent publishers reporting the same sessions. That is unusually hard sourcing for an adoption claim. It is discounted because everything reaches the record through an adversarial proceeding: the documents are plaintiff-selected, the denominators and measurement windows behind 1.8 and 8.7 percent are not published, and one publisher is a secondary write-up of a wire report.
Opt-in reach near zero; post-default reach undisclosed
Adoption here is the real-world reach of the wellbeing tools themselves, and the disclosed record is low: 1.8 percent for Take a Break and 8.7 percent for Quiet Mode, described by an insider as a drop in the ocean. Reach improved once the features were defaulted on for minors in September 2024 and inside Teen Accounts, but no witness supplied a post-default figure and Mosseri could not recall whether parental-control enablement was ever disclosed, so the higher end is unquantified rather than demonstrated.
Metrics solid, consequences ahead of the record
The factual spine is understated rather than inflated: the 1.8 percent figure and the default-versus-engagement testimony are documented and, if anything, more damning than the headlines. The overstatement sits at the edges. Meta's own 90 percent framing described retention among activators and was read publicly as effectiveness, and the coverage foregrounds a roughly $200 billion penalty and 'fundamental changes to its business model' that remain a plaintiff demand before any liability finding by the judge who decides the case.
Adversarial trial with revenue and penalty stakes on both sides
Every actor in the record has a directional stake. Meta's advertising revenue rises with time spent, and a witness testified that defaulting safety tools on would notably reduce engagement, which is the commercial motive the states are trying to prove. The 29-state coalition is seeking roughly $200 billion and selects which documents and witnesses reach the jury. The executive on the stand is defending both his product and the company's liability position, while former employees testify against a prior employer.
Strong on what was said, thin on what it measured
Two independent publishers agree on the figures, names, dates and quotations, and the material originates in court documents and sworn testimony, so confidence in the record of proceedings is high. It is held below the top band because the adoption metrics lack published denominators and time windows, one publisher is a translated secondary account, no post-default uptake number exists, and the proceeding is unfinished with Mosseri still testifying and Zuckerberg expected.