What stands out to me in cases like this is not that the companies had internal research - most large consumer platforms run these studies constantly - but how strongly the incentives push teams to ignore the results.
When your core metric is daily/weekly active minutes, anything that reduces stickiness automatically loses in prioritization, even if the internal research says it's healthier. At scale the problem becomes structural: product, ads, and recommendation teams are all optimized around the same engagement loop.
Regulators tend to frame this as a content issue, but the underlying mechanism is really a feedback-control problem. If you let a machine-learning system optimize for “time spent” on a population without mature self-regulation, you get exactly the behaviors described in the filings.
The interesting question isn’t whether these companies knew, but whether we can design incentive structures where “well-being” is a first-class metric rather than something that only appears in internal slides.
What stands out to me in cases like this is not that the companies had internal research - most large consumer platforms run these studies constantly - but how strongly the incentives push teams to ignore the results.
When your core metric is daily/weekly active minutes, anything that reduces stickiness automatically loses in prioritization, even if the internal research says it's healthier. At scale the problem becomes structural: product, ads, and recommendation teams are all optimized around the same engagement loop.
Regulators tend to frame this as a content issue, but the underlying mechanism is really a feedback-control problem. If you let a machine-learning system optimize for “time spent” on a population without mature self-regulation, you get exactly the behaviors described in the filings.
The interesting question isn’t whether these companies knew, but whether we can design incentive structures where “well-being” is a first-class metric rather than something that only appears in internal slides.
[dead]