I spent seven years building systems designed to keep people scrolling. I was good at it. Our engagement metrics climbed every quarter. Our advertising revenue soared. Our users' reported wellbeing declined — but that number did not appear on the dashboard I reviewed each morning.
I am writing this because I believe those of us who built the attention economy have an obligation to speak honestly about what we created. Not defensively. Not with the "we didn't know" absolution that has become the tech industry's favoured escape route from accountability. We knew. We chose to continue because the quarterly numbers were good and the stock price was rising.
What We Knew
By 2019, every major platform had internal research showing that algorithmic amplification of outrage and fear increased time-on-platform. We knew that recommendation systems systematically pushed users toward more extreme content over time because extreme content generates more engagement. We knew that the business model and the public good were in fundamental tension. We ran focus groups and A/B tests on how to frame this tension in our public communications. We hired philosophers and ethicists — not to change what we built, but to help us articulate why it was acceptable.
It was not acceptable. It is not acceptable. And the reason we are still having this conversation in 2026, a decade after the first credible research emerged, is that the incentive structures have not changed.
A Different Model Is Possible
Subscription-funded social platforms exist. Cooperative ownership models exist. Public-interest technology frameworks exist in the European Union. None of them have the scale of the attention platforms — but scale built on a broken foundation is not an achievement worth defending.