The Memo Was Never the Strategy

In late April 2025, Duolingo's CEO, Luis von Ahn, sent an all-hands memo declaring the company "AI-first." The company would gradually stop using contractors for work AI could handle, teams would get new headcount only if they could not automate more of their work, and AI use would be folded into hiring and performance reviews. Duolingo then posted the memo publicly on LinkedIn.

The reaction moved faster than any rollout plan could have. Users posted videos of themselves deleting the app and abandoning multi-year streaks, the company lost hundreds of thousands of TikTok followers within weeks, and longtime contractors heard, in the memo's plain language, that their work had become a line item to be replaced. A few weeks later von Ahn wrote a follow-up: one of the most important things a leader can do is provide clarity, and with the memo, he said, he hadn't done that well. Duolingo, he added, would keep hiring at the same pace and did not plan to replace its employees with AI.

I don't think the technology was the problem. Duolingo had real reasons to lean on AI, and it did. What I'd point to is the shape of the announcement: a policy dropped from the top that told people where the tool would go before anyone had learned what it could do for them. In my book I call that the engineering mindset (decide what should be installed, then install it), and its most reliable side effect is that the people on the receiving end hear a headcount plan, whatever the words say.

The cultivating mindset starts somewhere else: with what people in the system have already found worth using, and what they would want more of. The memo announced a destination. Nobody had yet earned the trust to ask a whole company to walk there.

From Dancing in the Wind: An Adaptive Leadership Framework in the Age of AI.

Is your AI strategy something you announced to your people, or something you learned from them?