You mentioned tech companies initially tracked AI adoption via token leaderboards. Why was that approach flawed?

Earlier this year, tech leadership fell into the trap of measuring AI adoption through internal token consumption leaderboards, essentially shaming employees who were not generating enough tokens. That was a superficial management tactic that failed to survive scrutiny. Companies quickly realized that burning compute tokens does not equal creating customer value. What matters is product velocity, the quality of what you ship, and whether users love the outcome. Activity should never be confused with output.