Cognition's CEO says token spend leaderboards are 'directionally correct' — but some people get carried away (www.businessinsider.com)

Cognition's CEO, Scott Wu, recently shared insights on the effectiveness of token leaderboards in a podcast discussion, highlighting their misalignment with meaningful employee performance metrics. While he acknowledged that incentivizing engineers to use AI is "directionally correct," Wu criticized the tendency to prioritize token spending over actual productivity outputs. Instead, he advocates for a focus on tangible results, such as project completion speed and cost savings, asserting that if engineers can deliver three times the work through AI, that should be the metric for evaluation. Cognition, a rapidly growing AI coding startup led by Wu and others, has raised significant capital, being valued at $26 billion post-funding. As token-maxxing—where teams inflate AI tool usage solely for leaderboard rankings—becomes a concern in the industry, other tech leaders share similar sentiments. They warn against incentivizing mere token spending without corresponding performance achievements, advocating for a more efficient approach to AI utilization that prioritizes cost-effective solutions over sheer volume of use. This conversation underscores the need for a reevaluation of how companies measure and reward productivity in the age of AI.