The Problem with Counting AI Agents

Boris Mann recently pointed out a critical issue with how we discuss AI systems: the tendency to use vague metrics like “11 AI agents” as if they conveyed meaningful information.

Mann analogized this to saying “I have 11 spreadsheets” or “I have 11 browser tabs” — it tells you nothing about what work is actually being done. The number itself is irrelevant; what matters is how these components function together towards a specific goal.

This observation highlights a broader challenge in AI discourse: we often focus on superficial attributes rather than substantive capabilities. When evaluating AI systems, we should prioritize understanding their architecture, functionality, and performance on real-world tasks rather than getting fixated on component counts.