Summary

The post applies Amdahl's Law and Gunther's Universal Scalability Law (USL) to AI agent workflows, arguing that adding more agents beyond an optimal concurrency point causes retrograde scaling — not just diminishing returns, but actual slowdowns. The author backs this with personal empirical data: 4 agents yield an 8.5x speedup over solo coding, but doubling to 8 agents would reduce monthly output due to cognitive context-switching overhead (kappa). The post provides a framework for diagnosing and optimizing agentic team throughput by minimizing contention and coherency costs.

Key Insight

The Universal Scalability Law exposes a hard ceiling on AI agent scaling: beyond the concurrency knee, coordination overhead causes retrograde performance, making architectural discipline and workflow partitioning far more important than simply adding more agents.

Spicy Quotes (click to share)

• 6

  So you would think if I doubled to using 8 agents I would produce 140k lines of code. Instead I would actually produce LESS code, only about 67K per month!

• 3

  When workers have to talk to each other and share space, adding too many workers doesn't just hit a ceiling like in Amdahl's Law, but it can actually make the whole team slower.

• 6

  Agentic scaling (or any scaling for that matter) isn't just about tokenmaxing. It's about how cleanly your system allows those agents to work in isolation without stepping on each other all the time.

• 7

  Coding is the easy part of writing software, it's everything else that takes the most amount of time.

• 4

  You can invite a million people and it will still take 10 minutes to bake.

Tone

analytical