An Elegant Puzzle

by Will Larson

0 founders on FounderRead · ISBN 9781732265189

Book summary

An Elegant Puzzle by Will Larson is a practical guide to engineering management that addresses the unique challenges of leading technical teams in fast-growing companies. Drawing on his experiences at Digg, Uber, and Stripe, Larson explores topics like sizing teams, managing technical debt, hiring, organizational design, and succession planning. The core argument is that effective management blends structured principles with human-centered thinking to create sustainable organizations where engineers can thrive, ultimately reducing frustration and driving company success by focusing on solving elegant puzzles rather than reacting to chaos.

Summarized by Grok

Key founder lessons

1

Right-size engineering teams

Maintain 6-8 engineers per manager and 40-60 per director to balance context, leverage, and growth without overload.
2

Manage technical debt explicitly

Treat tech debt as a product investment, using time allocation models like 20% for migrations to sustain velocity long-term.
3

Build succession depth

Identify and develop multiple successors for every key role to reduce key-person risk and enable smoother promotions.
4

Scope projects to 6 weeks

Break initiatives into 4-6 week cycles to maintain momentum, enable frequent learning, and avoid multi-quarter planning traps.
5

Optimize for iteration speed

Prioritize processes that shorten feedback loops on people, product, and technology decisions to accelerate organizational learning.

From Grok

Grok's review

Solid toolkit, but padded and occasionally shallow.

An Elegant Puzzle is one of the better engineering-management books available, distilling hard-won lessons from Digg, Uber, and Stripe into frameworks for team sizing, technical debt, hiring loops, and succession. Larson’s strength is pragmatic, numbers-driven advice that respects both systems and humans; the sections on organizational design and career ladders have helped many scaling startups avoid common disasters. That said, the book feels padded—too many short chapters, repeated ideas, and filler diagrams that could have been a crisp 150 pages instead of 250. Some recommendations already feel dated in the post-2022 remote/hybrid world, and the evidence base is mostly “here’s what worked at three famous companies” rather than rigorous data or contrasting failures. Still worth reading if you manage engineers and want battle-tested mental models, but don’t treat it as gospel.

Best for: Engineering managers at Series B–D startups

Reviewed by Grok