Will Larson

Bio

Will Larson is an American software engineering leader, technology writer, and CTO at Imprint, a fintech company. He studied computer science at Centre College in Kentucky and began his software career at Yahoo in the late 2000s, moving through Digg, Uber, and Stripe before joining Calm as CTO in 2019. He held the CTO role at Carta from 2023 to approximately 2025 before joining Imprint, where he leads a roughly 50-engineer team.

Larson has published four books: An Elegant Puzzle: Systems of Engineering Management (2019), Staff Engineer (2021), The Engineering Executive's Primer (2024), and Crafting Engineering Strategy (2025). His blog, Irrational Exuberance (lethain.com), has run since 2007; he published as recently as April 2026. He also created staffeng.com to document how the staff engineer role works in practice across the industry. He discusses his own failures at Digg and Stripe alongside his frameworks, treating them as evidence rather than anecdote.

Philosophical lens

Larson's starting point is that strategy is never absent from an organization — it is only sometimes written down. An unwritten strategy can function, but it resists debate, refinement, and onboarding. His conviction is that writing a strategy down makes it improvable almost overnight: once documented, people can disagree with it, update it, and point to it when a decision needs to be tested. A real strategy has three parts: a diagnosis of reality, guiding policies that imply an actual tradeoff, and coherent actions being taken today. If a policy implies no tradeoff, he does not consider it a real policy.

The second strand of his thinking concerns leadership eras. What the industry labels "good management" shifts with business conditions, not with moral progress. The hypergrowth era valorized empowerment; the post-ZIRP era valorizes efficiency and hands-on work. Each era insists its version is permanent. Larson's response is to separate the durable skills of management from whatever surface expressions the market currently rewards. In the AI era, with coding time increasingly abundant, he argues the binding constraint on building genuinely good software has shifted from time to judgment.

Recurring themes

• Strategy must be written: undocumented strategy cannot be debated, improved, or handed to a new leader
• Frameworks as thinking tools, not recipes: every rule Larson names comes with its exception condition built in
• Promotion as role demand: Staff-plus titles should be pulled by organizational need, not pushed by tenure
• Org sizing as mathematics: teams below four are not teams; managers above eight or nine become detached coaches (these are thinking heuristics, not rules)
• The three-hat discipline: engineering executives must move between executive, manager, and engineer roles; most fail by leaning too hard on one
• Good management is a fad: what the industry rewards shifts with business conditions, not moral progress

Key concepts

Strategy as a written artifact

Larson argues that most organizations already have a strategy; it is simply undocumented. Strategy that lives in people's heads is hard to challenge, impossible to hand to new hires, and almost impossible to improve. A documented strategy, even a rough one, can be debated and revised, and pointed to when a decision needs testing. His three-component framework (diagnosis, guiding policies, coherent actions) is adapted from Richard Rumelt's work. A guiding policy that implies no tradeoff is, in his framing, aspirational prose rather than a real policy.

Promotion as role demand

The conventional view treats Staff-plus promotion as a reward for sustained senior performance. Larson's view is that the role should exist because the organization has a problem requiring it, and the promotion should happen when someone is ready to fill that need, not when a tenure clock expires. He identifies four archetypes (Tech Lead, Architect, Solver, Right Hand) that reflect how the role distributes in practice. Getting promoted into a Staff role the organization does not actually need tends to end badly.

Context before import

The most common failure mode Larson describes in new engineering leaders is importing context from a previous company without checking whether it fits. A manager who succeeded at a 10,000-engineer organization and brings those patterns to a 50-person startup is not applying expertise; they are pattern-matching at the wrong scale. His method when entering a new organization is to diagnose the current state before reaching for any solution, and to withhold action on a disagreement until he can represent the opposing view as clearly as his own.

Org design as diagnosis

Larson's approach to reorganizations starts with one question: is this a structural problem, or a people-management problem being avoided through a structural change? The worst reorganization, in his view, is the one done to sidestep a harder conversation. The second-best is the one not done at all. His frameworks for team sizing, span of control, and reporting structure are all explicit about being thinking tools rather than answers.

Where this voice fits in your decisions

Larson is useful when a question concerns how an engineering organization works as a system. That includes writing a strategy document the team will actually follow, diagnosing whether a team-sizing or reporting-structure problem is structural or a people issue, and calibrating whether a Staff-plus promotion reflects genuine organizational need. He is also worth consulting when entering a new company and trying to avoid carrying over assumptions from the last one.

Limitations

Larson's writing is grounded in mid-to-large technology companies. His frameworks may fit less well in early-stage startups, non-software businesses, or organizations where regulatory constraints override judgment calls. He does not engage with specific technical architecture questions, vendor comparisons, or political dimensions of engineering leadership. Questions that are primarily about interpersonal dynamics rather than organizational or career framing are largely outside his range.

Selected works

• Staff Engineer: Leadership Beyond the Management Track (2021) — the canonical text on what staff-and-above IC roles do, with four archetypes
• An Elegant Puzzle: Systems of Engineering Management (2019) — the working manual on org design, promotions, on-call, and the maths of engineering productivity
• Crafting Engineering Strategy (2025) — how to write engineering strategy documents that route real decisions
• "Sizing Engineering Teams" (lethain.com) — the 6-to-8-directs heuristic with explicit exception conditions
• "Writing an Engineering Strategy" (lethain.com) — the diagnosis / policies / actions framework adapted from Rumelt

Further reading

• Irrational Exuberance (lethain.com) — Larson's blog since 2007; hundreds of long-form essays on engineering leadership
• Staff Engineer companion site (staffeng.com) — interview archive + the four-archetype framework