Overview of The revised rules of engineering management with Will Larson
In this episode, Gergely Orosz speaks with Will Larson about how engineering management, engineering careers, and the CTO role have changed since the 2010s. Larson argues that many of the assumptions that shaped management in the zero-interest-rate era no longer hold: managers should not try to shield engineers from business problems, senior engineers must care more about business impact and decision-making, and CTOs today need to be both strong technologists and effective operators of complex organizations. The conversation also covers Larson’s career path from Yahoo and Digg to Uber, Stripe, Calm, Carta, and Imprint, plus his views on middle management, career ladders, AI, and writing in the age of AI-generated content.
Main themes and takeaways
The old management model is fading
- In the 2010s, engineering management often emphasized:
- letting managers focus on people/process instead of technical decisions,
- optimizing for employee happiness and retention,
- buffering engineers from business pressure.
- Larson argues that model is increasingly outdated.
- Today, engineers and managers are expected to be closer to the business, more willing to solve painful problems, and more accountable for outcomes.
Business relevance matters more now
- Strong engineers are increasingly expected to:
- understand unit economics,
- understand product constraints,
- make trade-offs under ambiguity,
- work on the company’s most urgent problems, not just the most “fun” ones.
- Larson’s view: senior talent should not expect a guaranteed stream of interesting projects; they should expect to help the business succeed.
AI is changing software work, but not eliminating hard decisions
- AI tools are speeding up code generation and shifting work upstream:
- engineers start before PRDs are fully defined,
- engineers often move ahead of design,
- software delivery is faster, but coordination problems increase.
- Larson believes AI is changing software creation more than anything he has seen in his career.
- But he does not think AI removes the need for:
- technical judgment,
- trade-off decisions,
- product understanding,
- domain context,
- verification and reliability.
Career path and management lessons
What shaped Larson’s leadership philosophy
- Early career experiences at Yahoo and Digg taught him:
- incentives matter more than slogans,
- bureaucracy can create unintended consequences,
- technical systems and organizational systems both need ownership.
- Uber and Stripe taught him:
- how to hire at scale,
- how to manage through chaos,
- the value of durable teams with deep domain knowledge.
Durable, high-ownership teams are still the best structure
- One of Larson’s strongest beliefs is that teams should stay intact long enough to accumulate:
- domain context,
- trust,
- ownership,
- speed.
- He argues against constantly shuffling people onto “important” projects.
- Instead, he prefers:
- small, durable pods,
- clear ownership boundaries,
- protecting critical teams from distractions,
- letting them solve the problem end-to-end.
Why moving to Imprint made sense
- Larson intentionally stepped down in scale from Carta to Imprint to:
- get closer to the code again,
- relearn modern software development,
- adapt to the AI era,
- regain hands-on technical depth.
- He says this was not a downgrade, but a deliberate reset.
What makes a great CTO today
CTOs need both technical and bureaucratic skill
Larson’s core definition of a strong CTO:
- A strong technologist who understands what should be built and how.
- A strong operator/bureaucrat who can run a function, make trade-offs, and manage processes.
He stresses that:
- people who rise through middle management may become better bureaucrats but lose technical sharpness,
- people who rise straight from engineering may keep technical credibility but lack operating experience,
- the best future CTOs will likely be those who can do both.
Line managers may be better future executives than middle managers
- Larson thinks line managers often make better executives because they remain closer to the work.
- Middle management, especially in the 2010s, often optimized for:
- process,
- headcount,
- survey scores,
- shielding teams from business pressure.
- That created managers who were strong at bureaucracy but weaker at technical and product judgment.
Career ladders and engineering roles
Career ladders still matter, but old rubrics are outdated
- Larson still supports career ladders because they help with fairness and clarity.
- But he thinks many company ladders are stale and overly specific.
- Promotions at higher levels are often better understood as:
- better decision-making,
- better judgment in ambiguity,
- stronger business context,
- broader ownership.
Staff and principal engineers are about judgment, not just coding
- For senior IC roles, Larson emphasizes:
- navigating ambiguity,
- making trade-offs,
- building prototypes to test assumptions,
- understanding product and business implications.
- AI can help with implementation, but it does not replace decision-making.
Junior roles are still valuable, but expectations are shifting
- Larson is skeptical of overly rigid definitions of Eng1 vs Eng2.
- He sees those levels as more about:
- time in role,
- growing judgment,
- increasing context,
- developing reliability.
- AI may reduce the amount of code juniors need to write, but not the need for them to learn judgment.
On middle management
Middle management is under pressure
- Larson is blunt: it’s a hard time to be in middle management.
- The old model rewarded people for:
- scaling teams,
- managing process,
- solving through people and structure.
- The new model values:
- technical closeness,
- speed,
- direct decision-making,
- hands-on contribution.
His advice to people stuck in the middle
- Either:
- move closer to the work,
- take a lateral or backward step,
- return to line management,
- or accept that your current skill set may be less valuable in the new environment.
- He sees this as a chance to reskill, not a failure.
Writing, influence, and AI-generated content
Writing has been central to Larson’s career
- He has been writing online for nearly 20 years.
- Writing helped him:
- network early in his career,
- explain what he was learning,
- shape the industry’s thinking about staff engineering and engineering leadership.
- He sees writing as a way to change the field, not just produce content.
He’s skeptical of AI-written “slop”
- Larson does not believe AI should replace an authentic writing voice.
- His argument:
- if you’ve spent years building trust and a point of view,
- using AI to mass-produce generic content destroys that advantage.
- He prefers a small number of strong, original ideas over high-volume output.
- He believes the same is true for social media: one well-formed idea matters more than constant posting.
Notable ideas and quotes
Key ideas
- “Durable, high-ownership teams with domain context” outperform constantly reshuffled teams.
- The role of management changed because zero interest rates changed incentives.
- Strong engineers and leaders should focus on business impact, not just technical elegance.
- The future CTO must be both a technologist and an operator.
- AI increases the need for judgment, not just speed.
Memorable framing
- Larson repeatedly contrasts:
- interesting work vs important work,
- technical judgment vs bureaucratic skill,
- writing code vs making the right decisions.
Bottom line
This episode is a sharp look at how engineering leadership is being rewritten in the AI era. Larson’s core message is that the industry is moving away from management as shield-and-process and toward leadership as ownership, judgment, and business alignment. For engineers, managers, and aspiring CTOs, the practical advice is clear: stay close to the work, learn to make trade-offs, understand the business, and don’t assume old career scripts still apply.