LLenny's Podcast
← All episodes
Farhan Thawar (VP and Head of Eng)19 December 2024

How Shopify builds a high-intensity culture

9Frameworks
15Insights

Frameworks in this episode

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 4

Hot Take09:30

Looking Stupid in Public Is My Superpower

Farhan believes the willingness to ask dumb questions and look stupid is a genuine competitive advantage, and he tries to make it his whole team's superpower. He recounts being told 'that's the stupidest question I've ever heard' and even 'I've already explained this to you three times' — and shrugging it off because the goal is to understand, not to impress.

  • He deliberately asks questions others are too scared to ask, and often finds they had the same question
  • He built the resilience partly from retail and commission sales, absorbing constant rejection
  • When someone says his question is dumb, he just moves on to the next question
  • The point of asking is to understand the content, not to annoy anyone

not everyone can look stupid in public over and over but I believe it's my superpower and I try to make it my whole team…

Farhan Thawar · 09:30

my goal there is not to annoy the person but is to understand the content

Farhan Thawar · 13:00
#leadership#learning#psychology#confidence
Hot Take19:30

Do More Per Minute Instead of Working More Hours

Farhan reframes intensity: an hour is the same hour for everyone, so instead of expanding the workday, compress more output into it. He'd rather get more done in a 9-to-5 and then go take his kid to soccer than stretch work to 12 hours. Really good people can produce high-quality output 2-3x faster, a Parkinson's-law-at-scale effect most people ignore.

  • One hour is the same amount of time for everyone — the variable is calories expended in it
  • Compressing work frees you to do life outside work rather than expanding hours
  • Extremely good people output high-quality work in a fraction of the time a merely good person needs
  • You can cut time by 2-3x without dropping quality much

everyone says oh yeah like work hard and like do more hours when you're young whatever I'm like what if you just did more per…

Farhan Thawar · 21:00

one hour is an hour like it's the same hour

Farhan Thawar · 20:00
#productivity#intensity#management#focus
Hot Take45:30

The Metric That Matters Is Code Deleted, Not Code Written

Farhan pushes back on celebrating AI-written line counts: he tweeted that GitHub Copilot wrote a million lines of code for Shopify and it went viral, but what he actually wants is Copilot deleting a million lines. Shopify runs a delete-code club and hack days where a team hunts for code to remove, routinely finding a million-plus lines. A simpler codebase makes resilience, performance, and maintainability all easier.

  • Shopify has deleted millions of lines this year to rebuild things more elegantly
  • When people cite '25% of code is written by AI,' the real question is where's the 25% deleted by AI
  • The right code is never the voluminous lines-of-code metric — it's elegance
  • A delete-code club and hack days institutionalize the cleanup, with a manual for finding deletion targets

what I want to see is GitHub co-pilot deleting one million Linens of code

Farhan Thawar · 45:30

where's the delete where is the 25% of all kodas deleted by AI

Farhan Thawar · 48:30
#engineering#code-quality#ai#metrics
Hot Take1:03:00

Interviews Don't Predict Performance — Use Job Trials Instead

Farhan argues interviews are a poor predictor of job performance, and he'd rather assess people through real work. He uses a race-car-driver analogy: you can't really quiz a driver, you just put them in the car. His answer is job trials and the 30-60-90 day period, scaled through an internship program hiring 1,000 interns in 2025 — four months of real work product beats any set of interviews.

  • Studies and everyone's own experience show interviews don't predict performance
  • He once hired a coffee-shop enthusiast over a PhD, and the enthusiast is still a phenomenal ML engineer
  • Race-car analogy: to hire the best driver, put them in the car rather than asking questions
  • 1,000 interns in 2025 gives 1,000 real four-month job trials to hire from

interviews are not a good predictor of performance we know this we know this from studies

Farhan Thawar · 1:03:30

if I told you hey I want to go hire the best race car driver there's not really that many questions you could ask them…

Farhan Thawar · 1:04:30
#hiring#recruiting#interns#shopify

Explainer· 3

Explainer22:00

Why Pair Programming Is the Most Underutilized Management Tool

Farhan makes the case that pair programming isn't about lines of code — the throughput limiter is finding the elegant solution, not typing speed. Two engineers on one machine produce less code but move faster toward customer value, with higher happiness, more knowledge transfer, and fewer silos, at roughly a 20% cost. He likens it to the underhand free throw: statistically better but nobody does it because it looks dumb.

  • The bottleneck is designing the right solution, not hands-on-keyboard keystrokes
  • Pairing yields higher happiness, knowledge transfer, less siloing, and higher intensity
  • It's so intense that people new to it sleep 10-12 hours a night at first
  • Analogy: the underhand free throw is statistically better but looks dumb, so nobody does it (even Shaq refused)

it is the most underutilized management tool in engineering Bar None

Farhan Thawar · 22:00

the throughput limiter is not Hands-On keyboard

Farhan Thawar · 23:00
#engineering#pair-programming#management#productivity
Explainer29:30

The Six-Week Review Cadence That Creates Intensity

Shopify replaced its twice-a-year cadence with six-week reviews where teams walk through the roadmap, resourcing, and progress with leadership and Toby. Six weeks is short enough to remember context and long enough to get real work done. Combined with weekly GSD ('get shit done') updates to the whole company, the frequent check-ins invoke Parkinson's law — people want to show progress every single week.

  • Six-week reviews force teams to show what they learned and shipped since last time
  • The cadence balances remembering context (short) with getting a lot done (long)
  • Weekly GSD updates surface progress company-wide every week
  • Feedback isn't saved for the next review — teams iterate the very next day

we find six weeks is a very good Cadence because it's short enough that you can remember the context and it's long enough

Farhan Thawar · 30:30

whatever time you allot to something will be the time it takes

Farhan Thawar · 31:00
#management#operating-cadence#shopify#productivity
Explainer59:30

The Trust Battery: How Shopify Thinks About Trust

Farhan explains Shopify's trust-battery metaphor: your trust with someone is a battery that depletes and grows, and you should charge it when you can and spend the charge strategically. Toby reportedly starts everyone at 50%, but some leaders start people at 100% and only deplete it when things go off-alignment. It's become a shorthand for figuring out how to work with someone.

  • Trust is a battery that can deplete and grow over time, especially in remote work
  • Charge it when you can and use the charge strategically
  • Toby starts people at 50%; others start hard-charging teams at 100%
  • It's a shared vocabulary for calibrating working relationships

your trust with someone is like think of it as a battery that can deplete and grow

Farhan Thawar · 1:00:00

I'm pretty sure Toby says he starts everyone at 50%

Farhan Thawar · 1:00:00
#leadership#culture#trust#remote-work

Story· 6

Story13:30

A Different Billionaire Every Decade — Pair With Irrational Visionaries

Farhan noticed he'd worked for a different billionaire in each decade of his life, and the common thread is that they're irrational long-term thinkers who can see 10-20 years out. He can't see that vision himself, so he pairs with people who can, contributing his ability to push the ball forward 1% a week. He even wrote a list of every unreasonable person he knew in Toronto to find a co-founder.

  • The people he's learned most from share an irrational, long-term view of the future
  • He knows his strength (execution, moving forward) and doesn't try to build a vision skill he lacks
  • The move is to recognize your strengths and weaknesses and merge forces with a complement
  • He literally listed all the 'unreasonable' entrepreneurs he knew and met each to find one he was 'API compatible' with

they have an irrational view of what the world should look like over the next decade or so they're very long-term thinkers

Farhan Thawar · 14:00

I'm not good at seeing that Vision but I'm good at trying to move towards that Vision 1% a week

Farhan Thawar · 14:00
#career#self-awareness#collaboration#vision
Story23:00

Toby's One-Hour Rule: Delete the Code, Keep the Tests

Farhan tells how Toby and his CTO Cody would set a one-hour timer while pair programming, and if they hadn't finished the feature, they'd delete all the code, keep the tests, and start over — even if they were one minute over. The logic: if you can't write it elegantly in an hour, you're on the wrong design. Deleting is hard because of sunk-cost fallacy, but the shorter, more elegant solution appears after pathfinding.

  • Set a one-hour timer; if the feature isn't done, delete the code but keep the tests and restart
  • Inability to write it in an hour signals you're building the wrong thing
  • The best time to rewrite code is right after you wrote it, because you now know the problem domain
  • Deleting is hard due to sunk-cost fallacy, but it lands you on a much better solution

if we were not able to articulate and write the code for this feature in one hour we must be on the wrong design we…

Farhan Thawar · 23:30

code lives a long time code is a liability

Farhan Thawar · 24:00
#engineering#code-quality#pair-programming#shopify
Story37:30

Meeting Armageddon: Deleting Every Recurring Meeting Once a Year

Once a year at a random time, Shopify deletes all recurring internal meetings with more than two people, then bans adding new recurring meetings for two weeks. The point is to break the inertia of standing meetings that persist by default. That reset, plus moving announcements from Slack to Facebook Workplace, cut IC meeting time to about three hours a week — a 50-60% reduction.

  • Delete all recurring internal meetings of 3+ people (keeping 1:1s and external meetings)
  • Two-week moratorium on adding any new recurring meeting forces a rethink of what's actually needed
  • IC meeting time dropped to ~3 hours/week; managers to ~6 hours/week
  • Paired with moving announcements off Slack to a lower-distraction 'river of information' feed

once a year at a random time we will delete all recurring meetings that have more than two people

Farhan Thawar · 37:30

we're at something like three hours of meetings per week for an Niner contributor at shifi which is phenomenal

Farhan Thawar · 39:00
#management#meetings#productivity#shopify
Story42:30

Build Infrastructure, Not Features — Put Gas in the Tank

When merchants demanded NFT gating, Farhan estimated 2-3 weeks to build it, but Toby asked how long to build a platform layer where anyone could build NFT gating in an hour — 2-3 months. Toby said do the platform, because you don't know what people will build on top. He calls it putting gas in the tank, and habitually writes client code against an API that doesn't exist yet to discover what infrastructure should exist.

  • Toby reframes 'you could write this in a day' as 'what has to be true so you could write it in a day'
  • Building the infra layer is slower now (2-3 months vs 2-3 weeks) but unlocks unforeseen use cases forever
  • Shopify buckets work into experiment, feature, or infrastructure
  • Toby writes client code against a non-existent API first, then implements the server to match

if you put the gas in the tank people could drive on that gas for a long time going forward

Farhan Thawar · 43:30

what has to be true so that you could write this in a day what infrastructure do you need

Farhan Thawar · 44:00
#engineering#platform#product#shopify
Story1:21:30

The $18-Month Mistake: Punished for Hedging, Not for Failing

In his first week at Shopify, Farhan hedged on the point-of-sale rebuild — iOS in Swift, Android in React Native — to guarantee a shipped product while learning React Native. A year later React Native was the clear winner and they had to rewrite, burning ~18 months for 100 engineers. When he confessed, Toby's response wasn't anger — it was that Farhan's real mistake was not taking the risk.

  • He hedged to guarantee a product in market, but React Native was clearly the future platform
  • The wrong call cost roughly 18 months of time for about 100 engineers
  • Toby said he comes down harshly on people who don't take risks, not on those whose risks fail
  • Because he was going all-in on React Native anyway (core contributor, release captain), hedging was the wrong instinct

I will always come down harshly on people who do not take risks and you did not take a risk in this case

Farhan Thawar · 1:23:30

I hedged right HED is the worst

Farhan Thawar · 1:22:30
#failure#engineering#leadership#risk
Story1:35:30

The Waitress He Hired Who Became a Director of HR

Farhan spotted a waitress running a chaotic, busy restaurant with total organization, and offered her a job at Extreme Labs. She started as receptionist, became his admin, then a recruiter, and eventually took over an HR function — later becoming a director of HR elsewhere and finishing her degree. He kept hiring talented people from that restaurant, because everyone is smart at something.

  • He scans everyday environments for people who do their job exceptionally well
  • The waitress moved from receptionist to admin to recruiter to running an HR function
  • She parlayed the experience into a university degree and a director-of-HR role
  • Several other strong hires came from the same restaurant

she was kind of like doing an phenomenal job of organizing the entire crazy busy restaurant

Farhan Thawar · 1:36:30

everyone's smart at something right

Farhan Thawar · 1:38:30
#hiring#talent#stories#recruiting

Takeaway· 2

Takeaway00:00

Why Choosing the Hard Path Means You Win Even When You Lose

Farhan argues that when you face a choice, picking the harder option is asymmetric: if it works you did more, and if it fails you've still learned something, done hard things, and worked with smart people. The easy path that fails leaves you with nothing. He illustrates it with job seekers who blast out resumes versus building something on a company's API.

  • The hard path is a hedge against failure: you keep the learning and relationships even if the goal doesn't happen
  • The easy path that fails teaches you nothing because you didn't stretch
  • Instead of sending 10 resumes a day, build something on a company's API to show real learning
  • He deliberately took harder courses and minors (electrical engineering, financial engineering) to be around smarter people

if you do the hard path and it doesn't work actually you still kind of win because you've now done something hard you've probably worked…

Farhan Thawar · 00:00

are you really learning anything from sending out 10 resumes a day why don't you look at the API docs and build something even if…

Farhan Thawar · 00:00
#career#learning#decision-making#job-search
Takeaway50:30

There Are 10,000 Right Options — Don't Stop at the First One

When Farhan brings Toby an A-or-B choice, Toby sends him back to generate more options, because 'A or B' is just picking two of thousands of viable answers. Toby's philosophy: there are unlimited wrong options and roughly 10,000 right ones, but everyone stops at the first right option instead of searching for the optimal one. This pattern rewired how the whole company thinks.

  • Presenting only two options means you stopped exploring the solution space too early
  • The valuable work is finding which of the many right options is optimal
  • Toby refuses binary choices and asks for more options by default
  • Over time the team learns to spot the pattern and generate options themselves

there are unlimited amount of wrong options for any problem there's probably 10,000 right options but everybody stops at the first right one

Farhan Thawar · 51:00

I Don't Like A or B come back when you have something else

Farhan Thawar · 50:30
#decision-making#leadership#problem-solving#shopify