Why you should (or should not) work at a startup

Executive overview

Most people join startups for the wrong reasons — expecting rich outcomes, career clarity, or strong management. All three are unlikely. The real case for joining a startup is the asymmetric learning opportunity it creates.

Three concrete benefits justify joining: access to roles you're unqualified for, a fast track to founding your own company, and an accelerated slope of personal growth — whether the startup wins or fails.

The value of a startup isn't the destination; it's how fast the slope rises.

Reasons not to join a startup

• Management at early-stage startups is almost universally poor.
• Mentorship and direction are scarce unless you actively demand them.
• Getting rich from startup equity is statistically improbable — don't join for that.
• If you want a clear career path, five-year plan, or stability, join a large company instead.

Access to jobs you are not qualified for

• Early startups assign roles based on necessity, not credentials.
• Guillaume, a programmer recruited from a YC event, became sole owner of Justin.tv's entire Rails backend within a year — a top-20 Rails site at the time.
• That role led to a co-founder position at Social Cam (0 to 128 million users in two months), then co-founder of Triplebyte.
• The opportunity existed only because there was no one else to do it.

Startup experience as a path to founding

• Working at a startup puts you around people who want to build companies — proximity matters.
• "You are the average of your five closest friends."
• Finn Barr joined Exec from Groupon, wanting eventually to found something. He met his co-founder there, iterated through a failed startup, spent time in YC, and launched Shogun.
• The failed first attempt was the necessary step, not a detour.

Maximising speed of learning

• Kyle Vogt joined Justin.tv as an MIT student with no background in scalable systems. He had to architect a live video platform from scratch, learning in real time under constant failure.
• By the time Twitch sold to Amazon in 2014, that system was the fourth-largest bandwidth consumer in North America, handling 90 petabytes of data transfer per month.
• Justin's brother Daniel learned deal negotiation the hard way — closing a small, low-stakes acquisition of Exec while Justin was in Thailand.
• Two years later, Daniel applied those same skills as a co-founder of Cruise, which sold to GM for over a billion dollars.
• Learning happens on the way up and on the way down. A failing company teaches as much as a rocket ship.

The underlying principle

• Focus on slope, not y-intercept.
• The question to ask is: which environment maximises my personal rate of growth?
• That answer may or may not be a startup — but it should drive the decision.