Cofounders: how to find, test and keep one, the Paul Graham way
Paul Graham calls going solo the first of the 18 mistakes that kill startups: the work is too much for one person. Pick someone you already know well, ideally from building something together, who shares your values. Raise doubts before you start, split equity close to equally with vesting, and fix fights fast: by 2006, about one in five of the startups YC had funded had lost a founder.
Unofficial. Not affiliated with or endorsed by Paul Graham or Y Combinator. Every teaching links to the original essay or talk.
- 1
Why does PG say not to start alone?
"Single founder" is number one in "The 18 Mistakes That Kill Startups" (October 2006). His line: "Starting a startup is too hard for one person." Cofounders bring ideas, talk you out of bad decisions and pick you up on bad days. Going alone also sends a signal: PG calls it "a vote of no confidence", since it suggests the founder could not talk any friends into joining.
In "Why to Not Not Start a Startup" (March 2007) he says that when YC funded two solo founders, it asked them to make finding a cofounder their first job, and both did. In 2023 he told Tyler Cowen he still would not start one alone because it is so much weight to carry. "How to Start a Startup" (March 2005) gives the range: two to four founders.
Key ideaOne founder carries everything. PG’s advice is two to four.
- 2
Where do good cofounders come from?
From people you already know. In "A Student’s Guide to Startups" (October 2006) PG writes: "Co-founders really should be people you already know." College is the best place, because you see many smart people do the same tasks and everyone’s life is still flexible. The founders of Google, Yahoo and Microsoft met at school.
In "How to Start a Startup" (2005) he says a couple of days as friends tells you more than any company learns in interviews. His tests are simple: is the person genuinely smart, do they get things done, and can you stand working with them. Jessica Livingston says the same in a YC clip (posted 2017) from her 2014 Female Founders Conference talk: start with someone you know well, ideally from working or studying together, not someone who is just available.
- List 10 people you have worked with on something real: a project, hackathon, college fest, job.
- Mark who was smart, who finished things, and who you enjoyed working with. Only people with all three stay on the list.
- If you are out of college, PG suggests staying in touch with the best people from it, or joining a company full of smart young people.
- 3
How do you test a cofounder before committing?
Build something together first. In "How to Start Google" (March 2024) PG says it in one line: "You find cofounders by working on projects with them." In "Why to Not Not Start a Startup" he warns that meeting people at events can work, but you need to work with someone before you know if you want them as a cofounder. To Tyler Cowen he put it as: "You just can’t get co-founders instantly."
If no one around you wants to start a company, PG’s 2007 advice is to go where people do, or to switch to an idea people want to work on. YC also runs a free Co-Founder Matching platform. YC takes no equity. When checked in November 2026 the page claimed over 100,000 matches and listed about 1,900 active profiles in Bangalore, next to 3,200 in San Francisco, 3,000 in New York and 2,900 in London.
A cofounder trial, built on PG’s "work on projects together" advice Weeks 1–2: Talk to 10 possible users together. Write up what you each heard. Weeks 3–4: Build the smallest version together and get it to 5 users. Check: Who did what? How did you settle your first disagreement? Would you do 5 more years of this? - 4
What should you look for in a cofounder?
Shared values, different skills. Jessica Livingston’s 2018 advice is to find a cofounder with "complementary skills but the same moral compass as you." She and PG agreed on the big questions and deferred to each other on the small ones. In "Startups in 13 Sentences" (February 2009), "Pick good cofounders" is first, because changing cofounders is much harder than changing your idea.
In a 2018 YC conversation, PG said trust is the thing to look for, and that trustworthiness is harder to judge than intelligence. Find someone you trust, then ask them who they trust. In a tech startup, he adds in the 2005 essay, the founders should include technical people.
- Do not pick someone you dislike because they have a skill you need. PG lists this as a mistake.
- Do not add a friend or housemate just so they do not feel left out.
- Ask about money, risk and time openly. If you disagree on a big question now, it gets worse later.
- 5
How do you stop founder fights from killing the company?
"Fights between founders" is mistake 17 in the 2006 essay. PG writes: "About 20% of the startups we’ve funded have had a founder leave." Most of the fights he saw came from the people, not the situation, and most could have been avoided by choosing better. His advice is blunt: "Don’t suppress misgivings." Problems are easier to fix before the company starts.
Jessica Livingston says in her 2017 video that red flags about trust, work ethic or skill rarely go away on their own, so deal with them right away. In the 2005 essay PG notes that with two or three founders you must settle disputes fast or die.
Key ideaIf you have a doubt about a cofounder, say it before you incorporate, not after.
- 6
How should cofounders split equity?
PG’s 2005 rule of thumb: the split is probably fair when every founder feels they are doing slightly more than their share. It is hard to change later, so "try hard to do it right." He also says all founders should sign an agreement that their ideas belong to the company, and should check what their old employers own.
YC’s more specific advice comes from Michael Seibel (December 2015): split equally or close to it. "If you aren’t willing to give your partner an equal share, then perhaps you are choosing the wrong partner." Use four-year vesting with a one-year cliff, so a cofounder who leaves in year one leaves with nothing and a bad match does not wreck the company.
Seibel’s vesting setup Two founders, 50 / 50. Year 1: nothing vests (the cliff). Leave before 12 months, leave with 0. At month 12: 25% vests. Months 13–48: the rest vests monthly.
Straight from the source
What Paul Graham or YC actually wrote or said, taken apart, then what to do with it this week.
- 1 · Understand
- The first mistake on PG’s list is starting with one founder, and one of the last is fights between founders. Both come down to choosing people well.
- 2 · Break it down
- A startup is too hard for one person, and a solo founder is "a vote of no confidence": the friends who know him best did not join.
- About 20% of YC startups he had funded had a founder leave.
- Most disputes come from the people, not the situation, and could have been avoided with better choices.
- Do not choose someone you dislike because of a skill, and do not add someone out of politeness.
- 3 · Use it
- Before you commit, write down every doubt you have about your cofounder and talk through each one. PG’s rule is not to suppress misgivings.
- 1 · Understand
- Founder breakups devastate startups, so start only with someone you know very well and share values with, and deal with red flags immediately.
- 2 · Break it down
- Pick someone you know closely, ideally from working or studying together, not whoever is available.
- Complementary skills, same moral compass.
- Red flags about trust, work ethic or competence rarely fix themselves.
- In a two-person team, losing one founder leaves one person with everything.
- 3 · Use it
- Use her two filters on your shortlist: do our skills differ, and do our values match? Anyone who fails the second one is out.
Words to know
- Single founder
- Starting alone. Number one in PG’s list of mistakes that kill startups.
- Cofounder trial
- Working on a real project with someone before committing, which is how PG says you actually learn if you can be cofounders.
- Vesting
- Founders earn their shares over time (often four years) instead of owning them all on day one.
- Cliff
- A first period, often one year, before any shares vest. Leave before it ends and you get none.
- Same moral compass
- Jessica Livingston’s phrase for shared values: different skills are fine, different ethics are not.
Watch
Hand-picked from the people who explain this best. The video loads only when you press play.
Four minutes on why cofounder breakups hurt startups and why to pick someone you know well. Watch the first minute twice.
Michael Seibel on equal splits, vesting and cliffs, and common mistakes like extra equity for having the idea.
Put it into practice
Make your cofounder shortlist and plan one trial project
- 1Write down 10 people you have done real work with: classmates, colleagues, hackathon teammates.
- 2For each, answer PG’s three questions from 2005: genuinely smart? gets things done? would I enjoy working with them? Cross out anyone with a no.
- 3Pick the top two. Write one small project you could do together in two weeks, such as interviewing 10 users about a problem you both have.
- 4Write your three biggest misgivings about each person. PG says not to suppress them, so decide how you will raise each one.
- 5If your list is empty, create a profile on YC Co-Founder Matching and start a project with your first match before talking equity.
Check yourself
Pick an answer to see if you are right, and why.
Done? Keep going
Mark this lesson complete to keep your place. Progress is saved on this device, no login.
Questions people ask
Does Paul Graham think you can start a startup without a cofounder?
He calls being a single founder the first of the 18 mistakes that kill startups, because the work is too much for one person. He does not say it is impossible, but when YC funded solo founders it asked them to find a cofounder first.
How do you find a cofounder if you are not in college?
PG suggests staying in touch with the best people you knew at college, joining a company full of smart young people, or moving to a startup hub. Then work on projects with people before committing. YC’s free Co-Founder Matching platform is another route.
How should cofounders split equity?
YC’s Michael Seibel recommends equal or near-equal splits with four-year vesting and a one-year cliff. Paul Graham’s older rule of thumb is that a split is fair when each founder feels they contribute slightly more than their share.
Sources and further reading
- Paul Graham: The 18 Mistakes That Kill Startups (October 2006)
- Paul Graham: How to Start a Startup (March 2005)
- Paul Graham: A Student’s Guide to Startups (October 2006)
- Paul Graham: Why to Not Not Start a Startup (March 2007)
- Paul Graham: Startups in 13 Sentences (February 2009)
- Paul Graham: How to Start Google (March 2024)
- Paul Graham: How Universities Should Prepare Founders (August 2026)
- Paul Graham: What We Look for in Founders (October 2010)
- Conversations with Tyler: Paul Graham on Ambition, Art, and Evaluating Talent (August 2023)
- Y Combinator: A Conversation with Paul Graham, moderated by Geoff Ralston (2018)
- Y Combinator: Jessica Livingston Shares 9 Things She Learned From Founding YC (2018)
- Y Combinator: Jessica Livingston on Cofounder Disputes (2017)
- Michael Seibel, Y Combinator: How to Split Equity Among Co-Founders (December 2015)
- Y Combinator: Co-Founder Matching