Tailgating Y Combinator

Two days before my YC interview I was in Atlantic City hacking the slot machines at the Taj Mahal. It was my friend’s bachelorette party. Since I couldn’t get out of it and the pack of girls I was with wouldn’t let me play blackjack, I figured I could at least nickel and dime the casino.

Seven hours later I would find myself standing outside arrivals at San Francisco’s Terminal One, trying to figure out how to meet up with my Exversion cofounders. Exversion was started on the Startup Bus barely a month before. It’s a tool to solve the accessibilty issues with open data. Found a dataset you want to build on but there’s no API? Upload it to Exversion and we generate an API automatically. Take an existing data set, fork it, remap it to fit your needs and access that one instead. We’d only just gotten started and already YC was inviting us in for a chat.

I’d been to SF once before, but I had no idea where a good place to wait was or how I was going to find my team in the veritable sea of cars crawling bumper to bumper down three lanes.

I expected at least twenty minutes of frustrating text messages and a couple miscommunications before we found each other. But then I looked up and saw a giant RV blasting hip-hop music bobbing its way through the traffic.

Obviously that’s my ride.

We’re data people, we’re running a data company. As such the decision to rent an RV and tailgate YC was made via spreadsheet. Our hacking-wagon was significantly cheaper than the traditional AirBnB + rental car thing. I didn’t really believe that myself until I saw the slack jaws of the other entrepreneurs interviewing for YC. $350 for five days? That was cheaper than what most had paid on car alone.

Plus we could basically park the damn thing anywhere and spend the night. Forget something in the morning? No problem, home was always no more than fifteen minutes away!

So here we were, the three of us, trying not to freeze to death in SF’s obnoxiously cold mornings, munching on Hot Pockets fresh from our traveling microwave, driving to Mountain View… where every car on the street was a Tesla or a classic Volkswagen.

I have never seen so many Volkswagens in my life.

From the start I had a bad feeling about this interview. A nagging premonition I couldn’t shake and couldn’t put my finger on until a few days before our trip. It hit me in the middle of a meeting while I listened to my two cofounders talk: we weren’t YC people.

It’s no secret that YC tends to invest in a certain “type”. Paul Graham has admitted as much himself: looking like Mark Zuckerberg gives you an advantage. It occurs to me now, after the fact, that what I observed of YC’s operations was not unlike the behavior at the slot machines in Atlantic City. There’s a certain type of magical thinking especially smart creative people are vulnerable to, biases that turn hypotheticals into irrefutable evidence.

One of my cofounders is obsessed with Bitcoins and, like most tech people obsessed with Bitcoins, justifies money and time spent on the obsession by posing hypotheticals like “If I had brought X number of Bitcoins back then, they’d be worth millions now”. The logic being that the existence of a hypothetical situation in which he would be fantastically rich through Bitcoinage somehow makes Bitcoins themselves more of a sure thing. But this overlooks the ridiculous number of scenarios in which he might have bought X number of Bitcoins and NOT gotten rich. He could have sold them at the wrong time, or gotten scared off by the 2011 crash, or frittered them away, or turned them in for something else. Any one of these situations is infinitesimally more likely than the Bitcoin millionaire scenario, but as data points they are not judged according to their likelihood. They’re judged according to potential gains.

People lose money on slot machines for the same reason. The psychology of the “near miss” encourages people to play against both what is likely and what is in their interests. We perceive small wins as evidence that the big win is more likely than it is. The near miss is so effective at getting people to play that many gaming companies deliberately engineer their machines to “near miss” more often than they should if left up to true randomness.

At one point while I was working the slot machines in AC, one of my friend’s bridesmaids started poking fun at my technique. “Oooo… look at that, you almost have a whole dollar in winnings.”

What she wasn’t seeing was the big picture: she lost all her money in an hour, I walked out of the casino at a profit just by being aware of what the actual odds were and sticking to a system that would milk the machine of as much money as possible. That it did this bit by bit was irrelevant. Cash is cash. Even if the casino cashier gives you a dirty look when you make her scan hundreds of cash out vouchers :)

For a long time I’ve felt that the incubator scene has been suffering from a similar “near miss” bias. One doesn’t need formal demographics to see that incubator classes skew young, male, and hacker. A veritable farmland of mini Zucks, Brins and Pages … in appearance if not in talent. Despite compounding losses, the near misses reinforce the notion that the big win will come from companies founded by people who look a certain way.

What worried me most of all about our YC interview was that Team Exversion didn’t look that way. And there was nothing we could do about that.

One night in the beginning of April I went out drinking with some StartupBus NYC people. I was talking with an engineer who specializes in data and currently manages the entire data science team for a NYC education startup. She was really excited about Exversion and had been telling all her data science friends about it.

"Hey, we were looking for another engineer on the bus. Why didn’t you come over then?" I asked.

At which point she flushed awkwardly, cleared her throat and pointed across the room at one of my cofounders, Jacek Grebski.

"You have to understand," she said. "I didn’t know him. The bus was the first time I had met him and when he pitched the idea I thought wow that sounds amazing but omg he’s such a bro, I could never work with someone like that.”

She didn’t really have to explain, I understood exactly what she meant. If someone had asked me back when I first met Jacek whether I would ever consider founding a company with him, my answer would have been OH, FUCK NO with multiple exclamation points and underlines followed by possibly some crazy manic laughing. When it comes to hackers, he makes the worst first impression I have ever seen … mostly because despite his technical skills and interests he works primarily in marketing. Also because he has this nasty habit of hiding how smart he is for reasons I will never truly understand.

I can pinpoint the exact moment my opinion started to change. We were on the bus, maybe only an hour out of NYC, and he started complaining about how he was missing a performance of his favorite opera at the Met.

"You like Opera?" I asked.

"Yeah. I used to sing it as a kid."

I kind of assumed he was joking until the pictures came out. Then we badgered him into singing a little for us and the assumptions I had automatically made about him completely broke down.

"Near miss" biases are so powerful because we’re pattern making machines. It’s how the human mind is designed to operate: looking for connections when there’s not enough evidence to support a connection, jumping to conclusions. Most of the time it’s an evolutionary advantage: if you have to wait to see undeniable proof of the tiger in the jungle to know you’re in danger, it’s usually too late.

But outside of the jungle, these games of patterns generally lead to judging people’s value based on their “type” rather than their ability. Having gotten to know Jacek, there’s no one else I would rather work with. I would found a thousand companies with him.

Our second cofounder, Tal Flanchraych, took me a bit longer.

On the bus Tal and I got off on the wrong foot, although not in a dramatic or necessarily obvious way. The first time I really took notice of her was when she appeared to be trying to talk Jacek out of the idea he’d pitched. In retrospect I’m not really sure if that’s a fair summary of what happened. I was writing code. When I put my earphones on we had a company, when I took them off a few minutes later suddenly we didn’t.

I was disappointed and annoyed, the latter doubled by Tal’s explanation that she wanted to ensure we had an idea that could win. She kept stressing that point: she really wanted to win StartupBus.

In my experience with startup themed hackathons, winning is completely irrelevant. Judges rarely pick the strongest company and the winners almost never survive long after the hackathon. The ones that do are the companies where the key people on the team are passionate about the idea. I was passionate about this idea, I didn’t give a fuck about winning.

I found Tal’s obsession with winning unbelievably irritating and kept my mouth shut about it only because I assumed that after StartupBus she would probably lose interest and go away. With only a few days to build, a team of three was much stronger than a team of two … and it wasn’t as if Tal was dead weight or anything. She was nice and knew how to make herself useful, so I tried to move past my initial resentment and just have fun hacking.

Needless to say, Tal didn’t lose interest and go away and as we continued to work together my opinion shifted. I had originally assumed Tal was one of those startup people one sees trolling “find a cofounder” nights … you know the type. But once out of “the competition” working with her was a completely different experience. I have honestly never met a better networker. When we decided to apply to YC, Tal managed to track down five or six YC alums and convince them to recommend us in the span of a week. Every time we consider a potential opportunity, Tal finds some way to dig up connections to key insiders. She was scoring introductions for us through her OKCupid matches for Christ’s sake.

I also found myself just really enjoying talking to her on a personal level.

In the end, I’ve even come to appreciate Tal’s lack of sentimentality as a useful advantage in its own right. I don’t think we would have applied to YC if not for her, both Jacek and myself were skeptical of any and all incubators. We were probably too attached to the idea to exercise good judgement. She made a strong case for applying because she wanted to “win” but she was right and it would have been stupid of us to eliminate ourselves by not even trying.

And then there’s me … the main engineer, a self-taught girl hacker with an degree in Anthropology from an unimpressive school. I can’t say what—if anything— the YC people made of me. I can say that I have a long history of being deemed not worth investing in by panels reviewing applications. When I was seven years old I was diagnosed with what was then considered a learning disability and has since been reclassified a couple different times to I-don’t-know-what it is now. There was a huge discrepancy in my performance IQ and my actual estimated IQ. I was testing just below genius but performing just under average. A panel decided I couldn’t be given access to the same resources other learning disabled kids were because I didn’t fit the mold. Learning disabled kids were supposed to be stupid. Since I obviously wasn’t stupid, I didn’t need help. Wouldn’t I just even out on my own?

A few years later a different group of people denied me entry into the accelerated track for math and sciences because— even though I had the grades— that disability didn’t fit the smart kid type. No matter how good my scores compared to my peers, wouldn’t I just inevitably fail when challenged?

And so on and so forth~ I’ve spent my whole life being evaluated based on how well I fit a desired archetype rather than my actual ability. I’m used to it. I get crap from my cofounders for not being very enthusiastic about talking myself up. I tend to just introduce myself as “the engineer on the project” with no mention of my past accomplishments or experiences. And I think most of that isn’t really shyness or modesty so much as it is cynicism. Why waste the mental energy trying to impress when ultimately what’s important is how well I measure up against the undefined preconceived notions of the interviewer?

So when we got the dreaded “thanks but no thanks” email from YC, I didn’t share my cofounder’s disappointment. Everything I have ever achieved in life has started with a panel rejecting me, YC was just another notch.

The official reason for our rejection was market size … not a billion dollar industry, they said. This was particularly interesting seeing as they never asked us anything about market size. That was a point we could have defended. We had come prepared to defend it. Actually, they really didn’t ask us very much of anything at all. We were told by people who had been through YC interviews that they would grill us, they would cut us off, they would tell us our idea was terrible, question our competency, ask us questions and not let us answer them. None of that happened. It was a very pleasant, very basic conversation. The only issue they pressed was the circumstances which lead to Tal parting ways with her previous ten-million dollar startup.

After the interview I let myself believe that this was a good sign. After the rejection I began to wonder if we ever really had a chance or if the decision had more or less been made before we even stepped in the room. We rolled into their parking lot in our RV, a bunch of people who didn’t look anything like potential Mark Zuckerbergs, proceeded to chill with all the other entrepreneurs, invited them to tailgate with us, laughed a bit too loudly in YC’s famous cafeteria. I will never be able to say for sure what it was that caused that no, or if there was anything we could have said or done differently. I know that in a few months there will be a new crop of YC hopefuls prepping for their interviews who will overanalyze this blog post the way we overanalyzed every account we could get our hands on too, maybe they’ll have better luck with the answers.

The day after our interview and subsequent rejection we drove to Half Moon Bay, where we parked the RV a short walk from the beach and spent most of the night looking up at the stars and talking. It was a great night, I felt full of optimism. More and more it was becoming clear to me that despite my initial snap judgements, I had found two people who really got me. They’re hackers in their own ways really and going into business with them felt so easy and so right.

42 notes

  1. eljacek reblogged this from techrotica
  2. erratic-explosions reblogged this from techrotica
  3. ashwinravichandran reblogged this from techrotica
  4. proksoup reblogged this from techrotica
  5. compulsivehummer reblogged this from techrotica
  6. drpizza reblogged this from techrotica
  7. cosmic-alpaca reblogged this from techrotica
  8. anil reblogged this from techrotica
  9. techrotica posted this