Root > Articles > Entrepreneurship >

Why I lost faith in Y Combinator

23 Oct 2022 • Yousef Amar • 4 min read

Y Combinator (YC) has historically had a the reputation of being one of the best startup accelerators in the world. YC has a strong community, and in practical terms you get similar access to resources and support as most other great accelerators, but YC's reputation also makes it so that your seed valuation often has an extra 0, just by virtue of being a YC company.

My experience with YC

For full context, I'd just like to briefly explain my history with YC. We applied to YC in September 2020 (before their increase in batch size and investment amount), at a time where the economic climate favoured fundraising. We had a few interviews that didn't really go as we expected. Of all the things the interviewer could be hung up on, she kept coming back to one thing: revenue. We were a pre-seed company with an already relatively mature, complex product (uncommon) but at the stage of trying to find product-market fit.

Revenue was absolutely not a priority for us at that point and these interviews left me genuinely perplexed as to why YC kept asking about it. At the time, we were a marketplace with a transaction fee, and focusing specifically on retention, not increasing our revenue. I thought that maybe the interviewer's understanding of what we do and where we are was just different from reality, though I think we did a pretty good job of explaining that clearly and concisely.

We were eventually not accepted into that YC batch and told to apply again the next batch. Every batch since then we'd get an email saying we should apply, but the door had already closed since we were in the middle of raising our seed round even as we were interviewing with YC, and YC's terms were worse than our investors' in the end, so it was just too late. Anyway, that's not why I lost faith in YC.

Why I lost faith in YC

The main reason is that they started accepting absolute garbage startups. I don't know if this was a side effect of quantity over quality with their new batch sizes, but I knew that YC's reputation inertia was not going to last like this. Personally, I wouldn't want to be associated with the same accelerator as some of those startups. I know that evaluating the potential of a startup can often be quite subjective, but these startups were Juicero-level delusional with no redeeming qualities (like the team).

The example I'd like to dig into a bit more is a startup called DreamWorld, creating a sandbox MMO. There has already been a lot of drama around this game on IndieHackers and even on HackerNews) with people shocked that YC would invest in an obvious scam that would never pass technical due diligence with a clueless founding team. The following is a recent video that summarises this drama quite well. I recommend watching it.

Anyway, the reason I bring this up is because, as it happens, I'm uniquely positioned to be able to give an expert take on specifically the P2P netcode part, given I spent over a decade working on this problem and wrote my PhD thesis on it.

Long story short, it's not impossible, but it's certainly very difficult, and the upfront research and development is not commercially viable. When you want to build an MMO, your main problem is absolutely not going to be scalable infrastructure. In fact, there are quite a few companies that offer this as an affordable service (e.g. Improbable or coherence) so the argument of "don't they know how hard it is to do MMO netcode" doesn't really hold water, since they don't actually have to do it themselves. It's been done.

The hardest part[1] by far is to build a compelling experience that attracts and retains a critical mass of players. MMOs don't die because they have too many players and the infrastructure can't support it, they die because they have no players. Needing to make your game more scalable is a good, and tractable, problem to have. By the time you have that problem, chances are you also have more than enough money to solve it.

Ok, so they've made some promises that they're underqualified to meet. They've convinced a YC panel that either equally didn't know enough about the space to judge the viability of these promises, or had different incentives (I won't speculate on this, but the video above seems to think so). They should not get investment based on a unique tech differentiation, superstar team, or early traction numbers, because they had none.

Here's the irony though: to DreamWorld's credit, they were well on their way to solving a much harder problem, namely building community hype and a strong initial player base. Although it seems like most of this hype was as a result of moonshot goals without anything to back them up and the absurdity of the YC backing and Kickstarter, negative publicity is still publicity. A proper game studio could have taken this momentum and really built something good out of it, so it's a shame that DreamWorld blew it.

YC is not a success signal

It's always been apparent that startups not funded by YC can still become unicorns or be great. That much is a given. But I think over time it's becoming clearer and clearer that startups funded by YC can actually be pretty bad also, to the point where you question their selection process.

In the case of DreamWorld, while it doesn't make sense to me why they were accepted to YC (and I try to be unbiased about this, not thinking "why were they and not us") I really did want to see them succeed, for a couple of reasons:

I actually emailed the DreamWorld team, with a follow-up, offering to back them up on the P2P part in May 2021, but never got a response back. It seems that they eventually just gave up and switched to a standard client-server setup to solve problems with cheating (something I wrote a whole chapter on in my thesis). At any rate, maybe they could pull a No Man's Sky and redeem themselves?


  1. I've never built a large MMO myself, but spoke to people who have, and they confirm this anecdotally. ↩︎