This month, I interviewed Fred, who was the third person at a Y Combinator startup and had a very negative experience, including basically having his equity stolen out from under him. I don’t think this is universal – I’ve met others with much more positive stories – but it still seems important to share, with lots of fascinating details. “Fred Gardner” is a pseudonym to protect his identity, but I’ve met him in person and his claims seem credible.

Alyssa Vance:
So, you’re a startup founder whose background is in computer science, right?

Fred Gardner:
Yes, that’s correct. At this point, I’ve founded one startup, been on the initial founding team… one of the first couple of founding engineers of two more. My background is in CS, with a focus on machine learning and cryptography.

Alyssa Vance:
What appeals to you about being a founder, versus getting a job at someplace like Google or Microsoft, where you’re just working 40 hours a week and being comfortable?

Fred Gardner:
It’s two things. One is the real sense of autonomy that you get. As a founder, there’s a problem that you need to solve, and you have the whole wide space of possible tools in front of you, and it’s up to you to figure out the things that are best suited to solve that problem. That’s the type of problem I most like solving. And the second thing is having a really, really tight connection with the people who will actually be using your solution.

Before going into entrepreneurship, I’ve done a lot of work at R&D and defense companies. And the biggest complaint that I had about that type of work is I never knew whether a) my solution was useful, and b) who was going to be using it. This is a concern that I’ve had a lot about larger tech companies as well – that you’re so disconnected from the people who are actually using the solution; you’re building a product for a team that’s connected to a team that’s connected to a team, that’s connected to someone who’s attached to the product. At a smaller company, you’re really directly interfacing with people who are using it.

Alyssa Vance:
Cool. What got you interested in both computer science and entrepreneurship?

Fred Gardner:
The honest answer is going to be somewhere between:

– I was on the internet long enough that the descriptions of the work people did in CS looked really interesting to me, and I started doing it because of that, or;

– Seeing my brother start getting into it in late middle, early high school. I’m two years younger than him, so that would have been early middle school, I started getting into it as well, because it looked interesting.

In terms of entrepreneurship, that’s a much cleaner story. I spent the year between high school and college on a mountain in the middle of nowhere in Israel, and a friend of mine came to me with an idea that they wanted me to evaluate tactically. The details involved a novel way to do key sharing, and to enable some unique business use cases for data privacy. I looked at it and said, “Hey, this is a super interesting idea.” The company itself no longer exists, but I got into entrepreneurship when we were talking about this idea, said, “hey, wait, there’s actually something here”, and we decided to make a company out of it and see where it went.

Alyssa Vance:
So this company you founded, that eventually got into Y Combinator – what were the origins of that? How did you meet the other people, and what made you decide to get involved?

Fred Gardner:
I was part of a club at college, and one of the other founders of that club approached me with an idea, and said “hey, I want your technical evaluation of, if what we’re describing is at all feasible.” I met with them for a few hours, and it was like, “hey, I don’t think idea A is at all feasible, but here’s how you could sort of serve the same vertical, and change it into idea B.” And they were like, “hey, that’s actually really interesting. Would you like to come join us and help us build this?” So I was like, “Okay, who else is with you?” “Just me, the co-founder, and you would be the third person joining onto this thing.” This seemed interesting – it was a vertical that I thought was, if not the most impactful, at least somewhat of a niche vertical, I thought I could be of use – so I started working with them.

Alyssa Vance:
Okay. So you started this company and it didn’t go so well, is that right?

Fred Gardner:
Let me put it this way. We entered the summer with six people, four of whom were technical. We left that summer with three people, one of whom was technical; I was no longer with the company by the end of that summer. And that last technical person left within two weeks after that. So, “it didn’t go so well” is certainly one way to put it.

Alyssa Vance:
What were the different roles of people within the company?

Fred Gardner:
Forgive me if I don’t recall formal titles, but I’ll do my best. The two co-founders were the CEO and COO respectively, and they were mostly product design, a bit of product strategy, not really anything technical, talking to potential investors and customers. There was the CTO – who we later found out was lying about his technical work, but that’s a whole other story. There was me heading up backend development, and there was our principal front end engineer, who was heading up content development.

I was the first person that they brought in to do technical work. In particular, that meant that I spent a lot of time speccing out the system. I pretty much wrote the initial systems spec, built probably 30% of the first system with another couple of technical people.

The initial signed cap table had, of the six people, a pretty even split of 20% for each of the co-founders, and 15% for each of the other people, with dilution coming out of all shares equally. Part of the reason that I was initially comfortable joining was because we had signed this initial cap table. I was like, okay, the two co-founders aren’t technical, but they acknowledge this flaw and are giving equity to people commensurate with this. This, I assume, was the cap table that YC was initially shown, because this was the cap table that existed at the time that I interviewed.

Alyssa Vance:
You say it’s a whole other story, but how did you meet this person? Like, how did you hire this person?

Fred Gardner:
This person came pretty well recommended by the people at college. We picked him to be the CTO rather than me, because I made the choice that I wanted the CTO to be someone who had familiarity with the whole tech stack; he had significantly more frontend experience than I did. It turned out later that he had been lying in progress reports to the team about what stuff he had done vs. what stuff other people had done, and had been lying to each person on the team, claiming the responsibilities and Git commits of a different person.

Alyssa Vance:
Well.. okay. So the two people who initially started the company, or had the idea – it was a software company, and…

Fred Gardner:
Neither of the founders was technical, that’s correct.

Alyssa Vance:
Yes. What was their plan, exactly?

Fred Gardner:
In so far as I can understand it, it’s somewhere between… Dunning-Kruger where they’d done a bit of comp sci stuff, and they thought that they knew enough to do it, they’d taken intro CS classes and intro data visualization graphics classes and they thought that would see them through. And I think their other plan was to pull in a technical person and have the technical person do that side of things.

Alyssa Vance:
Then what happened next? Where did you go from there?

Fred Gardner:
We spent the next three or four months building. The first hint of unusualness that went on, was in the winter of that year. We applied to Y Combinator, we got the interview, and we had another potential investor/customer meeting that was the same day that we couldn’t move. I suggested that I and a couple intern style people go and handle this customer meeting, and that the other founders go to the YC interview, and that we tell YC “oh, by the way, we have another person, but he’s meeting with a potential customer/investor”.

The YC interview itself went well enough that we were accepted. But apparently, YC was very confused about the possibility that someone would have a customer meeting at the same time as a YC interview, and not move it. This was the first hint of strangeness. But after that, I forgot about that weirdness and things proceeded apace. We were building out the product, meeting with investors, meeting with customers, et cetera, until we got to that summer and went out to California to do YC.

Alyssa Vance:
So you were accepted into Y Combinator.

Fred Gardner:
Yep. We were accepted into Y Combinator; we went, and then we were informed that only the CEO, COO and CTO were going to be considered founders by YC, and that therefore only they’d be able to attend the majority of the programming.

Alyssa Vance:
But there were two people who came up with the initial idea, if I’m understanding correctly. And there were more than three people who had relatively equal-ish shares of equity. So how did YC pick those three?

Fred Gardner:
They were the three who had gone to the YC interview.

Alyssa Vance:
That seems kind of arbitrary?

Fred Gardner:
My assumption is that they were the three who had gone to the YC interview, and therefore, when YC had written down the three people that they were accepting, presumably, they’d only taken notes on those three. I’m not going to pretend to know what YC’s precise process was there. But my assumption was that somewhere along the way, something had gotten lost in the shuffle or not checked carefully, and it was just like, “Oh yeah, these two people are the ones who showed up for the interview. They’re the founders.”

Alyssa Vance:
Okay. Did you ask them about this?

Fred Gardner:
Oh I did. I asked them about this, a couple of weeks later, and they were like, “Sorry, this is not a thing we can fix. These are the people we accepted into the program, we didn’t accept you as a company.” I don’t know whether this is a truthful thing, or whether this is like… we want to save face because we don’t want to make it look like we didn’t do enough research into the company before doing the acceptance.

Alyssa Vance:
Y Combinator is obviously a very competitive investment program. I personally applied to Y Combinator, I got an interview, but then got rejected afterwards. They have an acceptance rate somewhere in the low single digits. Do you have a sense of why this company was accepted, whereas the vast majority of other companies were not?

Fred Gardner:
I think it’s a mix of two things. The thing that I was told by the people who were at the interview, was the interview was not going well, and then Y Combinator was very impressed by the demo of our tech, which I and a couple other people had built out over the past couple of weeks. So it’s either they were impressed by our tech, or… it’s probably some mix of that and somewhat credentialism, like, “oh, we have a bunch of people from Harvard and MIT on the team.”

Alyssa Vance:
At this time, were you getting investment or money from other sources?

Fred Gardner:
Yes. YC was not our first investor. We had a few angel investments from people that the CEO and COO knew. We finished our seed raise within the first couple of weeks of YC, and we started our Series A or secondary seed raise – it shifted over the course of our post Demo Day fundraise – maybe two weeks before Demo Day.

Alyssa Vance:
The main YC program lasts about two to three months, and the YC partners say that during this time, people are supposed to be super productive, and essentially do nothing but develop their product, get feedback from customers and sell it to customers. Is that your experience, your sense of how that went?

Fred Gardner:
The main difference in productivity, I think, doesn’t come from the fact that it was YC, but comes from the fact that we had gone from working on it part-time to full-time. I didn’t notice an extreme difference in productivity that can’t just be explained from going from working on something 40 minutes to an hour a day, to working on something seven hours a day. We were probably 7 to 10 times as productive as we had been previously, and this is explained by the 7-10X work time.

In terms of selling to customers and meeting with customers, we met with a few customers early and then a few more customers later. I talked a couple of times to YC partners when they let us in for full team meetings, and heard and read the notes that people took on YC meetings, and the advice we got was very wishy-washy, and was much less “here’s how we can help you talk to specific people”, and was much more just like, “Oh, go out and talk to people, We can’t help you until you’ve independently done that.”

Even after we’d gone out and independently spoke to people, they kept giving us the same advice, except for one particular case where a YC person tried to be helpful and inadvertently made a mistake that did not speak well for their competence.

Alyssa Vance:
Can you tell us roughly what size of the customer were you targeting?

Fred Gardner:
SMBs. The smallest customers we targeted were people who were running tens of thousands of dollars a year. The largest vertical of customer we targeted were large-ish retail providers who were doing hundreds of millions or billions of dollars of business. We targeted every type of customer within the range.

Alyssa Vance:
So you said that you weren’t superhumanly productive, but would you say things were basically going as planned?

Fred Gardner:
In terms of product productivity? Yeah. We were pretty good at consistently meeting roadmaps.

Alyssa Vance:
Then at the end of a YC, you did raise money successfully?

Fred Gardner:
At the end of YC, we raised money successfully, but from contacts that we’d had before YC, for the most part.

Alyssa Vance:
So you participated in the famous YC Demo Day?

Fred Gardner:
We participated in Demo Day, got two or three inbound pieces of interest. And I think none of those turned into actual investment.

Alyssa Vance:
Demo Day involves… I don’t know, but I was guessing around 500 investors, is that right?

Fred Gardner:
Yes. It involves like 500 investors. I don’t know how many of those were initial bites. All I know is when things moved far enough along that the CEO and COO shared their progress with the rest of the team… and that was like maybe two or three contacts that were made on Demo Day proper turned into people who were potentially interested in participating in a round.

Alyssa Vance:
It sounds like – at least from your experience, maybe others are different – that the hit rate from this was lower than the hit rate from just cold emailing people.

Fred Gardner:
I wouldn’t go that far, because the two or three people that I knew about were actually in talks for, at what valuation they’d be interested in participating in a round. So it’s probably about 10X the rate from cold emailing, but it’s significantly lower than the rate from already knowing some people. And the rate for Demo Day was worse than the rate which we built up by just being in the Bay Area and meeting with customers, and having those customers point us to potential investors.

Alyssa Vance:
Your customers pointed you to investors?

Fred Gardner:
Some of our customers were well off. Some of our customers had Corp Dev-style things that were interested in potentially going into rounds. In particular, one person who we were hoping to bring in as a high profile customer did end up participating in the round.

Alyssa Vance:
So what was the mistake that you mentioned earlier?

Fred Gardner:
Ah, yes, of course. As often happens with companies at YC, we were in stealth mode until we were running a launch. We were talking to one of the YC partners, and expressing our disappointment in the fact that they had not helped at all put us directly in touch with customers. And the YC partner was like, “Oh, my co-founder at a company is in the same vertical as you guys, let me put you in touch with him.” So we were in touch with him and he seemed very interested in our product. He was asking us all sorts of details about our tech, and all sorts of details about what we were doing, and we were like, “Oh, this is a wonderful conversation.” We were really impressed with her until one of us looked her co-founder up on LinkedIn after we’d had this conversation, and we realized that he was a board member at our largest competitor.

Within 45 minutes of this conversation ending, the CEO of that largest competitor sent a feeler email to get a sense of whether we were actively competing with him, and whether we were making claims. Over the course of that conversation, we made claims about failings in that competitor’s tech, as one does when you’re trying to sell to someone who’s like, “Oh, how are you different from these guys? They already exist.” So we were like, “Oh, their tech fails at X, Y, and Z.” He sent us an email pretty much saying, “if you continue making libelous claims about my technology, I’ll be forced to serve you”.

Alyssa Vance:
So he threatened a lawsuit?

Fred Gardner:
He threatened a lawsuit. And part of the reason we were in stealth mode, is that the company we were competing with already had a million dollar run rate, and could therefore afford a lawsuit. So this necessitated a large amount of damage control on our part. And then we went to the YC partner, and were like, “What’s going on?”, they were like, “Oh, sorry, I guess I probably should have looked up this person on LinkedIn before sending you guys a meeting”.

Alyssa Vance:
Okay.

Fred Gardner:
Yup. And this was not one of the newly brought in YC partners. This was someone who YC had referred to extremely positively in their external things, as one of the shining stars of their advisorial capacity.

Alyssa Vance:
So at the end of this, you’ve successfully raised an investment round. How large was the company at that point and after?

Fred Gardner:
The investment round was happening as everyone technical was leaving the company. That was as the company was transitioning from six down to two.

Alyssa Vance:
Why was everyone technical leaving the company?

Fred Gardner:
Unbeknownst to a lot of the technical people, over the course of the summer, YC was helping the founders reincorporate the company to drastically reduce everyone’s equity shares. The CTO went down from 15% to 10%. I went down from 15% to 4%. Another person was fired at the start of the summer for unrelated reasons – he actually had bad job performance and use of pharmacogenics. And the principal engineer went down from 8% to 2.5%.

Alyssa Vance:
Who got all the rest of the equity?

Fred Gardner:
The founder shares went up to 40% and 35% respectively. The two co-founders.

Alyssa Vance:
Why should they get all the rest of the equity?

Fred Gardner:
They did not give a justification for this. The closest they came to a justification was because they had taken a semester off previously when other people had not taken a semester off. And they had given some explanation of signaling, where if they didn’t have that much equity, it wouldn’t signal that they have control of the company. And they intimated that they were told this directly by YC investors.

Alyssa Vance:
Why shouldn’t I have the equity? I want equity.

Fred Gardner:
We had spent the summer being told that we were going to renegotiate our employment deals to account for the fact that the company had funding. We assumed that this meant that we would be given higher salaries. It turned out that what this meant was that our equity amounts were being drastically reduced.

So I was like, “Oh, this seems unreasonable”, and I didn’t think that YC would have been party to this. I sent an email to YC saying, “Hey, so this happened, and you guys are large investors in the company, and invested in it I assume in part because of the tech. And I think you would like to know about this.” The response I got was a diplomatic version of, “That’s really unfortunate, but there’s nothing we can do. And there’s nothing we would like to do, because any sort of struggle like this would destroy this company.”

Alyssa Vance:
Did it destroy the company anyway?

Fred Gardner:
Where the company is now, is they seem to have metamorphosed into a much less software-requiring version. In particular, I think that the current vertical was served perfectly reasonably by non-software versions of the same thing that relied on a lot of human labor; I think they have converted into an SMB that is human labor-powered in their vertical.

Alyssa Vance:
What happened to all of the engineers, and what happened to all of their technology?

Fred Gardner:
The technology they still have – there was patent pending technology that I had helped design that the company and I are joint authors on, I have no idea what happened to that patent. The engineers are… one is at Facebook, I’m currently at a different company, one is at – I forgot whether he’s at MIT or Stanford for grad school, but it’s one of those two – doing AI research. They picked up one person after we all left, who’s now the CTO. And the person who was fired at the beginning of the summer has now started a different social networking company.

Alyssa Vance:
What do you think you’ve learned about entrepreneurship?

Fred Gardner:
I’ve learned not to go into business with someone who I don’t really deeply trust beforehand, and also, that someone not being fully upfront with the exact process of any investor/equity negotiations is a massive red flag. I’ve learned a bunch of lessons about trust that are good to know for the cost of a summer. One additional thing I’ve learned is that if the reason you’re interested in in a service is intense personal attention and really high quality of educational interaction, maybe when that service expands 20X in eight years and has switched leaders twice, it probably no longer has the same quality of service and attention to detail.

Alyssa Vance:
Are you referring to your company’s service, or Y Combinator’s service?

Fred Gardner:
To Y Combinator’s service. That much of YC’s change in ethos, and in particular, the change of ethos to what matters being the YC stamp of approval versus the actual help YC gives, can be traced to the fact that YC has switched leaders twice, and has switched leaders to people who are relatively less tied to YC at its inception.

Alyssa Vance:
What does the YC stamp of approval get you?

Fred Gardner:
The YC stamp of approval I think… even now, it means you get past the initial rejection stage when you’re cold emailing people. Also, YC’s whole claim is that it’s significantly easier to be a B2B startup coming out of YC, because YC natively gives you this whole customer base of people that will purchase your product, and this does seem plausible though I have no direct experience with it. But it feels like YC used to be about a lot more than, we have a large network of people that have already done our service, and used to be about a lot more of, people telling people… pulling you into a room and giving you consistent concrete advice about what worked and didn’t work in your vertical. And it seems a lot of that has fallen by the wayside.

Alyssa Vance:
Paul Graham talks about in his essays… if you have a company and if you get the YC stamp of approval – or even going beyond that, if you are like Theranos and you raise hundreds of millions of dollars from the famous investors of the world – and then you don’t make money, or in Theranos’s case turn out to be a complete scam – you haven’t actually won anything. You are still left with nothing, or in Theranos’s case with federal criminal charges. So what is actually the benefit to you?

Fred Gardner:
I think there are two things here. One is that the YC stamp of approval probably makes it easier to get acqui-hired. The issue with “if you don’t actually make money, you have not won anything” is that that doesn’t apply if your company serves as an extended interview for a big tech company with a large signing bonus.

The second thing is that… one big thing that I can’t believe I forgot to mention about YC, is that there was one interesting thing that we noticed: in their words, in order to accommodate large cohort sizes, they decided to partition the cohort into several small groups. Now, while this was never officially confirmed, we did note that… the groups were partitioned by vertical, but also the groups were seemingly partitioned by which companies the YC founders thought had the best prospects. And that there was a mysteriously large quantity of people in group 1A who raised large amounts in their series A coming out of YC, and also those people met with the partners who had been there the longest. Then as we went down to lower and lower groups, these people got less time with partners and raised consistently less coming out of YC.

So I see two options. Either YC was A/B testing the necessity of partners, or YC has already pre-guessed what companies they think are going to be the most important to give help, and partitioned them with the cohort sizes based on that. Now, this is a bit of a tinfoil-hat theory and I’ll acknowledge this. But see if you can find any information… this is going to be hard to find without being on the internal YC forums, but if you can get your hands on any information about who is in what cohort, just see, of the people who raised large amounts coming out of the W20 and W19 and S19 Demo Days, which ones of them were groups 1 and 2, and which ones of them were in groups 8 and 9. I think that the correlation would be a better than chance.

Alyssa Vance:
There are thousands of people who have been through YC and would presumably have access to this, right?

Fred Gardner:
Ah, yes.

Alyssa Vance:
And a large percentage of them know how to use data science tools and SciPy, right?

Fred Gardner:
Yes.

Alyssa Vance:
So I’m a bit surprised if someone hasn’t already done this.

Fred Gardner:
I would believe that someone had done this. I guess my question is, what are the incentives to doing this? Why would someone have done this?

Alyssa Vance:
Understanding how your corner of the world works?

Fred Gardner:
Okay, I would not be surprised if someone had done this. I would be surprised if someone had publicized this. If you ever hope to do well in the Valley, there is a certain amount of “people not thinking that you’re conspiracy-theorizing about them” that needs to happen. If you publicize something, especially in a way that you can be tracked back to you, that’s like, this well-regarded accelerator is – even though it’s accepting everyone – ranking people based on who they think is going to be successful. This seems like something that would ruffle a lot of feathers.

Also, I have reason to believe that this is something that only started very recently, in response to the large expansion of YC cohorts. They did not have eight groups of cohorts in 2016. They had like two, I think that this is like… YC expanded what, like 8X in the past four years, something like this? Maybe 4X in the past four years? I think that this is only something that you a) want to do and b) need to do once you’ve expanded above maybe a hundred companies. And this, by their own admission, only happened in like 2017.

Alyssa Vance:
So the next question, if Y Combinator were trying to rank companies in advance, did those rankings have any meaning? Did they predict anything?

Fred Gardner:
I think they are predictive of how much the companies will raise coming out of Demo Day, and possibly how likely the companies are to face quick exit. I am not convinced that those companies are predictive of how successful they are likely to be, or how likely those companies are to end up on YC’s top valued companies. In particular, because if the cohort things are rankings, they must be happening before the YC partners have seen your work for 90 days – they must be based purely on your impressions in a 10 minute interview, which means that they’re going to be probably pretty terrible metrics for how likely these people are actually to succeed.

Alyssa Vance:
So after the initial 10 minute interview, how much time is spent between YC people and your company’s people?

Fred Gardner:
One or two meetings a week. This is after they’ve already cohorted you – between the 10 minute interview, and when they’ve accepted you, it’s zero time. After they have cohorted you, the YC program is one or two meetings a week with a partner. They have their Tuesday speakers where everyone comes in for like a couple hours on Tuesday, but you don’t really get one-on-one time with anyone. Then you have a whole group meeting once or twice a week. And then in theory, you schedule more time with partners, but that is a) very dependent on company and b) very dependent on how involved your partner wants to be with you.

Alyssa Vance:
Did you have a chance to speak to other people in the YC community?

Fred Gardner:
I did, there were a couple of YC parties that we went to, especially after Demo Day. It did not seem like YC was particularly more selective in terms of acumen, than the overall elite college population would be.

Alyssa Vance:
If you imagine Harvard or something, the amount of one-on-one time you spend with professors is normally pretty small. But then Harvard is this large community, and within that, they form all of these sub-communities and people get a lot of social connections from Harvard. How much of that do you think there is within YC?

Fred Gardner:
I experienced pretty much none of it. This might be an artifact of my somewhat outsider-ness. This also might be an artifact of, there wasn’t really a great native way… Harvard takes effort to make sure you form into your own communities. YC, as far as I could tell, did not take the same effort, and they sort of wanted everyone to be part of the one large YC.

Alyssa Vance:
Of course, that doesn’t scale.

Fred Gardner:
Yes. I don’t know if YC has taken… they may have changed, I don’t know if they’ve actively made steps to build sub-communities within the communities. One additional thing that would make this significantly harder for YC, at least for now, has part of their ethos being, “you spend the summer building something”, and spending your time building something means it’s very difficult to have the organic sub-community-forming interactions that you can have at a place like Harvard, where it’s not “spend nine hours a day working on one thing”.

Alyssa Vance:
If someone gave you a billion dollars and you were trying to design your own startup investment program, how would you do that?

Fred Gardner:
There are two options. Option 1 is, don’t scale – be what YC was trying to be circa 2011 and don’t scale at all, or scale up to like 20 and that’s it. Or 2, scale as rapidly as possible, and make my asset the fact that I’m capable of very effectively connecting people. I think that if YC threw themselves into a Hacker News-type community of sub-communities for startups, I think that this would also function better than their unhappy medium where they’re trying to be vaguely personalized attention, and also trying to be, “we’re a massive far-flung network,” and not doing either particularly well.

Alyssa Vance:
Will you apply to Y Combinator for your next company?

Fred Gardner:
I think I’ll apply, if nothing else because from what I can tell, the experience of the interview is extremely valuable.

Alyssa Vance:
Even though it’s only 10 minutes?

Fred Gardner:
Yes. Because it’s 10 minutes of being attacked on every facet of your company business plan, and being forced to consider holes in your ideas that you might not have considered.

Alyssa Vance:
Could someone sell that as a service, get a bunch of experienced people together and attack companies for a fee?

Fred Gardner:
The problem with startup red-teaming is that the companies at the stage that you’re doing that to, are not going to have the money to pay for it. But if there was a way that I could get consistent access to this thing, this is something that I’d definitely be interested in.

I will say that, to be fair to YC, they find themselves in a very difficult situation of having a valuable service – or having initially had a valuable service – and having a very, very large crowd of investors who want to help them raise their next round, and thinking that their service was valuable enough that they should spread it more widely. I just think that they didn’t plan well enough how to make sure that they were actually preserving what was valuable.

Alyssa Vance:
I know ordinary businesses that are in a situation like this. It’s like, okay, our product is hard to scale. There’s a lot of demand for it. We charge a lot of money. Also, our customers are annoyed because we’re not paying enough attention to them, because there are too many of them, even though we’re charging a lot of money. So what should we do?

Fred Gardner:
This is going to be more mean than I intend, but I can’t resist giving a quippy summation of what my beliefs are about YC – Paul Graham has a wonderful blog post called “Do Things that Don’t Scale”. I think YC read that post a little bit backwards, and read it as “Don’t Scale The Things You Do”.

Alyssa Vance:
But they say they scaled everything that they do. They say, at least, that every part of the program they have scaled up, so they can successfully run 300 companies per batch.

Fred Gardner:
Yes. So the issue is while they have in fact hired more partners, and pulled in more people to do interviewing, and have more people running cohorts… there is a difference, even if it’s the same proportions, if it’s Paul Graham in a room with Alexis Ohanian and Justin Kan and the Collison brothers and a few other people, that is very different from 40 people in 40 rooms with 300 companies, regardless of the fact that the proportions remain the same.

Alyssa Vance:
All we have to do is take the Collison brothers, and clone them 40 times.

Fred Gardner:
Yes, exactly. I think that the best way to do this is a simple piece of data analysis. Right now, YC has forty times more companies in the 2020 batch than they did in that first 2005 batch. They’re arguing their standard of quality has not gone down at all. Look at YC’s top companies by valuation, look at how they are scattered by year. It is approximately the same number from each year, barring the first year they did YC when they only had eight companies.

From the point where they’d scaled to like 20 companies, it’s the same number of companies from each year in that top valuations bar. This does not seem suggestive of the level of quality being the same at every year of YC. This seems suggestive of, there are 10 to 15 great companies in every batch, and they’re just adding more and more companies that are not going to be top valuation onto that scale.

Alyssa Vance:
How would you decide in the future, whether people you want to start companies with are trustworthy or not? Would you consider me trustworthy?

Fred Gardner:
The biggest thing that I’d say is trust but verify. If I’m convinced that you are not the sort of person who will directly violate verbal agreements – especially as they pertain to privacy – for something like founding a company, I will make sure to get evaluations from people if I’ve directly worked with the person, or evaluations from people I trust and know are good at evaluating people.

Alyssa Vance:
How do you know if someone is good at evaluating people?

Fred Gardner:
By seeing whether their evaluations of people pan out, and are accurate, and the way those people act in the future. At some point, this does bottom out at the problem of induction, right? But… the standards I used for deciding to join this current company was, I talked to a bunch of people that I knew well, and a friend of mine was currently at the company I’m currently at, and I had a much better experience. I think I’ve been vindicated thus far in my decision-making. For this current company, I had friends who both knew people who worked on this company and were directly working on it, and gave me much more direct positive reviews. That’s the sort of strategy I’ll be following from here on out – either that, or directly working with people who I know.

Alyssa Vance:
Is there anything that you’re especially excited about for your future companies?

Fred Gardner:
I have a bunch of ideas relating less to fully new technologies, and more to… My favorite kind of technological innovation is, two separate people have each solved one half of a problem, but neither of them knows the other half of the problem exists, so they don’t realize they could combine their innovation with someone else’s innovation to actually solve the whole problem. There are a couple of verticals that I’m interested in here, the primary one is privacy-preserving data analysis, but a secondary one is fairness in machine learning, where there’s a two-sided market where one side has been, “Oh, we can guarantee fairness under X conditions, but not without X flaw; and the other side is, we can guarantee X flaw isn’t going to happen, but we can’t guarantee fairness.” A lot of those sorts of things are on my mind these days.

Alyssa Vance:
Cool. Anything else you want to mention?

Fred Gardner:
Nothing comes to mind. One small thing is, I think that part of the reason that startup accelerators look different now than they do before, and VC investment looks different now, is that in 2008 the acqui-hire was not as big a thing as it is now. I think that that has shifted the industry, such that people in particular technical industries don’t feel as strong an urge to ensure their product is profitable.