When you interview a startup and think "they seem likely to succeed," it's hard not to fund them. And yet, financially at least, there is only one kind of success: they're either going to be one of the really big winners or not, and if not it doesn't matter whether you fund them, because even if they succeed the effect on your returns will be insignificant.

What this means is that YC is not looking for sustainable businesses, but homeruns. Which is entirely fair, that's the business they're in.

But you and your startup are in a different business: Your measure of success isn't the same as Ycombinators. If your startup ends up making you a million dollars a year you will probably be very happy and rightfully call yourself a success. But as the post points out that won't be enough for YC since they need to fund a lot of other startups that will inevitably fail out of their minority share. Thus they need a much bigger success.

If you get turned down for YC it might well be that your idea is just a sound business idea that YC doesn't consider just crazy enough that it might make them a billion dollars. But that doesn't mean that it won't make you a million.

Despite the rejection we are growing each month, turning down acquisition offers, our customers love our products, and at 24 years old, we are learning a ton. We have no investors, are profitable, and have a lot of time to grow the company and work on crazy ideas. I am having the time of my life.

It's really important to understand the business YC is in, and to separate your self-worth from it. It's also important to not hold any negative feelings towards them, since they are merely trying to maximize their outcomes, and they know the market quite well.

Go follow your passions and make stuff people love. The rest will follow.

> I'm not saying that the big winners are all that matters, just that they're all that matters financially for investors. Since we're not doing YC mainly for financial reasons, the big winners aren't all that matters to us.

Also,

> If we ever got to the point where 100% of the startups we funded were able to raise money after Demo Day, it would almost certainly mean we were being too conservative.

(and the entire rest of that section talking about how they can afford to take huge risks, and should be.)

On the whole, this piece sounds like some observations and thoughts from pg, and nothing more. I think founders shouldn't try to read too much into it, specifically whether or not YC would be interested in their business.

YC might choose to fund you just because they like you.

I see this a lot from various corners of the startup ecosystem, particularly 37 Signals and their followers. The problem is that it's not really true, by which I mean there are not very many examples of it begin true, and there is an excellent reason to believe that it may never be true. Which is: you need to find a market big enough to support "lifestyle income" (or your million dollars a year) but not big enough that a startup or growth technology company that is really good at doing things at scale isn't just going to eat it, and kill you in the process.

One of the reasons Jason Fried needs to yell so loudly about 37 Signals being the model for lots of other companies is because it's really not -- it's really rare to find a lifestyle technology business.

The traditional challenge you'd level here is "Name three" (37signals, Fog Creek, Balsamiq) but, due to the type of people I hang out with, I could get to fifty before having to slow down and start checking my Gmail. A friend of mine who is in the selling shovels business estimated that there are 30,000 firms selling SaaS. (That number struck me as crazy until I realized that, oh yeah, I'm routinely in rooms with several hundred of them at once.) The overwhelming majority will never raise outside capital.

So I'm just curious if Patrick was really meaning that he can without pausing name 50 non-VC backed lifestyle product companies that are making $1M in salaries and profits for their founders, or that he meant that he can name 50 companies that are generally well-off, and are making e.g. a few thousand dollars per year for their founders.

The difference is important in my opinion, as I can name a several that make a few thousand dollars a year by consulting, but I don't know people that make over a million a year for themselves with an internet product without an investment. Those that I personally know that earn $1m, are in more fishy type of business (E.g. quick SMS loans) and have a sizeable financial backing from more traditional investors.

Most companies that build and ship product can easily consult, so, any of those companies that continue to ship product for multiple years should cause you to ask how much more than $1MM they must be making.

"Salaries and profits" is an awfully weird metric, since salary is the #1 cost factor both for consultancies and product companies. Maybe you should just say "revenue".

I respectfully suggest that your radar is off here. No, you don't need to be in "fishy types of businesses" to break $1MM.

But that's not what founders themselves earn. Founder's personal salary and a slice of pure profits that are not reinvested to company growth can be considered a total that founders "earn" in lifestyle businesses.

Even with this metric, of course there is a lot of lifestyle internet companies in the world that do over $1M per founder. I'm just interested if Patrick meant that he can name 50 companies (presumably from his network) that are doing this well.

I worked for consultant who had a lifestyle business. He pulled in nearly half a million dollars in revenue per year, but I got paid a higher salary than he did. But I drove my own car, and he drove a company car. He didn't own a computer, but the company had a fiber connection, a server room and several very nice recent laptops. He didn't own a cell phone or a camera, but the company supplied him with a nice world smart phone and multiples of the latest cameras. He rarely took vacations, but the company paid for him to travel the world to visit clients and trade shows. Etc.

It's for this reason that revenue is often a better gauge of success for a lifestyle business, than profit or salary.

You... must live in a different world from me. I mean, I am a lifestyle "technology" business. Most of my customers are, too... a whole lot of them are so small that they still have dayjobs. For that matter, most of my suppliers are, too.

But think of all the web design firms that exist. All of the small-business IT firms. The small consulting shops of various stripes. There are huge numbers of these "too small for important people to care" companies. and so many web applications are thrown together by one person, just messing around.

Many, probably most of my suppliers and competitors are also small operations owned by one or two people. Linode, as far as I can tell, started a lot like I did; and so did most of my smaller competitors. (Slicehost is the counterexample; my understanding is that they started in a very startup-y manner.)

I mean, obviously, you get fewer firms as you raise your revenue cutoff. If you require millions of dollars a year in revenue, nearly all my customers fall off the list... but actually, probably not that many of my suppliers. I mean, I'm buying one rack from coresite, 5g from cogent, and 1.1G and 2 racks from he.net, but that's less than half my monthly outlay, and other than that, all that money goes to small private companies (and really, he.net might be considered a small private company that grew to be a not-so-small private company.)

Maxed out their credit cards, tried to find angel funding in St. Louis, and got told the going terms were "We'll get 50% of the company to cosign a loan for you" so they continued to bootstrap, managed to successfully structure a pricing model such that customers pre-paid for services (allowing them to service most of demand), rode on to acquisition by Rackspace... if I recall correctly. (Pours one out for Slicehost.)

They did seem to grow... very quickly. My impression was that they grew much faster than linode or I. Of course, I have no numbers to back that up either way.

http://37signals.com/founderstories/slicehost

I relate to "I have a limited shelf life in a big company" I mean, I'm literally 1/10th the size of slicehost when they were bought, but I have gotten a few companies courting me for a buyout... The thing is, at this point, they want the company for me, and it's pretty clear that I... well, I might stick around for a few years, but working for someone else just isn't what I want to do.

I mean, I can be bought, like anyone, but I imagine that because I know this about myself? I kinda think that my attitude would scare off anyone that might buy us. I mean, to save time, I've actually come up with a formula for how much money I'd want; some lump sum dependent on the size of my company, plus another large sum for every year they want me to work for them, plus another (smaller) sum for every year they want me to avoid competing with them.

I dono. This was mostly to save time; I was hit with two courtships within a relatively short period, and felt I was spending too much effort on it, but eh... I dono. I haven't gotten other nibbles for a while, so it's likely that it's rare enough that time spent on it isn't a huge deal.

I've always taken Linode as the model for where I want to take my company; of course, I've gone off in kindof a different direction; I want to own a network, and a datacenter, while Linode has preferred to have multiple locations. (I mean, it's one or the other.)

With regards to the new knowledge that 20k customers is possible, that just gives you something to shoot for next, right? There's something you can do that starts at 2k users today and gets you to 20k users. You know that is possible, because other people have done it. So work on that. (Knowing from previous conversations that you're rate limited on customer acquisition by processes other than marketing I'd suggest, as the Slicehost guys did, "automating the "%#)(% out of everything" and then focusing on customer acquisition as your primary job, if that is a huge priority for you.)

More broadly: philosophically, if we consider ourselves failures for not doing things other people have done, then we'll all be failures essentially all of the time. That strikes me as an unhappy bit of philosophy to adopt then.

Well, yes. Mostly I'm just amazed how they did that so quickly. That is admirable. But, I guess most of my growth has been in mad dashes, too... mad dashes followed by lulls where subscriptions keep pace with cancellations. I can see how it'd be possible to keep up those mad dashes, if your automation was better to start with, and if your support was better, and the marketing. All these are things I can improve; but it's admirable how they absolutely nailed all those things in such a short period of time.

>I'd suggest, as the Slicehost guys did, "automating the "%#)(% out of everything" and then focusing on customer acquisition as your primary job

Yeah. This. I mean, I make the excuse that I'm too busy 'shooting alligators' but you really have to set aside time to drain the swamp (and let's be honest here, I've started a bunch of different projects... time that would have been better spent on draining the swamp by automating more of this stuff.) I did actually make a lot of progress automating the killing of old accounts the night you wrote that. (It's still a manual process, but I wrote some scripts that took it from being an all-day thing to clear out everyone to it being a 10 minute thing to clear out everyone.)

>More broadly: philosophically, if we consider ourselves failures for not doing things other people have done, then we'll all be failures essentially all of the time. That strikes me as an unhappy bit of philosophy to adopt then.

The balance between arrogance and considering oneself a failure is a tricky one. I mean, I know that the only thing that has come close to destroying me was that confidence. (Well, after getting out of school. That's a long story; but there certainly were plenty of strong voices saying that I can't do it then; they were all, of course, trying to get me to put in the effort to get into a reasonable college, the upshot, of course, being that nearly everyone told me I was going to work at 7-11 if I didn't go to college, and it was obvious that I wasn't going to do so. I think a lot of my arrogance is innate, but a lot of it also comes from just how easy the real world was for me, compared to school.)

But yeah, PG wrote about confidence in a way that I related to, perhaps better than anyone else I've read:

"But you should treat your optimism the way you'd treat the core of a nuclear reactor: as a source of power that's also very dangerous. You have to build a shield around it, or it will fry you."

That line, really, is why I got interested in his writing and in this community. I've gotten burned badly, several times, by raising the control rods too much. I know that for most people I know, their big mistakes usually have to do with hitting the metaphorical SCRAM button before the reactor even gets warm, but for me? The big problems have always been that over-reach; the over confidence.

Of course, Graham goes on to say:

"The shielding of a reactor is not uniform; the reactor would be useless if it were. It's pierced in a few places to let pipes in. An optimism shield has to be pierced too. I think the place to draw the line is between what you expect of yourself, and what you expect of other people. It's ok to be optimistic about what you can do, but assume the worst about machines and other people."

which, of course, I believe is Very Bad Advice if you are a flawed individual. Self-doubt is essential to survival. Personally, I like the control-rod metaphor. When things get too hot? you lower the rods and let things cool down. Need more power? raise the rods.

But always be aware that while having no power is a problem, it's easier to clean up after a SCRAM than a meltdown.

Of course, much like how it may be best for the military to select fighter pilots with irrational levels of confidence, I can see how it would be best for an investor to select founders with irrational levels of confidence. If they try hard and fail? eh, well, there's always another one waiting on the sidelines. And in both cases, there's a supervisor that tries to make sure they don't get in too much trouble.

When it's just you, well, you've gotta consider the consequences of failure. And sometimes the consequences are small; if you have a corporation, and the credit is all in the corp's name? sure, take big risks and fold it if you screw up. But, make damn sure ahead of time that is possible.

Those mistakes (of overconfidence in cases where the consequences of failure were non-zero) have cost me years. And nearly all of the strategic mistakes I make now are attributable to the same source. Most of my tactical mistakes are attributable to, well, my brain not functioning normally. whatever you want to call it; attention deficit disorder; being lazy, having difficulty focusing? whatever you want to call it. It does respond fairly well to a battery of only moderately unpleasant medications, so I can't complain too much. - but the thing is, I know that about myself. It had a lot to do with why I didn't do well in school. I should be able to include this margin for error in my planning; but no. that confidence takes what I can do on the days that everything is just right and plans using that; forgetting that even with medication, I get maybe three of those days a week, when I'm really paying attention to all the inputs, working out, eating right and sleeping well; Otherwise, three of those days a month is a more realistic estimate. And I really am effective on those days; even if I planned for three of those effective days a month, I seem to still meet bay area sysadmin standards, so with half a decade, I ought to be able to come up with what is maybe ten days of good solid work on this problem.

I mean, I personally think that the biggest thing holding prgmr.com back the most right now is the automation, and I think the primary thing holding that back is that I've been spending my good days on other projects... that yeah, are interesting and might be worth something at some point in the future, but ultimately should take a backseat to the butter and guns.

Do you have a "business guy" helping you? Have you ever? If not, why not? (Never met one, didn't know you needed one, etc?)

There was one guy that I turned down that I probably shouldn't have turned down. But he was more technical than I am, too. That was probably a mistake, but he's working on his own thing now.

The thing is, nobody else would have had the follow through to keep bleeding money on the thing for as long as it took me to figure out what the hell I was doing. I mean, would that have been better? maybe. Almost certainly my lifetime earnings up 'till now would have been greater. Or maybe a business person would have kicked my ass 'till we came up with a solution that kindof worked even though I didn't know what I was doing? Or maybe they would have come with enough money that I could have compressed those lessons into a shorter timeframe? (Most of my early mistakes had to do with hardware and the cheaping out on thereof. With sufficient capital, I could have moved on to making other mistakes earlier.)

But really, having a business guy would have been counter to most of the reasons I went into business in the first place. I thought I could do it better. If I really am wrong, I need to accept my place and get a job.

The thing that is holding you back is that you are very honest and conscientious. All you have to do to determine that is to read your page on "colo" at your site where you are very frank about the fact that the rent might go up. That is very atypical in business. I've seen the same ethics hold many people back. I'm not saying at all that everyone who makes it big is unethical because of course that's not the case. But I've seen many people who haven't made it who are just to damn honest and I've seen many people make it that are just unethical and easily pull the wool over people's eyes.

Now of course everyone fibs a bit here and there. It's not like a person is either totally honest or a total cheat. There is a gray area and it's a continuum.

Anyway that's my take on things. I can point to the obvious example of Jobs and Woz (of course Woz had no business sense but the fact is he really didn't even know it was possible to be screwed by Jobs so he couldn't even protect against it.)

I think you have a great thing going. Personally I feel you should move away from the hacker market and move more mainstream and raise your prices. Or perhaps develop a different site to address that market. If you are interested, I'd get involved in that possibly.

I don't know if that's true or not, on either side, but it's certainly a great compliment. I definitely want to be seen as honest to a fault. (the counter argument is that I do often over promise. I do not always temper my confidence with enough reality. I'm not trying to be dishonest in those cases, but the effect is the same.)

Really, though, saying "I'm not successful because I'm too honest" sounds a little too much like sour grapes for me to accept it as the truth. I mean, I'm a cynical person, but it just doesn't ring true. Markets are more complex than that. Yes, there is a constant arms race between buyer and seller, an arms race of deception, but participating in that arms race is very expensive. I believe this is a lot of why "cloud hosting" and other transparently priced services are so much more expensive than colocation. People are willing to pay extra to avoid that expensive bullshit. (the other reason is that it's so easy to raise prices on hosted products... I will explain near the end of this message.)

Well, and more to the point, I'm mostly doing this for fun. I can make a lot of money working for other people, I mean, before I was published I was working at what I consider absolutely ridiculous rates. there's way less upside risk; but making bay area computer nerd wages is plenty comfortable. If I have to do things that make me feel bad to work for myself? eh, I'll just go back to working for other people. I mean, I'm not saying that I wouldn't do something I felt unethical if I were starving or I couldn't provide for my family or whatever... but I'm not in that position. My 'plan b' in case of failure is downright cushy.

>All you have to do to determine that is to read your page on "colo" at your site where you are very frank about the fact that the rent might go up.

Co-location is run like commercial leases; e.g. when it comes time for renewal, the amount the rent goes up is determined by how hard the landlord thinks it will be for you to move. You can have empty units on either side of you and the landlord is gonna say "Yup, market rents have gone up!" if the guy knows that you've put down roots and moving would be difficult.

The thing is, co-location, by it's very nature, is difficult to move, just like commercial real-estate. And if you say it's extra hard to move? well, I told one provider that it would be difficult for me to change IP addresses. Within a week (and this was about a month after I signed a 2 year contract /after/ feeling like they screwed me while I was on a month to month, thinking the contract would protect me.) bam. Suddenly I'm paying a buck fifty per month per IP address (that's a reasonable price for one IP. but I had 768. ouch.) Now, I don't know the whole story; they say their provider (the ones that actually own the IPs) started charging them and they are just passing that through, so who knows, but I certainly felt like I was getting screwed.

So yeah; a rational actor expects providers to screw him as much as possible, and does not expect contracts to mitigate that risk very much.

If there was a way to honestly signal that you were not going to do that, I think you could attract customers.

One possible way of doing this is charging all customers the same amount for the same services; If I did that in a way my customers could verify, they would know at least that I could not raise prices more than what the average customer would tolerate (e.g. the customers that make it obvious that it's hard for them to move would not get extra screwed, they'd be protected by the customers that could easily move.) which is something. This is what I'm trying first, but it's complicated by the fact that everyone is used to getting custom configs in co-location. Salesguys love custom configs, because it makes it harder for customers to compare prices.

Really, though, this is not a complete solution. I don't know of a complete solution yet. It is a hard problem.

(interestingly, REITs own most of the really big data centers, and they do a similar thing, only their charge per squarefoot is staggering. Like $60 per sqft. then you pay a (quite reasonable, really) fee for power and cooling on top of that. This is why you see people wanting those ridiculous blade setups where they use 20Kw in one goddamn rack. If you are going to charge me $60 per sqft, then reasonable prices for power, I'm going to want my money's worth.)

>Personally I feel you should move away from the hacker market and move more mainstream and raise your prices.

For the co-lo market, that's a reasonable thing to do... well, maybe even for Xen. I mean, supporting non-technical users is dramatically more effort than supporting users that can help themselves. It's completely fair to charge a large premium for that support. But, it's also a whole lot of work. oh boy. I've actually been talking about a product in this direction; targeting developers that did not want to be SysAdmins. The idea being that we'd constrain what packages they were allowed to install, (e.g. RHEL6/CentOS6 base/updates only) and define what we'd back up (e.g. we'd take nightly dumps of your MySQL or PostgreSQL databases) but we'd take backups, install updates, and generally insure the thing keeps running. I mean, that's still not non-technical users, but it is moving more of the work on my side of the line. I think I could do that well, but I need a lot more infrastructure and (not many, but a few) more people. It's lower on my priority list than my 'super shitty storage' project.

For the Xen market... I have been raising my prices. You raise your prices in the hosted services market by holding your prices the same. How long has it been since I increased the ram per dollar? a long time. Moore's law abides. And my bandwidth allocations are 'pretty good, for 2006'.

I'm working on upgrading my existing customers to twice the ram for the price; I'm starting with the customers that are on those expensive IP addresses (now, I started this... a long time ago, and some of them are angry because they've moved IPs and I haven't made good on my end of the bargain. I should not have made that promise until I had the move scripts ready to go. But it's happening. I'll be emailing two more servers that the move script is ready to go this week.)

But yeah, I /have/ been raising my prices, by keeping them the same. And really? it's worked out far better than I would expect so far; Apparently I have something of a reputation because I'm growing in spite of a recent crop of cheaper (and better automated) competitors that sprung up in the last year or two.

One thing I wanted to add though on the issue of your costs increasing and the "retail" market. The retail market hates to switch (as much as you hate to move colo) and will stay with their current solution to avoid having to make changes (and that goes for the "tech" guy that works for the retail customer it's not to his benefit to switch hosting if he can simply pass the cost on to his customer or his customer pays directly).

So while your costs increase on a wholesale level the amount your customers can and will pay will increase as a much larger number. (If you want.)

I would explore a deal with your landlord that allows you to pay him a % for your success which takes the uncertainty out of your costs. This is similar to what is done at a shopping mall. The tenant pays for "sales" in addition to a base. While it might seem counter intuitive to do a deal like this I know I wouldn't want the uncertainty of pricing increasing and having to be negotiated in real time. That way your interest and that of the landlord are in line. And anything can be negotiated.

Generally in retail realty, a store will sign a lease and have options to renew at a preset rate that protects them. Starbucks doesn't sign a lease nor does the local pizza shop and then be held hostage. They do agree (nnn lease) though to cover any increase in costs that the landlord has in taxes etc which the tenant knows can be verified and make sense. Generally. It won't be a deal where the landlord sees they are successful and then makes a decision that they can't move and they play this game and jacks up the rent. That is what appears to be happening with your situation so I would work to get around that.

I'd be glad to do strategy with you on this further if you want. Feel free to contact me.

The reason I'm not rich is because these niches are hard to break into. It's all about enterprise sales to obscure niches. In fact, we need a dating event to pair enterprise sales people and tech founders. I tried enterprise sales, but within a month I wanted to kill myself. It takes a special breed of human to do that.

Ultimately, I'd agree that successfully getting a 37 signals, Balsamiq, or Fog Creek off the ground is difficult, but compared to the VC model of "DropBox or bust", it seems like a much more attainable goal.

I think the main lesson is that there's really no free lunch in the startup world - we've all heard it a hundred times before, but there's no silver bullet that guarantees success, or even a decent chance at success in this business. It's not quite the lottery, but a 1 in 50 shot probably isn't unrealistic odds.

Another reason: marketing. People don't get that.

Not really. He's saying that going by the numbers _should_ be only looking for homeruns but, for a number of different reasons, they don't do that.

I think it means more than this - it means if you invest in companies the homeruns will dominate your returns no matter your preferences for it to be otherwise. So there might be no model of systematic investment that makes sense investing in companies with no homerun potential. At least at the risk levels of software startups.

Granted when they initially pick your business it is probably by because you explained what a great lifestyle business it will be.

Put an evil twinkle in your eye and make internet VC think you want to rule the world and will lie, cheat and steal to achieve this, when truly you're just an honest kid with good morals who'd be happy living a modest but financially secure lifestyle.

Founder 1 Internet VC 0

One intuition is that people do not believe in a power law distribution. They intuitively don’t believe that returns could be that uneven. So when you have an up round with a big increase in valuation, many or even most VCs tend to believe that the step up is too big and they will thus underprice it.

Which is intersting, because he actually talked a little to pg about this during the class where he said it:

Peter Thiel: Do Y-Combinator companies follow a power law distribution?

Paul Graham: Yes. They’re very power law

But Thiel was not happy just stating the phenomena, but he also states on why most people discredit it, when he talked about secrets[2]:

The power law secret operates similarly. In one sense it’s a secret about finance. Startup outcomes are not evenly distributed; the follow a power law distribution. But in another sense it’s a very human secret. People are uncomfortable talking about inequality, so they either ignore it or rationalize it away. It is psychologically difficult for investors to admit that their best investment is worth more than the rest of their portfolio companies combined. So they ignore or hide that fact, and it becomes a secret.

The distribution secret also has two sides to it. Distribution is much more important than people think. That makes it a business secret. But it’s a human secret too, since the people involved in distribution work very hard to hide what’s going on. Salespeople do best when people do not know they’re dealing with salespeople.

[1] http://blakemasters.tumblr.com/post/21869934240/peter-thiels...

[2] http://blakemasters.tumblr.com/post/22866240816/peter-thiels...

Perhaps thats why index investing is so succesfull, because you get the benefit of the outliers.. the bad ideas, or long shots, that suddenly skyrocket..

No! That is different.

These are public companies, and the fact you are seeing amazing share price growth from Apple is an exception.

Usually[1] on the public share market you'll see growth rates of ~8% pa, with some slightly above that and some below that.

The continued rapid rise in Apple's share price in exceptional, and is having an effect on your portfolio that is unlikely to be seen again in our lifetime.

That is completely different to pre-A-round investing, where it is expected to see (say) 1% of companies return 1000%.

TLDR; Apple is an exception. Don't think PG's essay applies to public markets.

[1] "Usually" in the sense of the pre-2008 sharemarket.

I certainly wouldn't be surprised if it was 1000% per year at some point in the tail, if not at 1% then for the top .01 percent of companies.

I'm having trouble parsing that sentence. I think you are saying that you think 1000% per year returns are normal for the top 0.1% of companies.

This is absolutely not the case. Even Apple (by far the best example of rapid share price growth in a large company) might, maybe manage to increase its share price 100% this year (low of ~$374 in Nov last year, currently at ~$662). That's exceptional - companies like Standard Oil, Exxon, etc never managed that.

(Occasionally you may get a smaller resource oriented company that fids oil, gold reserves or something and sees a 1000% increase. Or a small drug company that has a successful trial. These are very unusual too though, and more similar to VC investing that the typical public markets).

[1] http://observationsandnotes.blogspot.com.au/2009/03/average-...

[2] (pdf, sorry) http://www.risadvisory.com/images/uploads/Rydex_Historical_t...

I appreciate the links, but didn't find information about the extremes. What I'm looking for would be something more like this paper [1] on Extreme Value Theory but with more pretty pictures.

I got lost in this one soon after the introduction, but was interested in their statements "cross country evidence that the tail behaviour of returns is leptokurtic" and "the tail distribution is of the Fréchet type, hence fat-tailed". I was hoping for a cartoon graph showing just how fat that tail is. 

I think you are saying that you think 1000% per year returns are normal for the top 0.1% of companies.

Essentially, but subtly different. Saying that I "wouldn't be surprised" was more to express the degree of my uncertainty than to state my belief. And I posited .01%, rather than .1%. To put numbers on it, it strikes me as plausible that 3 out of the 2700 listed Nasdaq issues would be up 10x for the year.

Occasionally you may get a smaller resource oriented company that fids oil, gold reserves or something and sees a 1000% increase. Or a small drug company that has a successful trial. These are very unusual too though

How unusual? I'd like to put a number on it. Is a gain of 10x over a year a 1 out of 1000 event, which would make it likely for a couple Nasdaq stocks a year? What about 100x returns over a larger number of years? I'd guess that it's happened at least a few times, but don't know.

My instinct would be that it's a fat tail, but not as fat as the VC market. Rather than hoping for 10x over 10 years (26% year-over-year) with carefully chosen startups, with a broad market index you'd probably lucky to hit 3x (12% compounded).

But what would the long term expected returns be for a broad portfolio of mining companies, pre-trial pharmaceuticals, and internet IPO's? And how would it compare to an average VC firm? I have no idea.

[1] http://www.hec.fr/var/fre/storage/original/application/3b27b... (PDF)

I don't know, but the data is out there. You can buy the complete stock history of the Nasdaq fairly cheaply. If you only want day's end prices it might even be available for free.

A quick search found some links for best performing stocks per year[1][2], which indicates that 10x is rare enough that it only happens once every few years.

[1] http://www.dailyfinance.com/2010/12/08/2010-top-10-stocks-sa...

[2] http://money.cnn.com/galleries/2011/markets/1112/gallery.for...

For instance, whereas Airbnb can admittedly be seen as a questionable idea (but not outright bad), Dropbox (the idea, before implementation) sounds like a very good idea. Maybe not a $7b idea, but still very good.

Of course, among the YC funded startups, there may be a lot of ideas that sounded bad at first; but if all those startups are in fact financially irrelevant, should they be used to try to build a theory of success?

- - -

It's interesting to learn that if YC funded 10x more startups they would be just as successful (and maybe more, since there could be a big success in the startups that are left out) -- it means that if the selection process rate is around 10%, YC could do without it entirely, with no significant effect to its bottom line.

Also, I took this picture this summer http://i.imgur.com/B30hL.jpg

I don't think so. My first encounter with DropBox was very similar to my first encounter with Google... but it was "Oh look - another way to share files" rather than "Oh look - another search engine".

Or indeed many, many people's reaction to the release of the first iPod ("Oh look - another mp3 player").

Predicting their current level of success from the point of initial investment - pretty close to impossible without the benefit of 20/20 hindsight. Their success depending so much on how well they executed and the smart changes of direction that those companies made along the way.

(with, maybe, the exception of the iPod which was a bit more obvious for those that paid attention to iTunes)

E.g. Would DropBox be anywhere close to their current position if they hadn't figured out their freemium / recommendation based model for customer acquisition?

"I can't sync more than one folder? It syncs everything to every device? (not anymore). It doesn't use WebDAV? Oh well, at least it's simple. I'll just use it until something better comes along."

And they win.

Of course they have plenty of features, an API and much more flexibility that before if you want it, but none of that complicate the core function - a directory that syncs.

I dreamed about something like Dropbox before Dropbox existed; I always have had many different machines that needed to be "synchronized" by hand. I carried around hard drives, and Iomega disks and whatnot, and used "Beyond compare" to sync all of those and it was a nightmare.

The recommendation model of Dropox had nothing to do with me adopting it -- I didn't receive a recommendation and didn't send any. But I was very excited when I was first able to use it, and still find it amazing.

Before I used it DropBox was just another file sharing/syncing software. One of a whole stack of 'em that all seemed to suck in one way or another.

I'm sure that they all promised that they would be bringing cloud file storage to the masses as part of the initial pitch.... and I tried them all because, like you, I wanted this service before any of 'em existed.

Why did DropBox win and all of those others failed (or, at least, didn't succeed so wildly)? Why was it obvious to investors that DropBox was going to win, and the others "fail"?

Dropbox is not another file sharing software; the application form to YC doesn't even mention file sharing:

http://dl.dropbox.com/u/27532820/app.html

Dropbox isn't even about cloud storage.

Dropbox synchronizes your files between your different computers -- silently, automatically, without you doing anything except turn those machines on.

Nobody did this before Dropbox and still nobody is doing it now (except maybe AeroFS, which is much more difficult to setup and use -- but certainly not iCloud or any other "solution" that is restricted to one OS or company, and certainly no backup solution either).

That's why it was a fantastic idea... which has since been coupled with a brilliant execution, yes. But the idea itself was amazing.

That isn't to say Dropbox didn't blow them out of the water in regards to execution, but it wasn't an idea that was completely without precedent. Ideas rarely are, even if they seem like that in retrospect due to one company out executing everyone to an insane degree.

This part is important. The Microsoft service only synced to your other Microsoft things. iCloud only syncs to your Apple things. Dropbox syncs everywhere.

Dropbox is to file sharing, as facebook was to social networking as Google was to search.

When I try to IM a person a picture, they may be on any one of a dozen IM systems (almost all of the compatible with Adium) - and my success in DMing them a picture is <10% trying to get through firewall. Email used to be my goto approach, but that took a bit of the spontaneity out of it.

Dropbox gives me the ability to drop an image on our shared folder and "real time" have it pop up on their side. I do this all the time, and it's just one of many, many common uses of Dropbox.

Easily the most useful new utility that I've added in the past three years to my OS X system.

But - your perspective on Dropbox - is precisely why it was so hard to predict - even after using it, who on earth would have know that it would have taken over the file sharing space so quickly? And _everyone_ thought google was going to get into this space much, much earlier.

As it is - on the surface, google offers better value and more space for your money - but I don't have a single friend who has switched over to their shared drive. We've all stayed on Dropbox because of the network effect (we've all got shared files via Dropbox - don't want to add yet another file sharing system to slow down our computer.)

We'll see if that works out in the long term - it certainly did with search.

Cross-platform internet file sharing in a transparent way is a (surprisingly?) hard problem. Before Dropbox there were many companies who had tried to make a success of it and they had all failed[1] in one way or another. (Not cross-platform enough, not seamless enough, reliance on ads for income etc etc.) DropBox succeeded because they took that hard problem and made it look easy.

1. Keeping all my documents in sync between several computers, so that I can pick up work on any one of them at any time

2. Off-site incremental backup

3. Sharing and working on a set of project files with non-technically-minded collaborators

4. Sending and receiving large files

Moreover, it just sits there quietly solving problems 1 - 3 day-in, day-out, and I never have to think about it.

Think about it. Sharing files between 25 club members, using our own bandwidth, our own disk space. And Dropbox thinks we should pay them big bucks for this. For what? Storing our files on their server, insecurely? If they had an option to stop doing that, I would select it.

SO no, it doesn't solve any hard problems for us.

The reason that storing files on their server (and using their bandwidth, as well as yours) is important is that it means I don't have to keep all my synced computers on at all times. That's a big deal for me.

It sounds like yours is another problem, which Dropbox is not well adapted to. Perhaps you'd be better off with PowerFolder or similar.

My problem is, Dropbox scales the cost as the number of people looking at a folder increases. As an Engineer I see that as marketing, their cost doesn't increase incrementally in this case. It seems unnecessary and blocks me.

Btw you would only have to keep 1 synced computer on, some of the time. Not a big deal actually. And why keep my data around on their server after we're synced? Simpler for them I suppose, but insecure for me.

I could try and get everybody in the club to install another tool; might look into that, thx.

I often work on my Macbook on the train, then when I get to the office I switch to a Win7 PC and continue working on those same files. There is zero hassle and I can hardly imagine a better solution. ()

As a bonus I get access to all my files from my smartphone and ipad. And with the ipad being such a pain in the ass to synchronize, you really need some kind of dropbox-like solution.

() - Maybe if you did all your work inside VMWare and had the state of the OS image automatically synced between computers that would be a nicer solution... Then you would not have to close your project files on computer #1 and reopen them on computer #2.

The returns aren't as dominated by just a few big successes, though. Although I suppose one-hit wonders are a manifestation of something similar: All the returns come from one song out of who knows how many dozens a band or artist may have come up with.

Publishers, from what I can tell, have the same problem. One thing that I'm struck by is how many novels that we now consider classic, or novelists who we now consider important, barely scraped into publication. Tolkien famously saw Lord of the Rings in print because of Rayner Unwin, the then nine-year-old (I think) son of a publisher liked LOTR. John Barth and William Goldman have both written about how close they were to pursuing other opportunities—a PhD and insurance, IIRC. A Confederacy of Dunces was only published after O'Toole killed himself. Melville's poetry was self-published IIRC. Virginia Woolf needed to start her own press.

There are probably others with similar stories.

Not only that, but I have to wonder who barely didn't make the cut. Given the large number of writers who skated into print, there must be at least an equally large group who "should" have, but left no marks on history sufficiently legible to trace.

I think about these issues a lot for two reasons. The first is that I've had many close calls with literary agents, all of whom eventually said, "I like you, but not in that way." The second is the technological environment: now that ebooks mean self-publishing is much more pragmatic than it used to be, people who really want to publish have a means of going outside the conventional system. Some power-law-style stars have already emerged (Amanda Hocking, the 50 Shades of Grey author). Others probably will. Maybe I'll be one. But if I'm not, I don't think I'll be too bothered: I mostly want to write.

Yeah, 99% of self-published books are probably un- or poorly edited dreck, but that 1% count for a lot.

(A side note: I wrote about this a little more at the bottom of this post: http://jseliger.wordpress.com/2012/07/29/links-the-time-for-...).

Namely, that record companies are like VC funds, and the massive cash they make from Def Leppard (or Lady Gaga or whoever) pays for the 1000's of other acts that they fund and then fail.

"The underlying thesis behind Andreessen Horowitz’s investing strategy is that in any given year only 15 companies will make up more than 90 percent of the returns. So it pays to get into those companies at almost any price."

(Source: http://techonomy.com/2012/08/the-andreessen-horowitz-effect/)

I guess once you figure out who the big winners are going to be, getting in is worth it at basically any cost when the returns are going to be so high. If others aren't comfortable with that much risk, you can win big because the price isn't driven up to expectation, leaving room for more profit.

It seems like if an entrepreneur is a little different (black, hispanic, women, little eccentric, older, etc.) he/she need to act and behave like "white young male" in order get noticed and funded. However, in that case he/she might be hiding the characteristics which will make them "rock star".

Have later YCombinator classes had a larger percentage of homeruns? If you have more applicants and have gotten better at selecting, this should be the case.

Is it about selecting winners, or weeding out losers? Are there companies that you know will fail? Who would be least likely to succeed as an entrepreneur?

Home runs are so rare that it's not a matter of percentage per batch. A batch will have 1 or 0, and probably 0. It will be a few years before I can tell if the rate is increasing.

Selecting winners and weeding out losers seem the same thing to me. There are companies we think will almost certainly fail, but we can never be sure. An ineffectual person would be least likely to succeed as a startup founder.

With start-ups I think it would be too expensive an experiment to fund.

On October 7, 1998 the Journal presented the results of the 100th dartboard contest. So who won the most contests and by how much? The pros won 61 of the 100 contests versus the darts. That's better than the 50% that would be expected in an efficient market.

This is wrong. If the dartboard consistently underperformed most stockpickers (i.e. say it ranked around the 30th percentile year over year) then you could make the case that (some) people add value. If, on the other hand, the dartboard is near the mean of the distribution of outcomes, you could make a case that it's all luck.

No such market exists for startups - so selection is required.

Well, you could try ranking results and calculating a score based on how each person outperformed the dartboard, then calculating a score over time.

The hidden premise of that statement, appears to imply there exists a universal intuition to which this is "counter". I'm not sure that is completely accurate, and is perhaps more useful to think about it in terms of a bifurcated process that can roughly segment the populace into two modes of preferred intuitive operation, but possibly more.

Without going into detail, both types of intuition make subconscious predictions, yet do so in markedly contrasting approaches, which may explain the root cause of arguments between a priori/a posteriori, bayesian/frequentist debates. The underlying ideologies may not necessarily be the effect of environmental upbringing as often assumed, but could instead be structurally rooted in the fundamental wiring and decision making of our cognition.

I'm sure some will disagree, but I naturally find frequentist methods rather challenging to understand, and typically intuit in a distinctly bayesian manner (which is probably less common, typically VC's tend to be good pattern matchers and skilled numeracy - aka natural frequentists). Coincidentally what PG thinks is counter-intuitive (like in this article), could seem rather intuitive to some. Likewise there are almost certainly other insights that are mundane, boring obvious and intuitive to PG, yet are highly counter-intuitive (and interesting) to others.

Of course it's all just anecdote and opinion for now and probably a controversial position, but hoping advancements in cognitive neuroscience research, would one day be able to test this empirically.

Paul, you're sounding like a venture capitalist who is worried about whether he can find investors for his next fund.

I would posit that the people whose opinions you should care about are potential founders; and that their primary concern is themselves, not the performance of a fund (oops, I mean class) as a whole. You're damn right that it would hurt YC's brand if 70% of each class didn't survive past Demo Day -- because for an individual founder, success is pretty much binary, and having a 50% chance of becoming a millionaire is more attractive than having a 5% chance of becoming a billionaire, despite the 100-fold reduction in mean wealth.

You may be in in the business of farming black swans, but if they're all you worry about you'll find that all the swans end up laying their eggs elsewhere.

It seems like it would be difficult to just have a public conversation about a topic. Do you think about that much when you write?

After I wrote this one, I had to go back and armor it by pre-empting anything I could imagine anyone willfully misunderstanding to use as a weapon in comment threads. The whole of footnote 1 is such armor for example. I essentially anticipated all the "No, what I said was" type comments I'd have had to make on HN and just included them in the essay.

It's a uniquely bad combination to both write essays and run a forum. It's like having comments enabled on your blog whether you want them or not.

At least some of the people who are questioning you are probably doing so out of intellectual curiosity, rather than animosity. You probably know that intellectually, but for many people—including me—it's sometimes hard to remember that in the heat of the reading moment.

Of course, my blog posts don't get nearly as much attention as your essays, and I don't have the problem of having people try to draw attention to themselves in the hopes of being remembered when applications are considered for the next YC round.

Your posts are more technical in nature, and can benefit from debate. Other people might write things that are more personal, or opinion, or thinking-aloud, and while a little bit of good-natured feedback from trusted people might be appreciated, lots of nitpicking and debate and very public arguing is not.

The point I was trying to make was that I would be disappointed if what pg clearly saw as a problem was "fixed", because for me it's a feature.

(I'm not picking on you btw, I have a tremendous amount of respect for you. But I don't think your desire for debate on what you write is suitable justification for other people putting up with the same, especially when it's discouraging them from writing.)

This is perhaps the worst thing about discourse on the internet. If you and I were having a conversation in room, you'd never pretend to not understand me to rip into me (maybe to understand my point better.) But on the internet, so many people are just trying to score points that it's nearly impossible to have a conversation.

1, that's terrible. So many benefit from the ability to read (and interact with) your opinions. It's so much more valuable than a random once-off Reddit AMA. It's a shame the value is reduced through the actions of a few.

2, this community attracts young smart people, exactly those who might want to match wits with you. Some of it is valuable, some is just annoying. I'm reminded of the niceness value in discussion; if everyone was nice about it, maybe you wouldn't feel as apprehensive or besieged and we could still have a good debate to extract the maximum benefit.

In the end, I hope you find some way to care less about the opinions, because what you do and write about are so valuable. In the spirit of "If you aren't writing enough wrong stuff, maybe you're being too cautious."

You don't give your detractors enough credit. Many of these misunderstandings are not willful, and actively seeking to avoid them is almost always good practice on your part. (In my opinion.) For the record, I found footnote 1 illuminating.

I'm sure that the intense scrutiny you get would annoy any writer. But I think it genuinely makes your essays stronger, too, and I hope you don't hate it too much.

I think at least half of the misunderstandings that arise in HN comment threads are honest misunderstandings.

The essay reminds me a bit of the argument that the Native Americans actually got a great deal when they sold Manhattan for $24, because that money would be worth about $30 trillion today if it had been invested at a return of just 7.5% for the past 385 years. I think it's interesting and worthwhile to explore these sorts of mathematical ideas, but at the same time it's impossible to make good business decisions without fundamentally basing them on human needs, human scales, and human timeframes.

What would happen if you split up the batches into two groups, YC Classic and YC Black Swan, and placed founders into the groups post-interview? People who were placed in YC Classic could continue to have the expectation of ~100% demo day success they do now, while the people placed in YC Black Swan would be told "I find your idea interesting and we'll let you in but don't expect funding after demo day." People could take the Black Swan offers as rejections if they wanted.

There are a bunch of problems with this approach, but I wouldn't be totally surprised if way more than 30% of YC Black Swan was funded.

While a theme was that the outsized returns can come from ideas which sound bad, it doesn't necessarily say that all outsized returns come from bad-sounding ideas.

Not only that, but I think reading the essay will show that the outsized returns are not known until several years after demo day.

Sometimes bad ideas are just bad ideas. In the Venn intersection between bad-sounding-ideas and good-ideas - the 'bad sounding and not good' is a much bigger area.

To me, the entire essay is about making sure that institutionally, the bad-sounding-but-ultimately-good ideas are not left out. Trying to further identify and silo them at application stage would be even more fraught.

In a world where potential founders are making a choice between YC or doing the startup anyway without YC, sure. But the world is more complicated than that, and I'm sure there are plenty of potential founders whose decisions are informed by what they hear from YC.

[...] split up the batches into two groups, YC Classic and YC Black Swan [...] People could take the Black Swan offers as rejections if they wanted.

This sounds like a very good model, in that it would separate the "is this something YC wants to invest in?" axis from the "does Paul Graham think that I'll succeed?" axis.

Let's say that p_accept is the probability that YC accepts the founders of the next Dropbox. YC itself cannot optimize for p_accept because of the factors mentioned above: instead, it has to optimize for p_apply * p_accept * p_fund, the product of the chances that those golden founders will apply to the program, be accepted, and find the funding they need to grow and thrive.

With the hypothetical YC / YC Black Swan split, the original YC can optimize for p_apply * p_fund, and YC Black Swan can optimize for p_accept. Not only that, but since all Black Swan candidates would have started as applicants to YC, Black Swan's p_apply would equal that of the original YC. Numerate investors with the same sense of the power law as pg would take care of Black Swan's p_fund.

Thus, all the prestige, cultural appeal, and midsize exits would derive from the original YC, but all the power-law returns would emerge from Black Swan.

So, pg, new business model?

The main difference in branding for KV's funds is really only to the LPs, though they do mention it to founders on their website. They want to cover themselves in case the seed funding both doesn't return anything and looked to be imprudent in retrospect. Luckily I think they're doing well.

Vinod is actually quite explicit about trying to find black swans in his pitches :-)

The group-think effects are to be considered, before adopting this model. Am obviously not predicting anything, except perhaps this move might make the Black swans even more unpredictable.i.e: if you look at the Tableau of payoffs here(http://edge.org/conversation/the-fourth-quadrant-a-map-of-th...). this move can push investments from complex payoffs to very complex payoffs easily. Personally, i am likely to even shoot for the YC Classic group deliberately or in rare optimistic moments, the Black Swan group.

Would be interesting to learn why the black swans have done better then the normal batch, or why the normal batch did better or why it was a 50/50 split.

You might even learn that because a startup is considered as a black swan founders in that group work harder and thus have a higher success rate. This would really put into light debates about how much intelligence vs hard work may effect success rates in the world of startups.

We can afford to take at least 10x as much risk as Demo Day investors. And since risk is usually proportionate to reward, if you can afford to take more risk you should. What would it mean to take 10x more risk than Demo Day investors? We'd have to be willing to fund 10x more startups than they would. Which means that even if we're generous to ourselves and assume that YC can on average triple a startup's expected value, we'd be taking the right amount of risk if only 30% of the startups were able to raise significant funding after Demo Day.

So- if a VC can triple a startups expected value that's great, but you'd need to do a bunch of them. I think this is essentially what Dave McClure is doing- making lots of smaller bets to "hit singles" as he says.

Reminds me of my college days playing online poker. The best players would have a 20% ROI at the $55 10 person tournament tables, and each game would take an hour. If you just play one at a time, you'd make about $10/hr. That's why everyone played 10 tables at a time- we made 10 times as much.

The implicit assumption is that the population of startups is uniform across both the set that YC funds and the set that YC does not fund, such that startups in the latter group have the same chance of being a big hit (modulo YC's mentoring, which is accounted for with the "triple a startup's expected value" clause). But the implication of that assumption is that YC is picking startups at random, and that their filtering process is totally useless!

This sort of math comes up all the time whenever there's a screening process. Imagine that you're hiring for a large tech company, you currently hire 1% of applicants, and you find that among the employees hired, there is no correlation between your interview scores and the employee's eventual job performance. Can you conclude that your interviewing is useless? Should you ramp up hires so you get more workforce of equal quality?

Well, maybe. Because there are a bunch of possible hypotheses that could give this result. Perhaps your interview process is designed to weed out false positives more than false negatives, so it's accurate to the 99th percentile, but then gives no discriminatory power. (Many IQ tests are like this; they're highly correlated with life outcomes up until an IQ of about 140, but beyond that they break down entirely and there's often an inverse correlation with income, happiness, etc. past that). Or perhaps your applicant pool is bimodal: 1% come from other employers and are fully qualified, while 99% are the same jobseekers that every other company rejects. Or perhaps your interview process is broken, and you would do better to find a new one. Or perhaps your interview process is okay, but a number of well-qualified applicants are not even applying to your company.

Which of these is correct? You can't know without randomly sampling the population that was rejected and making an estimate of their quality. This is why all decent scientific experiments have a control, and why financial models get backtested on data that was not part of the training set. Even then, there're lots of things that can go wrong in experiment design, and lots of different ways to interpret data that don't necessarily mean "Fund 10x more startups."

Except that YC doesn't do this. They aren't funding 10x more startups and pg says he avoids finding out how many get funded afterwards because it's the wrong thing to optimize for.

To put it another way, there are a lot of ways to bring down post-Demoday funding to 30% and most of them are not going to be helpful. The observation just points out that given their high VC funding rate YC is probably not optimizing for the homeruns as well as it should from a financial perspective.

"We'd have to be willing to fund 10x more startups than they would."

I guess that if currently nearly all of your startups are getting VC funding, there are two ways to end up funding 10x more startups than VCs: 1.) Make the VCs fund 10x fewer startups or 2.) Fund 10x more startups yourself. (Or various combinations, of course.) #1 seemed absurd to me, so I read the thought experiment as suggesting #2.

Edit: Hmm, I guess another alternative is to fund startups that "look worse" to VCs, which is basically #1 but doesn't seem so absurd. So perhaps I was misreading the quote.

"We'd have to be willing to fund 10x more startups than they would."

The contracted 'we would' is key. They WOULD have to be willing to do that IF they were going to pursue that goal. They aren't. He's positing the idea that if they were properly optimized for the black swans then they would see a much lower funding rate. Not that they should aim for a low funding rate for it's own sake. But, as he discusses, actually performing such an optimization is hard, not to mention the fact that making money isn't their only reason for running YC.

This is quite different from assuming a high failure rate. The marginal cost of having say, 10% more failed startups isn't a big deal when 95% of them are already failing, and the model accounts for that high failure rate. The black swan is not the unexpected failure that wipes you out, but the unexpected success that outweighs all the failures.

Obviously, you'd still be screwed if none of the stuff you invest in pays off. But that would be the case even if your model was more pessimistic; if even 5% success was too much to ask for, perhaps the market for VC is busted, and there was no hope for profit anyway.

I don't think pg is implying that screening processes are useless. I'm sure many obviously hopeless ideas get filtered out. It's more that, given the current state of knowledge, everything beyond that basic filter is by nature difficult to discern and we may as well assume a random fat tail distribution.

The problem is that when the business grew and everyone entered, it changed the assumptions that the models were based upon. A certain percentage of mortgages would blow up when subprimes were 1% of the market. The fatal mistake was assuming that the same percentage of mortgages would blow up when subprimes were 10% of the market, because the process of going for 1% to 10% means writing many more loans and extending credit to buyers who should never have been buying houses in the first place.

There's another problem - the "market share" of subprime loans was underreported by Fannie and Freddie (and they were the largest single buyer of the relevant securities). So, even if your model of forclosures depended on the share of subprime mortgages and was perfect, you got the wrong results because the inputs were wrong.

A large part of what made it blow up is leverage: that fixed sum was often more than the entire capital of the fund. That doesn't apply to YC, I think; to my knowledge, they don't borrow any money to fund the batches. The main investment PG et al makes is in time, energy, networking, and brand name. And that's where it could blow up: fund 10x more startups, most of which fail, and suddenly the YC brand isn't worth anything with investors, PR, employees, etc. Perhaps that's why PG doesn't do it. In this case, his intuition is a perfectly rational response to unintended consequences that are outside the awareness of his conscious thought process.

(Looking upthread, I wonder if your objection really is to the idea that more risk automatically means more reward, so therefore YC should fund 10x more. But clearly the argument is more subtle than that – otherwise, they should just fund everyone who asks them.)

Edit: I have another argument, or perhaps just article of faith: I think there's a major pool of unexploited talent out there, basically wasting away (mostly in corporate jobs) at far below its potential. If many more startups get funded, many more creative endeavors will get going. Most won't be black swans but that doesn't mean they won't be side effects of great good.

If I'm wrong about that, then maybe there is an analogy with subprime, where lenders' (investors') zeal caused them to lend to (fund) ever-crappier borrowers (founders). On the other hand, if I'm right, then this is more about correcting an inefficient allocation of talent in our economy.

Though PG didn't mention it, from the title it's clear he's alluding to Taleb, who would decidedly not justify the mortgage bubble. The idea is not simply to expose yourself to risk but to make sure you are exposed to large positive outliers while making sure you are not exposed to the negative ones.

As resbear said, when you leverage yourself you are increasing your exposure to both positive and negative fluctuations. The long-term success of such a scheme relies on a) accurately estimating the probability of positive vs negative, and b) being able to take the negative ones without blowing up. Neither of those conditions were fulfilled.

If I had to teach middle and high-school kids just one thing about entrepreneurship, this would be it!

That we can accept and reject trades (or ideas) based on their agreeability, or dispersion of opinions, is not often considered as a metric. So long as the difference between eccentric crap and alpha is difficult to discern and costly if one goes all in on the wrong bet the uncertainty should be dealt with by diversifying across only disagreeable trades.

As a former quant trader who left a large bank and found the centre of Silicon Valley uncannily similar to the heart of Wall Street, this essay is illuminating. Wall Street is being too conservative, consigning itself to blindly jumping between mediocrity (low expected return) or bubbles (strongly negative expected return) in the top two quadrants. Just like Silicon Valley, though, at its edges it allows itself to be different. It's interesting seeing how similar mantras, based in sound financial theory, change in their aesthetics as they cross domains.

It's telling, though, that the concept of a "black swan", implying a tail event generally unforeseen by the relevant population, is met with trepidation by much of Wall Street and corporate America yet seized with zeal by a select few.

This is partially why I think the YC application process is flawed. If you think about the idea for 20 seconds or so, what good is that? You may be wasting 20 seconds trying to rectify an idea you can't logically determine to be good.

A better process is a continuous one: you'll get an excellent stream of actual data and not just made up answers to "What is your best non-computer hack?" The YC partners could start out the morning looking at their Twitter-like feed of what have potential startup teams done today? At some point a partner would say, "We have to pull these guys into our next batch."

Incidentally, this would be better for startup teams looking to break the threshold of project to startup. By updating this proposed feed, the startup feels obligated to not only continue to put out data, but good data. Eventually the team would realize they can't cut it or they will keep mutating teams or changing ideas until one has great prospects.

EDIT: I'll buy someone a smile and a coke if the above hypothesis is implemented and tested. The following is exactly why: Instagram is the one we'd most likely have missed. It all depends when we'd talked to them. They were a kind of overnight success in traffic. If we'd talked to them even a day after they launched we would certainly have said yes. But before that it might have seemed too speculative.[1]

[1] http://news.ycombinator.com/item?id=4497561

As a thought experiment, I would love to hear what pg and team would have thought about the following companies, had they applied to YC before they grew in popularity (assuming YC existed when they were starting out):

PayPal, Twitter, Pandora, SalesForce, Instagram, FourSquare, and Pinterest. And perhaps a few others as well.

I feel pretty sure we'd have been impressed enough by Max and Peter to fund them regardless of the idea (which initially had almost nothing in common with Paypal).

I knew Ev before Twitter so I'm sure we would have funded that.

Pandora I know nothing about, so I can't guess there.

SalesForce I'm pretty sure we would have funded because Benioff radiates "winner" in much the same overwhelming way that Zuckerberg does.

Instagram is the one we'd most likely have missed. It all depends when we'd talked to them. They were a kind of overnight success in traffic. If we'd talked to them even a day after they launched we would certainly have said yes. But before that it might have seemed too speculative.

I don't know about FourSquare. I've never met the founders and don't understand the business.

Pinterest we definitely would have funded, because Ben is a two time YC alum (with different companies alas). We knew he was good because the first time he was part of a startup that as an experiment we didn't make move to California. As Demo Day approached, they were in terrible shape. But Ben swooped in at the last moment and gave one of the most convincing Demo Day presentations.

Does this also mean that the "big winners" will not come merely from a pool of what appear to be bad ideas, but from a subset of that pool: bad ideas put forth by founders the VC thinks are impressive? That too, seems somewhat obvious.

Taking it one step further, can you separate the "idea" from the "founder" in assessing the project's chances of being a "big winner"?

Maybe that was the goal of requesting that people should apply to YC even if they had no "idea".

After all, the supply of "bad ideas" is potentially limitless.

Have you ever experimented with an opposing strategy? (Funding "non-technical" founders with "weak" personalities.)

Probably not, lest you would risk your reputation. Hence, you might not have evidence to show that this approach would not yield suffcient numbers of "big winners" to justify the investments. For example, if someone sugested the hypothesis that from a set of a given size of non-technical founders with weak personalities, some "big winners" will emerge, could you disprove that? If you wanted to be scientific, you would have to test it, repeatedly.

From my intepretation, I think one of the points in the pg essay is that human "intuition" will only take you so far toward the "big rewards". By its very nature it steers us away from the counterintuitive and protects us from taking what others would perceive as unreasonable risks. It will stop you from adopting a strategy that no one else is using that might get you labelled as foolish (but, as history shows us time and again, might actually yield an enormous reward).

In hindsight we'll continue to see that some of the biggest winners would have been viewed as unreasonable risks by many VC. Foresight won't allow us that vision. So-called intuition will stop you from investing in a "big winner". Hard to accept, but true.

It might not yield any "big winners" (then again it might) but it would in either case produce evidence that we could use to try to disprove certain theories.

"highly profitable or highly entertaning to watch..."

or highly informative

I sincerely hope you're being flippant and don't mean that as a serious answer.

He seemed to never lose an ounce of morale from anyone telling him "no", he seemed to be doing this pretty naturally by never faulting or doubting himself but rather placing the blame on the nay-sayer. The take-away for me was someone whose morale cannot be broken is an unstoppable train. I have to be honest, it seemed full-fledged delusional to see someone never question them self based on other people (and some very important/influential people's) judgement. But that seemed to be him.

I have not met either of them. And while I have met a few individuals who have achieved success I can honestly say I have never met anyone who I thought radiated success or was otherwise predisposed to success.

Have you never met anyone with an unusually high mix of intelligence, determination, focus, and charisma (to randomly pick a few traits)? If you have, are you more likely to bet for or against them in their chosen field?

If you haven't, that is entirely possible. Such people are, by definition, rare.

Deal with people you already know.

If the project turns out to be a dud, the past history of success of the people involved negates any arguments that it was not the right decision to fund it. No one is going to question that the people involved were not "winners". They had proven that already; that's why the project received funding: because those particular people were involved with it.

To use an oft regurgitated title: "No one ever got fired for funding a project that involved [insert so-called born winner name here]"

Just for fun: How about Andrew Mason, the guy behind Groupon. How far do you think he would have gotten by taking the YC route? Is he a "born winner"? Groupon made some investors very happy.

One might wonder why no one can tell us who the "born winners" are until after they've had some "victories".

"I have not met either of them. And while I have met a few individuals who have achieved success I can honestly say I have never met anyone who I thought radiated success or was otherwise predisposed to success."

Where in the above two statements do you contradict the idea that Benioff and Zuckerberg do give this impression in person? Notice that pg doesn't say that about everyone who is successful.

It's just something you are. A bottom of the org chart stinky sysadmin doesn't just wake up one day and have the personality and determination of a world class CEO.

In a more grim outlook, it's genetic. The disconnect is when people want to be 'winners' but are, at heart, bottom of the org chart stinky sysadmins. You can work your way up to lifestyle business, but the next step of multi-billion pants in the air growth company is probably out of reach.

"Rookie team, regulatory nightmare"

http://www.bvp.com/portfolio/antiportfolio

"Cowan’s college friend rented her garage to Sergey and Larry for their first year. In 1999 and 2000 she tried to introduce Cowan to “these two really smart Stanford students writing a search engine”. Students? A new search engine? In the most important moment ever for Bessemer’s anti-portfolio, Cowan asked her, “How can I get out of this house without going anywhere near your garage?”"

It would be interesting to do an analysis of those companies that were rejected by YC, and see how many "black swans" were present in those rejects, as well as maybe relative value of the rejected startups vs the chosen ones to see how effective the YC selection process is.

Take the swans you get an enjoy them. :)

Would you have invested in Wikipedia? Tim Berners-Lee's WorldWideWeb project? WikiLeaks? Linux?

I guess I'm interested because it's not clear to me that any of the founders of these things radiate "winner" in the same way that you seem to look for in your founders. Or maybe they do?

Better transportation infrastructure, especially high speed rail. (Notice how the French have better track than the Germans because the Germans put off investment while playing with maglev; Elon Musk might be at risk for repeating the same mistake in the US with Hyperloop.)

Fiber optic cable to just about every building in the world, owned by people who charge at a rate that isn't much higher than the construction cost and don't try to keep charging more and more for higher data rates.

Possibly some more good parks.

We need cheap solar and batteries too, but I think those things are likely inevitable given enough time at this point.

The biggest wins have been to invest in advertising and walled-gardens. I much prefer the thought of investing in people though, that which can make the greatest difference to individuals.

I don't have money, but I do have time and skill. And my money is where my mouth is, the idea I'm working on does need to be profitable to self-sustain, but the goal I have is to change lives of individuals.

To me it doesn't matter that others don't define this as winning. I think their definition is faulty.

So instead you treat it as a calculus problem. The metaphor Thiel used was that of space travel and the Apollo missions. You assume you know exactly what's going to happen and behave as such. In a nutshell, "no-one would want to ride in a statistically, probabilistically-informed spaceship."

It's a hard task spotting black swan founders (so instead you just leave yourself as open as possible to them). But the conviction of calculus (in them and in you) is almost certainly a necessary ingredient.

You could stretch a Wilde one-liner around this: "We are all in the gutter, but some of us are looking at the stars."

Another thing to consider is that the best teams often pivot into a big success rather than start with it. There are countless of such examples, and it just goes to show that even starting with what appears a relatively safe but limited idea can eventually grow into a huge success. Sometimes founders need to get their hands dirty in the market to realize what is the real opportunity. If you pass those teams up because you think their idea cap is too small, you'll be missing a lot of big hits.

What it all comes to for me is investing in people. People make big hits, not markets. Markets can grow and startups can span multiple markets, but it all starts with the people who direct it forwards.

Take the "seems like a bad idea" bit. Many of the big hits are really well-worn ideas done with better marketing, better timing, and a user experience which makes it available to new markets.

Dropbox? File sharing. Facebook? Geocities. Both are interesting in that they take something that was a giant pain in the neck but useful enough to put up with anyway, and then they make it usable enough that virtually anyone can do that stuff routinely.

But the meta-cognition isn't all that important. What's important is that they're getting out there and helping new companies learn and grow. They can be wrong all day (or not) about the why because they've developed a process which provably works, and which they can measure and improve upon.

Also, the fact that Dropbox and Facebook were "unoriginal" is exactly what made them seem like bad ideas at the time. I saw Dropbox present at demoday, and the main question on my mind was, "How is this different from the million other attempts at online file storage? (xdrive, etc)".

It seems to me that both Dropbox & Facebook are example of companies pursuing good ideas that lots of other companies were also pursuing. They just happened to be the ones that executed the best (and possibly also had the best luck) so they were able to win the market.

If this is the case then your job as a funder of startups might be to try to identify the companies capable of out executing everyone else on a good idea that a lot of people are working on.

This is much different than the great idea that looks bad phenomenon that pg discusses in his essay.

I don't have the time though, so if any bloggers/journalists are reading, this could be an interesting article, or even book, if you get enough info.

Whether or not this info would actually help another team become successful is another story though.

The big successes that I've seen all have exceptional founders, but that's clearly not enough. Timing is a huge component as well. Often a good idea will fail simply because the market or technology isn't ready. My own experience with Gmail reflects some of that. When we decided to write the whole frontend in JS, everyone said that it was a bad idea doomed to failure. It had been tried before (e.g. desktop.com) and had always been a disaster. They were right about the past, but wrong about the future. We released Gmail right around the time that browsers were finally getting good enough, and were were very careful to keep our code as fast and slim as possible.

The actual formula that academics use is Performance = AMO, ability * motivation * opportunity. I think there are a couple of modified versions but I can't remember what they are off the top of my head.

This is to say that my experience confirms that there probably is some completely subjective formula that will fool people into thinking there is some reason behind my good luck. :-)

I wanted all of those services before they existed, which to me says that they never seemed to be bad ideas, unless you were unduly afraid of implementation/regulatory difficulties. It may not have been obvious which competitor was going to win, but it seemed likely without the benefit of hindsight that somebody would.

Facebook and Twitter seemed like fundamentally bad ideas to me at the time (in fact, they still do, but the world appears to disagree), and I just don't get instagram, but those are the exceptions that prove the rule.

With paypal it seemed crazy that sellers would hand over control of their bank accounts to this unqualified, unregulated company. Given paypal's history of account freezes, it still does.

Video sharing sucked in 2004, damn right. Microsoft and Real and Aol and Apple were pouring big money into making it work. Could we've guessed that someone was going to make a lot of money in internet video? Sure. Could we have predicted that it would be youtube? Much tougher. Dropbox was the same; there were hundreds of other products making much the same claims about how they could synchronize your files easily. Likewise skype; we'd had netmeeting for years (even today, polycom are still in business selling software that works less well than skype for hundreds of thousands).

Can you predict that a given service could be done much better? Maybe. But picking which startup is going to win based on that is a whole lot harder.

Or does PG mean that the idea can seem bad because the market is already very crowded? I remember talking to William Morgan about the early days of GitHub. He knew the founders from Powerset, and he told me that at the time he thought to himself, "Really? Another hosted version control startup? That seems like a bad idea."

http://news.ycombinator.com/item?id=8863

Turns out my memory was only half right. DropBox had many people that loved the idea and thought it would revolutionize filesharing. It also had a large number of people who pointed out all the reasons why it wouldn't work.

I wonder if there's a lesson here in that good ideas that seem bad tend to be highly polarizing. I've recalled Paul Buchheit say here that GMail met a lot of internal resistance, with many Googlers saying it was a distraction and would never work. I also recall Larry saying that there was significant support for GMail at all levels of the company, going up to the founders, and many Googlers loved it.

Yes -- exactly -- and they tend to generate a lot of heat in group discussions. One of the indicators we watch for are people getting visibly angry during the discussion -- either angry that other people aren't "getting it" or angry that other people ARE "getting it".

To be fair, it's not actually possible to present a new idea on this forum (or any forum with technical-minded people), without getting a flood of reasons why it won't work.

1) Execution driven success in a winner take all market. The idea seems good, but there is lots of competition so it's hard to see how the startup will break through & win the market.

2) A genuinely "good idea that initially looks bad" like you describe in your essay.

It's hard to think back and decide which, now successful, companies are 1 vs 2 because of the memory distortion effect, but it really does seem to me like #1 is more common. What do you think?

If you're absolutely certain of yourself, chances are that there are others out there who are just as certain, so you will have competition. If you're uncertain, and feeling lonely, there's a better chance that you have the field to yourself.

My experience of this is being involved in the development of the first WiFi (802.11a) system. It was only with hindsight that the significance was clear. The reality at the time was an isolated toil in the dark, not a high flying roller coaster ride.

Best strategy is to "win" in other ways, and let the cards fall where they may.

Taleb's new book coming out soon will describe all this in a lot more detail -- should be very interesting.

But whats the thought process and how does it change given this insight ? It should still be the case that as an investor you still should focus on the best opportunities and try to maximize a decent return from the majority of your investments then focus on picking that facebook, dropbox or AirBnB right, just as you would if you were a trader in the stock market ... otherwise your just gambling ?! Focus on the technicals and data that give you the best odds and just take those trades cause taking trades where you are waiting on information from the fed chairman or draghi without insider info for example (even though they have big pay offs if you swing the right way) will earn you pretty much the same amount of returns as losses in the long term.

If the potential to win big time is counter intuitive what do you trust from data in front of you to make the decision to invest other than the "gut feeling" ?

In any case, i would think with dropbox it never really sounded like a bad idea and the founder/founders (Drew and Arash) sounded very promising even from the start. If this is the case even good bets do pay of big time to :) maybe these are easier to pick then the facebooks where at the time it does look like a bad idea.

Basically, Im curious as how this insight in PG's essay would change how an investor should invest or is it just an FYI and something to keep in the back of your mind as an investor ?

That's not quite right. It's more a question of taking enough of the right kind of chances in funding the companies that you do fund, as opposed to all the companies you don't fund.

I think of it as the concept of "search space" -- you need to define your search space properly, so that you have a good chance of having some big winners in the search space. You have to define your search space because you can't invest in everything. The investors that do the best job of defining their search space (having the most accurate criteria, best judgment, best mental models, etc.) have the best chance of funding some of the winners -- almost as a side effect.

When you talk to old successful VCs, what they tell you is that they had almost no ability to judge which of the companies they funded would go on to be the big winners in their portfolios at the point of funding them. They only learned that later. The challenge was to get enough of the right kinds of risk "above the line" and then give the portfolio time to develop.

Don't you think you can you pick out some that had zero chance of becoming massive-scale homeruns, because they served sort of a niche market? There are none that were obviously (to me, or apparently to pg and co) going to be unprofitable, but if you want the next Facebook, you're not going to get it by investing in, say, Codecademy, which, as laudable as its goals are, only appeals to newbie coders, who are a tiny slice of the population.

Well you would just never know. Computer science and software engineering could become part of the basic school curriculum in the next few years. In the news you see that "the 14 year old boy developed an X for the iPhone and is now a millionaire". Enough of these stories could prompt the education system to deem coding as a basic skill like math, english and science. Code academy could end up being the platform that schools use to teach it.

But i guess the chances are slim because as it stands now the population is small though the possibility is there.

If there's only one winner in a class, at best, does it matter if the class is 10 or 50? 50 isn't going to give 5 winners, is it?

Is the idea that 50 gives a 5 times higher chance of finding the 1? Is that really true, though?

Would focusing on a smaller number of potentials actually be better for finding the 1? Thinking through the scenarios more thoroughly?

During YC's history so far, we know (a) is true and (b) seems anecdotally to be true.

  > the one thing you can measure is dangerously misleading [Demo Day fundraising]

Footnote [4] http://paulgraham.com/swan.html#f4n says airbnb did get funded; and dropbox was unanimously voted the best in their class (according to one of them http://www.quora.com/Y-Combinator/When-Dropbox-first-present...)

So this is an intriguing thesis, but not supported by the (admittedly scanty) evidence of the two biggest outsize hits. Do the other outsize hits support it (i.e. by not raising funds after Demo Day)?

I expect it will take a few more years but it will turn out to be financially rational to invest in teams with repeat YC founders.

If you could theoretically "randomize" start up ideas, founders, business models, and every other primary ingredient in current "start-up theory", like a giant multivariate test... would we not see similar results to now?

Or similarly, if you closed your eyes and invested "blindly" how would you expect your results to differ (ROI-wise)?

It seems like investing results more closely resemble a lottery than a predictive model, based on your description.

Sort of a scary thought: what if current investment criteria (and therefore start-up goals) are largely irrelevant?

For example, IQ is distributed in the same way as winners and losers in start-ups. What if that is a key factor,(purely hypothetically, of course)?

In theory, shouldn't "good criteria" for investing generate progressively better results within smaller and smaller samples?

Clearly you don't look at YC in purely financial terms, but it would be an interesting thought experiment to consider whether there would be more efficient ways to generate returns if you did.

Among other things, there may be more efficient ways of testing variations of ideas around a theme, or testing/ranking skills, or assessing likelihood of success for various experimental team combinations.

To use a casino metaphor, it seems like you're taking a 50-100 completely binary and largely independent bets per class and hoping to beat the house on average. As opposed to, say, trying to create the MIT blackjack team to beat the house systemically.

The reason that angels and VCs show up is because YC is providing a service for them, so from their perspective you would have jumped the shark. Each angel only has so much money to invest, and VCs feel it's necessary to provide value-add in other ways, so because their ends don't scale I'm having trouble seeing this purely mathematical strategy benefitting anyone besides YC. If there were a way for follow-on investors to benefit in a way that helps improve the overall startup ecosystem then it makes a lot of sense, but I think that's a case that needs to be made that this essay didn't touch on.

If YC had funded Paypal, they would have made a nice return with the $1.5B eBay acquisition. But their valuation would've ended there. In reality though, they currently make up 44.9% of eBay's revenue. eBay is a $63.45B company.

I don't know YC's policy on holding stock in the acquirers company after someone in their portfolio is bought out, but assuming they keep stock, a company acquired by Google a few years ago which allowed YC to accrue a few million dollars worth of Google stock could very well be their biggest financial gain in 50 years.

My point though is that in 50 years, their minute percent ownership in Google they received via an acquisition could be worth far more than their much larger percent ownership in Dropbox.

So which was the better investment?

* Dropbox used rewards for more storage, as well as its fundamental file-sharing tools, to encourage users to sign up their friends and family.

* Airbnb gamed Craigslist [1] to encourage property owners to signup on Airbnb. With a strong supply of rental properties, Airbnb was able to build a true case for its value in the minds of travelers.

[1] http://venturebeat.com/2011/06/02/airbnb-admits-gaming-craig...

That said, it often isn't obvious from the beginning. When I met Dropbox at YC, I don't recall them having any great plans for how to grow. I think that was discovered later on.

Also, that often-repeated Craigslist claim about AirBnB is actually false. They got very few listings from Craigslist.

So yeah; right now? PG is absolutely doing the right thing. But if I'm right, after the crash? assuming he comes through financially able (which I assume he will; he doesn't seem the type to bet the farm on facebook stock) and wanting to continue investing, at that point? he will look for sustainable companies. If you are trying to start a traditional sustainable company, in many ways, it's easier to do during a downturn. There is little competition for inputs, so you can get everything you need much more cheaply, and there is a lot less competition, so if you come up with something new, you won't have 5 other 'me too' companies copying you in the first year.

Really, there's less opportunity cost for the founder, too. I mean, right now, the temptation to get a job at facebook or google is really strong, even for me; I'd get to work with some really great people, and probably make about 3x more 'profit' than I do now.

During the downturn, getting those plum jobs that pay really well while letting you work with really great people is a whole lot more difficult, so taking the pay cut while you get your business off the ground is not as big of a deal.

This was basically the message of PG's recent leaked "bad times" email:

"The startups that really get hosed are going to be the ones that have easy money built into the structure of their company: the ones that raise a lot on easy terms, and are then led thereby to spend a lot, and to pay little attention to profitability. That kind of startup gets destroyed when markets tighten up. So don't be that startup." [1]

What I take away from all this is: have a revenue model built into your startup from the start. Use it. Generate profits of some kind. Not only will this help when you need investors (since they like businesses that can stand alone), but it will help you even more if they don't come.

[1] http://www.businessinsider.com/facebook-fallout-y-combinator...

I mean, twitter didn't introduce advertising at all until they were absolutely huge. I had the impression that facebook was similar (though I could be off; I'm not a regular user of facebook) - I mean, in a very real sense, it's easier to grow if you don't monetize (assuming you can still keep the lights on.)

This would basically be a test of the notion that "ideas don't matter", and it would allow you to correlate how even knowing a founder's idea might bias you towards believing they might succeed.

But many startups exit without making it big and return 2-3X to their investors. If the median startup were doing 3X, you'd be making a great return. I don't hear much about this strategy. It would be interesting to look at the remainder of YC's portfolio with Dropbox and Airbnb removed to see what that distribution looks like.

If we just examine YC exits to date, Heroku and OMGPOP are around $200M each, $400M in total, which is likely over 50% of the rest. If we remove those too, there are a few exits in $30-60M range (Loopt, CloudKick etc.) which are quite likely to be over 50% of the rest. At the lower end of exit price curve, the power law might not hold, as acqui-hires typically have a rather standard price ($1-2M per an engineer)

The "not merely not" phrase here is a little confusing. What PG is saying is that the probability of it succeeding is inversely related to the probability of it being a big success?

This reminds me of Arthur Rock's heuristic: he said he invested in startups that were "open ended with no limits to their success" as opposed to ones that seemed likely to succeed.

Solve a big, annoying problem for lots of people. Dropbox was obviously a great idea; look at how much trouble people went through to share files (using flashdrives, CDs, email, spammy uploading sites or ones that enforced a delay) before it got big. Just because there are existing players in the space doesn't mean the problem is solved. There are so many other major inconveniences that people have to put up with that are incredibly ripe for disruption. Healthcare - why is it so damn hard to get an instant, cheap medical opinion on a non-urgent but still worrying issue if you don't have a doctor in the family? Apartment rental (for both owners and tenants).

The hard part is not telling whether or not something is a good idea. It's evaluating whether or not the team can pull it off.

I remember hearing PG comment that as a founder you need just a few million from a startup. That's as true today.

Perhaps the relationship between exit valuation and wealth created is elastic. If the goal is to make wealth, perhaps lots of 10-100M startups has a bigger impact on society than a couple of billion-dollar startups.

In terms of dialing up risk, I think you're looking @ it backwards. You have a finite life, which means you are slowly eroding to zero. You have nothing to lose by taking risk because you are already eroding towards zero.

Even worse, the remaining days you have left are a greater percentage of the remaining days in totality of your existence, so therefore each day is eroding FASTER than the previous one.

So, relatively speaking, you should be increasing risk over time in any case because the rate at which you are eroding is accelerating.

Basically, rather than looking @ risk as what you have to lose, really its about what you have to gain. There isn't any true risk because the end game is the same.

YC would be a pretty lonely place if we only had one company per batch. And yet it's true.

The point is that not only would it be lonely, it would also be impossible to pick that one perfect company per batch. (Well, not impossible, but vanishingly unlikely.)

And I think pg understands this well, as shown by the rest of the article - a person or investment team can't possibly pick that 1000-bagger without it being a complete lottery. So you invest in a large number, knowing that most of them won't do too much, a few will tank, and one in a hundred will be a grand slam that pays for all the others.

I think I understand the intent of what you're saying here, but it could also be read to mean that you can increase your reward simply by taking on more risk. It might help to clarify that whenever a market isn't perfectly efficient (which is apparently always, unless P=NP), there will be investments that carry more risk without offering a higher return (i.e. that have a negative alpha). Perhaps it's better to state that as the potential reward (or return) of an investment increases, you can tolerate more risk.

The point is, profit has to be booked to be realized.My question to PG is how does YC decide when to sell?

Now, I wonder whether the understanding of both aspects is enough to put him in a position to recognize founders on the other side of the table capable to provide a solution to random black swan farming?

Eventually more interesting is the question what would be his best guess from which field of science he expects useful outcome enhancing the criteria to make more educated decisions on founders?

Am I alone with the impression the essay sounds stucked?

The fact that the best ideas seem like bad ideas makes it even harder to recognize the big winners. It means the probability of a startup making it really big is not merely not a constant fraction of the probability that it will succeed, but that the startups with a high probability of the former will seem to have a disproportionately low probability of the latter.

It could be that it's a bit late, and I am tired. But it sounds a bit awkward.

Anyone care to clear that up for me?

Thanks!

If the public Demo Day is now fraught with too much expectation perhaps it should be treated as a PR event rather than a goal for the startups to shoot for.

This doesn't solve the other problem you mention, which is that you'd be surrounded by people flaming out, which would be stressful in a lot of dimensions.

Are the returns going to remain of this order over the next 20 years ? I understood that VCs rule of thumb was 10x return paid for the other 9. is this difference earlier stage, bubble or other?

I am still surprised how YC hasn't transitioned to that entirely. They already have the money; optimizing for that is tryin to play a financial game that others have already mastered.

That's a lot like distilling whisky - you have to wait at least a decade to find out how good a batch was.

Assume you could separate the funding vehicle from the YC reputation, having it as a separate unassociated entity.

That is, assume the "reputational" cost could be managed. And no one would accuse you of being "insane".

I liked this essay. I think it's sufficiently honest.

http://blakemasters.tumblr.com/post/21869934240/peter-thiels...

if there is a 500 pound supposition in the room it's liquidity. if the only way an investor can achieve it is through a massive "event" such as an ipo or sell out then any otherwise successful company is a non-entity for investors.

if y combinator were holding securities with some liquidity it needn't devolve into a 20th century vc. there are many ways to achieve this, all disruptive yet also quite well suited for this institution and its community.

(http://on.fb.me/Pb3DeP)

Worse, even -- innumerate and/or people who don't have domain knowledge. Probably considerably over 99% of all people.