[pg]

Skimming this, I noticed a few mistakes.

I was 30, not 31.

My degree is in CS, not computer engineering, which is a hybrid of CS and EE.

We were already working on another company when we decided to work on ecommerce. We were making software to generate web sites for commercial art galleries (who didn't want web sites). So we didn't suddenly decide to write software to make stores. It was more a question of switching to a market that wanted what we could make.

Robert's apartment was not in NYC. I was the one who lived in NYC, and I was visiting him in Cambridge.

I didn't wake up with a specific sentence in my head. I just woke up with the idea that we might be able to control the software on the server by clicking on links.

It's an overstatement to say that the idea of Viaweb was too strong to fail. We came close to failing several times. It is true though that the best thing we had going for us was the quality of the software (rather than, say, marketing, or connections).

We weren't the only ones "insane enough" to make web-based software. At least one of our competitors did. It was a big help though that our most dangerous competitor took a long time to grasp the idea.

The story about the origins of YC omits Jessica. We decided to start it one night while we were walking back from dinner in Harvard Square (http://ycombinator.com/start.html).

Robert never devoted his time exclusively to YC. Jessica and I did, but he and Trevor have always done it part time.

As far as I know, no one has ever tried to put a valuation on YC. You could value our current assets fairly precisely, but that would come to less than 500m.

[zatara]

I am almost afraid to ask you this, but here it goes.

On the last few weeks/months before starting Viaweb, did you consider yourself a failure for being almost 30, well-educated but out of the formal career track, "poor" and unmarried? If so, was that the fuel behind your many amazing achievements later on?

[pg]

No, not really. I'd written the two Lisp books, and people liked those. Not a lot of people, but they were people whose opinions I cared about. Actually Viaweb felt like more of a compromise than the way I'd been living before, because it was something I was doing mostly for money.

[sayemm]

So, you finally had your first taste of startup success at age 34. And you started Y Combinator at 41.

Think your story, along with many others in the Valley (e.g. Jim Clark), goes to show that this is a long-term game, and it only gets better with age and experience.

[edw519]

Skimming this, I noticed a few mistakes.

Also, the story started a little too late for me. I would have loved to have heard about pg's earlier inspiration...I often wonder if he drew it from the same places as me: the classrooms of Doc Shultz, John Drumm, Adrienne Kapisak, Laura Mikesell, and Linda Sproull, and of course the computer lab and chess club at Gateway High School. I wonder how he ended up in computers instead of history, art, or literature. Funny how time flies. We've both come a long way since Monroeville, Pa.

Perhaps a future blog post?

[pg]

I did write about that once: http://paulgraham.com/heroes.html

[sharkweek]

We were making software to generate web sites for commercial art galleries (who didn't want web sites).

Do you think this is where your affinity for people who might be pitching a bad idea but seem smart / have good work ethic comes from?

[pg]

Yes, actually. I feel like one of our advantages is to understand how lame founders can seem initially and yet ultimately succeed, from having been there ourselves. Or more precisely, which dimensions it's ok to be lame on, and which it isn't.

[sriramk]

If you had to fill out a YC application for ViaWeb back then when you initially had an idea and a working demo, how do you think you'd have fared?

For example, while it was very obvious that your team had high skill (given your academic credentials), you didn't have domain expertise in online commerce.

Would you have picked you? :)

[pg]

That's a question we ask regularly.

We did have domain expertise of a sort: we were experts in generating web sites, which back then was a comparatively rare type of knowledge.

I'm pretty sure we would have invested in us, because we had a very clear thesis about what needed to be made, and ecommerce would have seemed a promising market.

[technotony]

Could you fill that out a bit more? What dimensions do you think are critical to have right, and which can you fix later (ie be lame at initially)?

[pg]

You can be clueless about business and uncertain whether you want to start a company. You can't be lazy or stupid.

[capdiz]

"It's an overstatement to say that the idea of Viaweb was too strong to fail. We came close to failing several times." Sounds polite and down to earth but its very true. Google, Microsoft am sure they all had times when they came close to failing. That's the good thing about being successful. The hard times are forgotten by everyone (mostly the naysayers) else apart from the person/people who really felt them most thus making that success bitter sweet.

[ry0ohki]

I've never had an article written about me or my products that didn't have some mistakes, I'm not sure why it's not standard practice to send a draft to the subject for fact checking...

[snprbob86]

I've had a few reporters send me quotes and excerpts for fact checking. They then proceeded to ignore my corrections, since their fabrications better matched the target narrative.

Once you're on the other side of the news, even a little bit, you begin to understand just how broken it really is.

[ry0ohki]

Sometimes quotes are sent yes, but not the rest of the article. Usually the quotes are not off but the facts of the actual story.

[rchi]

I had the same experience on fact checking

[jcarden]

As have I

[arbuge]

Besides the remarkable insight that went into Viaweb, which I think this article captures, the thing I most admire about what PG & RTM accomplished there is the hustling and knocking on doors they carried out to sell their application after developing it. Quite inspiring for modern-day B2B entrepreneurs I think...

ps. "You could value our current assets fairly precisely, but that would come to less than 500m." <--- elegantly understated; this is what success at the seed investing game looks like :)

[outworlder]

My memory might be betraying me, but I do remember a lot of hype about how people would be able to visit 'virtual galleries'. That's how stuff like VRML was usually pitched, but people didn't actually know how the technology could be used. Which reminds me of the whole "housewives will be able to store recipes in their personal computer" thing.

That could have been pg's motivation. I could also be reading to much into it.

[hnriot]

I remember viaweb, or webgen as I think it was called at the time, I was at Pointcast then who were also skirting around Yahoo. Back then I am not sure there was much of a distinction between computer engineering and either of CS or double E. It seemed prudent in the 90's to also get a good grounding in the digital systems side of computer science because it was a time where knowledge of the hardware gave software a significant edge. Which this is still true in some areas today, for most people, concepts like byte alignment and endian'ness are irrelevant.

From what I remember there were quite a few making web based software, Hayden's Critical Path for example springs to mind, and of course Yahoo and Nutscrape (as it was fondly called by those unfortunate enough to have to work on it)

[leadholder]

>generate web sites for commercial art galleries (who didn't want web sites).

Were you generating the sites in sort of a 3rd-party, no-connection way, or did they just want someone to create their sites for them?

[emiliobumachar]

There are a few details in this essay of him:

http://paulgraham.com/bronze.html

[hiddenstage]

$500 million valuation?

YC could probably sell their stake in Dropbox alone for more than that.

[rdl]

YC's stake is common stock. It was presumably somewhere between 2-7%, like is advertised for current YC companies, back then. I assume it got diluted a lot since then, as Dropbox has had multiple rounds of financing.

Dropbox is worth maybe $5b. I think YC's stake comes out to 1-2% after the rounds. So, off by an order of magnitude or so. Just based on published info about the distribution of YC companies and their value, it's probably more like $200mm total fund value. Given that money-in was say 300 x $20k, and maybe triple that to account for salaries and other overhead, say $20mm, it's a 10x return so far.

Which is pretty baller for a VC fund.

On the other hand, I suspect PG etc. could have created a single startup of their own worth >$300m x 2 (to account for dilution) and probably exited it, in the same amount of time.

[YuriNiyazov]

They might have actually done that anyway. Like PG said, YC itself is worth ~ $500m when marked to market, and I imagine they get acquisition offers.

[SatvikBeri]

I wonder. YC's initial stake is usually <10%, right? Dropbox got $1.2MM right after YCombinator, which to me suggests a high rate of dilution. So let's say YCombinator owns 5% of dropbox. Dropbox is probably worth slightly less than $6bn according to pg's comment that Dropbox + AirBnB are worth about $7.5b. That would value YC's stake in Dropbox at about $300MM.

Of course, this is just a guess-I have no idea how much Dropbox was diluted in their early funding rounds. Even so, Ycombinator's stake probably couldn't be worth more than $600MM if we assume Dropbox is worth $6bn.

[hiddenstage]

I figured that much, but on the market I'm sure someone like Google/Apple/Microsoft would pay a premium to current valuation for 5% of Dropbox.

[bibinou]

> Unix, the open-source operating system

edit : what UNIX did you use ? Linux ?

[jsyedidia]

freebsd.

[tedmiston]

Was the book written with your involvement or without?

[pg]

Robert interviewed me.