Let me just jump in on a topic that I've seen a few times, and one that I think throws programmers for a loop. It's the whole "premature scalaculation" thing.
Did you know the 3 guys who built YouTube initially didn't start it out as a video site? It was a dating site! Even more, did you know they didn't even use a version control system? The philosophy was "push once a day, make some type of change that the user can see." They also were focused on building out features, not worrying too much about scalability at first.
The guy who started PayPal said that one of the most painful parts were the times where he had to delete massive amounts of his own code. He said one of the biggest flaws of engineers is to write "the perfect code." And yet, 90% of it gets destroyed shortly after it's written.
I'm not an advocate of crap code. In fact, at our startup, we write very good code and follow the API models of companies like Google and Flickr so as to have a scalable product and architecture with clean, readable, commented code. But when you're a startup, you have different priorities.
I'd say maybe the first 6 months (an arbitrary number) don't worry about scalability. You don't even have a user ;) Why not worry about how you'll get your first 100 users, or better yet, how you'll get your first 100 users happy, as Paul Buchheit puts it?
On a final note, I read someone else mention about the CEO thing. In my opinion, you don't need a CEO until you at least have user growth down. Yes, you want investors, but try to offset most of that work to PG or whoever else is investing. It's in their interest that you succeed too. You should be working day and night with your engineers, building the product. It will make the future fundraising and meetings much easier, just think about Google and Andy Bechtolsheim.
I really enjoyed reading this, and I think you're going to go far in the future. Sometimes we get dealt some bad cards, and the only thing you can do is to fold 'em and wait for the next hand. Stay persistent, stay focused, and kick ass.
> Did you know the 3 guys who built YouTube initially didn't start it out as a video site? It was a dating site!
Really? Source? I knew that the they'd created the "I was at a party and realized there was no good way to share videos" story for the press. But I thought the actual story was that they'd done market research about trends that were changing the technology landscape, and came up with the rise of cellphone video and broadband access. From there, it was a short leap to "let's put video online".
There a lots of stories of how the company started, the press not being the most reliable. Check out the ACM talk one of the founders gave earlier at uiuc.
I believe the idea was sparked after one of the founders read a particular wired article.
The genesis section starts around 25 minutes in. Looks like we're both right: they mention the Wired article at 25 minutes in, then the cell-phone video aspect (via Indian Ocean tsunami) at 29:00, and the other enabling technologies (including broadband) at 31:41.
Lol you can see the two kids in the orange sitting on the stairs in the middle of the screen. I'll say that the stairs were very uncomfortable.
I saw an e-mail that one of the founders wrote, and I can't remember where it was.. maybe at the second meeting which was private/not recorded. The contents of the e-mail were asking friends to tell their friends, and they offered to pay girls to sign up on the site and upload videos (thinking it would draw in guys).
It's kind of silly to argue how YouTube started--the argument I was trying to make initially was that your code will probably be erased shortly after it's written, especially if you're not sure what your company will be working on.
Let me just jump in on a topic that I've seen a few times, and one that I think throws programmers for a loop. It's the whole "premature scalaculation" thing.
Did you know the 3 guys who built YouTube initially didn't start it out as a video site? It was a dating site! Even more, did you know they didn't even use a version control system? The philosophy was "push once a day, make some type of change that the user can see." They also were focused on building out features, not worrying too much about scalability at first.
The guy who started PayPal said that one of the most painful parts were the times where he had to delete massive amounts of his own code. He said one of the biggest flaws of engineers is to write "the perfect code." And yet, 90% of it gets destroyed shortly after it's written.
I'm not an advocate of crap code. In fact, at our startup, we write very good code and follow the API models of companies like Google and Flickr so as to have a scalable product and architecture with clean, readable, commented code. But when you're a startup, you have different priorities.
I'd say maybe the first 6 months (an arbitrary number) don't worry about scalability. You don't even have a user ;) Why not worry about how you'll get your first 100 users, or better yet, how you'll get your first 100 users happy, as Paul Buchheit puts it?
On a final note, I read someone else mention about the CEO thing. In my opinion, you don't need a CEO until you at least have user growth down. Yes, you want investors, but try to offset most of that work to PG or whoever else is investing. It's in their interest that you succeed too. You should be working day and night with your engineers, building the product. It will make the future fundraising and meetings much easier, just think about Google and Andy Bechtolsheim.
I really enjoyed reading this, and I think you're going to go far in the future. Sometimes we get dealt some bad cards, and the only thing you can do is to fold 'em and wait for the next hand. Stay persistent, stay focused, and kick ass.