VCs = Angels = YC = Agents They all finance and support young entities for a % of a future asset.
Compare a very scalable model, IMG Worldwide. Take baseball for comparison sake.
Talent Scout = PG colleague who reviews applications using performance criteria = bank using advanced credit scoring system to approve or reject loans
Reddit = Alfonso Soriano = Successful loan applicant
Paul Allen = PG = Drew Rosenhaus
YC is popular not because they aren't a VC but because they are a VC who provides more perceived value for what they charge. That means just like IMG, their model is scalable as long as their value proposition is maintained and continues to be favored. All YC needs to do to fund 1,000 companies next year is to hire/acquire/develop 50 more PGs with specialties in more focused markets and sub-markets. Not simple but definitately doable. Also fully automate the application and evaluation process so that "people" only touch the top 3%. Continue to be careful with funding amounts but generous with time and connections. Continue to distribute and increase your value through your customers by financing their meeting each other and adding value to each other. Basically, YC is a social network with most of the value right now distributed offline through their customers. News@YC will change that to be sure. I think it's not only possible to grow up real big, but the direction you are already headed. Just make sure your scalability has a bit of Ben & Jerry's thrown in, they did a great job of keeping their value system in place long after they grew. PG, you mentioned it's the scaling of the partners responsibilities that is hard. If the partners maintain the integrity of the barriers to entry (active role in developing and maintaining your application evaluation tool) and support (ensuring key events and high-level connections are solid and cascaded down from their organizations to yours) you can do it. Just my $.02.
So what your suggesting is creating mini YCs with specialist interests?
For example you could have a mobile apps 'agent'. This person would seek out the best ideas for mobile apps, find the best teams, maintain connections with mobile blogs & mobile network operators and pass the best teams back up to main YC.
O'r you could have geographical specialists who, for example, look out for startups in the UK and send the most promising over to Boston/Mountain View.
Yeah Dan, that's right on. The basic business practices employed today would flow through to the tentacles of specialization and the over reaching controls and management would still reside at the top like any other holding company structure. It's likely you could improve on the typical holding company structure but something along those lines. I work at the Group level of a global financial services firm and I think my stripes are showing a bit here :) Not just mobile apps but also a p2p apps expert, b2b apps expert, transaction processing expert...the options are as broad as the marketplace. I'm thinking big.
I'm planning on flying out from Connecticut to join the meetups and try and get to know other people who share my goals even though I won't be attending the actual startup school event. I look forward to being able to tell the story about how I met my cofounder while crashing a YC "everyone is welcome" social session.
Here is a point I made at that blog...One very small piece of this puzzle I havent read here I work for a global investment bank and here we manage risk. Paying one person 3x to do the work of 5 headcount is great until that person gets sick, moves to Iceland to get married, quits to go on Beauty and the Geek, gets burnt out, falls off a bridge, etc. Then you are screwed. That is a huge problem from a management perspective.
I wouldn't say you're screwed. I'd say you just need to get another equivalently high-powered person to sub in. It'd take them a while to load the code into their head, but so does any replacement programmer. And, if you do know how to get Eric, getting just one more is probably not that hard.
ecuzzillo, your point is valid but in the environment of a large company, hiring can take months, there is a huge administrative expense associated with the hiring process and the top 1% is still the top 1%...hard to find even if you had already found the first Eric.
In the environment of a normal large company, the giant administrative expense of hiring people (and in general the suckiness of being in a big company) is a far bigger problem than trying to find a good programmer.
I think that ties into what the author said about the advantage of mediocre teams for long term projects. Its sometimes better to having something kludgy and understandable than an elegant hack that no one else will understand.
You're presenting a false dichotomy. You don't get smaller code by making it incomprehensible. It gets smaller because it's cleaner and clearer.
Think of the problem in terms of compression. You can compress via syntax or by semantics. In the first case you end up with gzip and an unreadable compact mess. In the second you end up with a better, more elegant program, expressed via core, largely orthogonal concepts. For some reason you seem to think that a good programmer works like gzip. That's simply wrong.
Of course the other side of that is the fairly common situation where someone in the team of 5 is doing 5x the work for 1x salary, and management has no idea until months (if ever) after they leave...
I agree with everything PK says and would add one more thing: JL does an amazing job of staying out of the way of the answers. Most interview style books allow way too much truncation of the answers by the ego driven interviewer, usually right when you can tell the answer was about to get interesting. Well done.
She's a superb interviewer. This isn't your 60 Minutes style of interviewer where she's the star and the subject is simply the vehicle to make her look smart, etc. Also, Paul Graham's forward is one of the best essays I've read in a long time. And, perhaps, the best opening paragraph. Damn, the man should be hired just to do that. The sprinter metaphor: genius.
At the end of the day this is a public explanation of a private decision. I'm guessing that the amount of thought that went into this decision is massive and began on some level long before PG made a comment. Without personally knowing Andres, I think it's safe to say as a physics expert he possesses the analytical skills required to go into this with a clear head. I for one give him credit for acting on his instincts rather than playing it safe. All the posts below are proving that point. How many people would be claiming that he made a mistake if he continued to pursue Physics and Octopart failed? No one because who can criticize the pursuit of a doctorate? Chances are his dedication to Octopart will lead him in yet another direction, one that most likely would not have reared itself in the basement of the Physics lab.
As for the women...well that is a whole other conversation. Andres, I know lots of women now that I'm married, let me know your type.
Compare a very scalable model, IMG Worldwide. Take baseball for comparison sake.
Talent Scout = PG colleague who reviews applications using performance criteria = bank using advanced credit scoring system to approve or reject loans
Reddit = Alfonso Soriano = Successful loan applicant
Paul Allen = PG = Drew Rosenhaus
YC is popular not because they aren't a VC but because they are a VC who provides more perceived value for what they charge. That means just like IMG, their model is scalable as long as their value proposition is maintained and continues to be favored. All YC needs to do to fund 1,000 companies next year is to hire/acquire/develop 50 more PGs with specialties in more focused markets and sub-markets. Not simple but definitately doable. Also fully automate the application and evaluation process so that "people" only touch the top 3%. Continue to be careful with funding amounts but generous with time and connections. Continue to distribute and increase your value through your customers by financing their meeting each other and adding value to each other. Basically, YC is a social network with most of the value right now distributed offline through their customers. News@YC will change that to be sure. I think it's not only possible to grow up real big, but the direction you are already headed. Just make sure your scalability has a bit of Ben & Jerry's thrown in, they did a great job of keeping their value system in place long after they grew. PG, you mentioned it's the scaling of the partners responsibilities that is hard. If the partners maintain the integrity of the barriers to entry (active role in developing and maintaining your application evaluation tool) and support (ensuring key events and high-level connections are solid and cascaded down from their organizations to yours) you can do it. Just my $.02.