The discussion of biotech was interesting. It made clear that Y Combinator is replicating the biotech model of mass production of innovation.
Paul said (more or less) that the risk investing in Biotech was in some ways lower (because it's easier to identify the 20 leading scientists from whom the next wonder-discovery will emerge). I think this correct, but the comparison between that and what YC does is misguided.
Those leading scientists are like the companies who've already been through YC, presented on a platter for the big money to get at.
So what's the YC equivalent? Paul said it himself - grad school.
For Biotech, a PhD program is the way to mass-produce seed-funded innovation. You give someone a few years, let them attack a problem and generate data. The best PhDs coming out get labs and professorships (equivalent to venture funding worth $1-$2 million), etc.
Hacking is not science. Therefore a seed-funding model that works for science does not work well for software innovation. YC is the replacement - it's your PhD in software innovation. People go in, work with mentors (I meet my PhD supervisor once a week, and regularly hear talks from the best in the field, just like YC founders). The best innovators come out and get funded for $1-$2 million and hopefully go on.
If there were flies on the wall, they wouldn't be nearly as interesting. Just corner some of the speakers elsewhere and put a few drinks in them, and you might get similar results.
Another technique I found works great is to "be wrong on front of an expert". Make uninformed but curious statements before knowledeable people and I will take the time to tell you how it really is.
Two of my favorite topics, entrepreneurship and economics, being discussed by leaders in each respective field.
It's when you get crossovers such as this when you get into the field that Joseph Schumpeter and Clayton Christensen really mainstreamed:
1. "Creative Destruction": The concept of new technologies destroying the older ones.
2. Innovation as the driver of creative destruction.
My only real critique of Y Combinator and Graham's model in general is that it's still very tech/web-oriented. I get the cost and exit benefits, but the noise around this sector has changed people's ideas of startups: Lemonade stands, clothing lines, etc.
Most surprising takeaway is at the end, where once again immigration limitations are shown to stifle economic creativity and growth.
My only real critique of Y Combinator and Graham's model in general is that it's still very tech/web-oriented. I get the cost and exit benefits, but the noise around this sector has changed people's ideas of startups: Lemonade stands, clothing lines, etc.
Ycombinator is web-oriented, but the model isn't (necessarily). You could have a company similar to YC in any field with low initial overhead and high scalability. Unfortuantely, I can't think of any that can match web software (maybe finance, actually: there are already hedge fund incubators, and they might scale down average funding to scale up the number of people they fund).
You could probably easily do the model for consumer products if you would be able to figure out a test distribution mechanism, such as a partnership with a grocery store chain, or some such. Yeah, it would be something akin to the American Inventor show, but the key is to be able to do serious in-store or TV ad (informercial or 30 second) testing.
I agree that this model works for web software and tech in general because the nature of the industry nowadays allows for lower startup costs. A lot of industries are still capital intensive.
A similar model that specifically selected companies in non-tech industries whose competitive advantage/business model centered around some low-cost, innovative solution to a formerly capital intensive problem could be successful if you could find a number of startups with that focus.
one of the things mentioned towards the end is that they "don't get" the time sink application. that is because they are producers. the internet has made visible the vast unproductive masses. paul doesn't like timesinks because his time is valuable, most people's time is not valuable.
I simply loved listening to this. It was more like a fascinating conversation with Paul Graham that we got to listen in on than an interview and it was extremely exciting.
Thanks for the fascinating insights, essays and interviews Paul.
I've read Paul Graham's essays and followed his projects ever since I stumbled upon his essay on Java (as an ambitious teenager dreaming of hacking), but I've never actually heard him talk. I was surprised, gathering from the manners of expression and enthusiasm, how well he fit my image of the hacker and nerd - although I had all the reasons to expect that. It's not the voice I've heard in text.
Anyway, it got me thinking how much being surrounded by young (be it yet intelligent and stimulating) hacker people affects the way you think and talk.
I did. But this is a prepared speech, by PG the essayist, of prose.
I guess the novelty factor for me in the interview was more of a social quality, through communication. The part of a person that does not show in prose.
By far, the most important statement (although it was all quite interesting) was this: "There's nothing better than to have sophistocated users. And conversely, it's dangerous to build a product for giant corporations or, god help you, for the defense department, because then the stupidity of your customer will make you stupid. You want to build a product for clever, impoverished outsiders."
Gives me hope that they'll see that in my startup's YC application.
"worse than random" - great way to describe the likely result of big-corp/government investment in IT projects. The YC model (and VC model to some extent) works precisely because of the unpredictable nature of innovation - it is hard to predict 'the next big thing', 'the next google' etc. but you can increase your chances by investing thinly across many startups. The returns for those that succeed will far outweigh the cost of those that don't.
Paul said (more or less) that the risk investing in Biotech was in some ways lower (because it's easier to identify the 20 leading scientists from whom the next wonder-discovery will emerge). I think this correct, but the comparison between that and what YC does is misguided.
Those leading scientists are like the companies who've already been through YC, presented on a platter for the big money to get at.
So what's the YC equivalent? Paul said it himself - grad school.
For Biotech, a PhD program is the way to mass-produce seed-funded innovation. You give someone a few years, let them attack a problem and generate data. The best PhDs coming out get labs and professorships (equivalent to venture funding worth $1-$2 million), etc.
Hacking is not science. Therefore a seed-funding model that works for science does not work well for software innovation. YC is the replacement - it's your PhD in software innovation. People go in, work with mentors (I meet my PhD supervisor once a week, and regularly hear talks from the best in the field, just like YC founders). The best innovators come out and get funded for $1-$2 million and hopefully go on.