How My Brain Kept Me from Co-Founding YouTube
prog21.dadgum.com
prog21.dadgum.com
Thankfully, like the writer, I eventually realized the stupidity of my ways, but I think many of us are susceptible to this bug at least once ;-)
In 2 years, that will be 30 years ago. If RAM doubles every 1.5 years, then in 2 years a new machine should have 16G RAM (x1,000,000). At the moment they have about 4GB, so it's not far off.
The thing is, 16GB RAM will soon be ordinary. Every PC will have it. What things become possible that are unthinkable today? Maybe multiple virtualized OSes could become standard.
It is a philosophical one instead. The ideas of tomorrow must look beyond what is possible today into anticipating what may be possible tomorrow.
The execution of such an idea would then hold the promise of innovation to break today's barriers.
On the flip side, it also opens opportunities that people in the industry simply didn't entertain, because they were literally unthinkable and therefore unseeable just a few years ago (and I literally mean literally). I spoke to the inventor of a new method (a few decades ago) of hardware multiplication that became the standard for a while; he said it was made possible by an increase in silicon. It was theoretically possible before then, but no one looked.
Had he had the foresight to avoid those limiting beliefs, he could have invented something like YouTube.
He wasn't in a position to co-found youtube, not to mention there were already plenty of video sites around before and during youtube's rise to prominence.
It's a classic case of "well I had that idea I could of done it" - doing it is 99.9999% of the exercise, having the idea is 0.0001%.
Given the current state of copyright law it was pretty much insane for them to expect to be anything other than a lawsuit magnet; unless they had a plan to deal with it.
In other words, just because they sold to Google, doesn't mean other companies didn't also offer money that would have made the founders millionaires multiple times over if they had accepted any buyout offer in the previous 6 months to a year before that. Most any startup that sells for millions or billions of dollars didn't just accept the first buyout offer they got and it happened to be for a lot of money; no, they had been negotiating and evaluating serious offers from day one, all the while improving their software for months or years.
Remember the days before flash video? Remember having to use quicktime and realplayer? Those days sucked.
YouTube democratized online video. The founders may have been at the right place at the right time with the right execution. However, do not discount the fact that YouTube's founders came from PayPal. I'll try to find the link later but in an interview one of the YouTube founders said being on the ground-floor at Paypal and watching them build the runway for the company to take-off inspired them because it made them believe it was possible.
The only people who've made money off YouTube are the founders. I'm glad Google are kindly subsidising it for us, but then I'm not a shareholder or a recently laid-off employee. As a business it's a dog.
He then goes on to equate this to developers not wanting to have their users download 64MB runtimes or 2GB IDEs. These two things are entirely different.
A developer knows that, while Joe User may have no problem waiting for a 500MB video of some chick topless to download, the same user isn't going to want to wait to download a 50MB installer for a program that makes Bingo cards.
As something of the resident expert on this topic, let me tell you: users are certainly willing to wait for a 12MB installer for a program that makes bingo cards. It actually converts (very marginally) better than the 500kb installer which needs Java as a pre-req.
The default VisualWorks image is 12MB. And I also worked on one production app that was 110MB deployed. Most of that was data. We pre-cached all of the queries required to populate all of the drop-downs, so that after it got done loading, users found it to be "fast."
I thought of something like YouTube. "Flash video? No self-respecting techie would use Flash!"
A fuckton of people have come up with the idea that "hey, since a 4gb address space is really plenty, why don't we try using an n/2 address space and have a whole cons fit in one word?" I'm sure the same thing has come up plenty of times before.
It turns out to be not so hot to avoid directly using the OS' VM and the hardware's MMU.
A mind easily blown. Stand back anytime this guy is near a website that is anything more than an aggregate of img tags.
Flickr is fucking atrocious for actually viewing someone's photos. It's about as shitty as they could possibly get away with.
On the other hand, it has absolutely terrific support for communities, with all sorts of little touches to encourage their formation. It also manages to simultaneously appeal to US Grandmas, Arab twentysomethings, and the requisite "self facilitating media nodes" that fuel WEB 2.0 hype.