It seems as though popular websites, books, songs, movies, etc offend a lot of principles of quality and taste. What's more, it seems to be by design. It is. The funny thing is that this is exactly as it should be. It can't be any other way.

The reason is that people are more different than you usually like to think. The form of a message has to be tuned to the intended effect, to the semantic receptors of the audience. When your audience is large and the common thread between them so thin, the form of a successful message becomes self-parody, an appeal to the basest desires. Nothing escapes this rule. Just look at the highest of highbrow-but-popular stuff (say, Shakespeare or Cervantes) and count the profound ideas shoved next to diarrhea jokes.

I spent a few years designing ads for the phone book and direct mail. It's a no-nonesense business. Personal aesthetic takes a backseat to performance. In a way it's great: you have a single test of effectiveness (i.e., the phone rings or does not ring), reinforced by fast iteration and feedback. Direct Marketing is a real-world example of design by genetic algorithm.

This kind of design does not always produce ugly results. It seems that way because, by definition, you see more wide-audience-targeted material than other kinds. Yachting and Spa ads are pretty posh but odds are good you're not on the list.

Nor is popularity a good measure of quality. The books Tom Sawyer and Huckleberry Finn have the same characters, same kinds of silly scenes, pirates, farce. Yet Sawyer is a kid's book while Finn is a masterpiece. The difference is the stuff Mark Twain wove in between the crowd-pleasers.

I sympathize with esthetes. I am am one. But if you are trying to make a popular website, remember that you are a commercial artist. Commercial art is the most demanding because of its constraints: the worst being that you do not have final say. The audience does. Would Monty Python be as popular as it is without the flying sheep or the road-testing they did on each sketch?

Also: An essay on Whitespace by Mark Boulton

1) Any theorem that is disproved by contradiction must be thrown out.

2) ...except for Bivalence and LEM even though Russel & Gödel kicked the bottom out of them three generations ago. It's just too scary.

3) Probability, that fluffy blue blankie we use where bivalence breaks down, is axiomatic even though it produces no significant results.

4) A logical tautology cannot also be a factual truth.

5) ...but the Principle of Induction always holds, except when it doesn't, and when it doesn't it means there is something new to learn. Why? Because that's the way it has always happened in the past.

The accuracy of a given company's mapping application is 100% for the company headquarters, greater than 75% for the homes and hometowns of company employees, and undefined everywhere else.

Well, I've made it into Startup School this year. I'm genuinely excited, not just for the speakers but to meet the other 649 hackers.

I'm curious to meet Paul Graham in particular. Over email he's always gracious & thoughtful. But --and this is a little seed of a thought that's been growing for a couple of years-- I wonder about the effect he's having on the industry: whether YCombinator is doing good for the art of software as it is for the business.

Take music, the other creative pursuit linked with garages. People learn it on their own or at expensive schools. Groups condense out of the social soup of young people, and it tends to clump in certain cities. They have goofy names. They do it because they love it. The best are able to work magic. A lot of them fail, some make a decent living, a very few become rock stars.

That's the myth. But that's not what actually happens. In the background are competing factions of promoters, managers, coaches, ghost writers, hired experts, financial backers. Without the cooperation of this network your chances of making it are thin enough to cut your wrist. Without them the Beatles would have just been a really good bar band. Generations of jazz and blues geniuses died in poverty because their looks were not marketable. Talent is not closely correlated with success.

So ask yourself: Who makes money in this regime? The backers, the promoters, the hired guns, the music schools, and a pinch of the talents. How does this regime further the art of music? Do you have a radio? Have you used it lately? The internet is a great end-run around the music cartels, but your startup is already on the internet.

YC is backed by serious & smart people. The pitch is soft, the benefits are compelling. They didn't create this situation. They are doing a lot of good: directly to the kids coming to Boston and as inspiration to everyone else. But if you love the art of software, keep your eyes open. Don't assume that what's best for your success is best for your art, or what's best for them is best for you.