90% of Everything Is . . . Take a Guess

Last week I spent some time sitting on Mr. Andreessen’s back porch pondering the magic of 90% and how it created Silicon Valley and continues to rule it. “In venture capital, you think a lot about so-called adverse selection,” Mr. Andreessen says. “ ‘Why am I so lucky as to be the person you’re trying to raise money from?’ You want to get to positive selection, the best people coming to you.”

But how? Mr. Andreessen starts with the replication crisis in scientific studies, especially in psychology—over half of studies can’t be replicated. I suggest “studies show” are the two most dangerous words in the English language. Mr. Andreessen quickly adds, “The corollary is ‘experts say.’ ”

Mr. Andreessen’s friend in the scientific research world told him about a historical study of heart and lung drugs that were approved but were not effective. Mr. Andreessen learned that “one of the things you do to counter a replication crisis is a ‘preregistration of hypothesis’—instead of pretending after the fact that you have a hypothesis, that you’re cherry-picking data to prove.” The result of this preregistration? There were fewer new drugs approved because researchers could no longer fudge the data. “Of course, what this implies is that most drugs that are already on the market today probably don’t work.” His friend agreed and said forget 50%, it’s 90% of research that is bad to begin with.

Then Mr. Andreessen says, “I had heard of this, aha, Sturgeon’s law!”

Theodore Sturgeon

(1918-85) was a science-fiction author annoyed that people were saying all sci-fi was bad and wanted to stand up for the 10% that was good, saying, “90% of everything is crap.” Mr. Andreessen says “90% of music is crap.” The same is true of “paintings, writing, TV shows and movies.” I would add ideas, stocks, opinions, politicians—the list goes on.

That is “the nature of creative work. There are only a few people in each field that know what to do,” Mr. Andreessen says. “The reason I’m so fascinated by this is the ethic of our times is egalitarianism. Everybody’s the same, everyone is equal. The conceit of the times, the ethos that if only you put the work in, you’ll get great results”—Malcolm Gladwell’s 10,000-hours nonsense. “It’s not just effort. It’s not just accidental. There’s something else going on. In these domains, we have a very small number of people who know what to do. And we have a much larger number of people typically laboring under some set of delusions—generating crap. It is what it is. I wish there were more quality painters . . . or entrepreneurs.”

How does this explain Silicon Valley? After

Franklin D. Roosevelt’s

science adviser,

Vannevar Bush,

conceptualized modern university science—“scaled science”—metrics were put in place: number of papers, grant money, number of students per professor, citations. Universities created an implicit scoring system that was easy to game. University research is a “self-accredited cartel with no market pressure,” Mr. Andreessen notes. Hence the replication crisis.

“Super-talented” people, Mr. Andreessen figures, leave academia or big companies because they realize they’re “swimming in an ocean of mediocrity.” Hence the positive selection for venture capitalists. “Doesn’t this take a bold person to make the jump?” I ask. “Yes, an aggressiveness and intolerance. Or aggressive intolerance!” Mr. Andreessen says. “But it’s not the fault of the people, it’s the fault of the systems.” Academic research and big companies like

IBM

“keep running the same scripts.” They need a new model. Take note,

Apple,

Amazon,

Facebook

and Google.

So Silicon Valley became the positive attractor for talent. “Venture capital got to meet the very best, those in the 10%,” Mr. Andreessen says. Thank you, Sturgeon. And away from the East Coast, it was “the vanguard of the frontier.”

But doesn’t Sturgeon’s law also apply to investing in startups? A typical venture fund has a home run, a few companies with middling returns and lots of smoking holes in the ground. “We have a limited partner that has comprehensive data,” Mr. Andreessen explains. “For top-end venture funds, the good news is that it’s not 90% failure rate, it’s 50%. This is for top-decile venture.” So maybe Sturgeon’s law doesn’t apply. But wait. The 90% of funds below the top decline do worse. Sturgeon’s law rules. Mr. Andreessen pauses, raises an eyebrow, and nods in agreement.

Does this law apply everywhere? Or perhaps 90% of all laws, including Sturgeon’s law, are crap.

Write to kessler@wsj.com.