The Next Mark Zuckerberg Is Not Who You Might Think

Of course we all do, but that's pure ego. Each one of them was brilliant, there is no discounting that, but you also can't ignore the role that luck played. You can't just be brilliant and you can't just be lucky, you really have to be both to have the kind of success they did. It's not surprising that they are anomalies, hoping to recreate their success is like hoping to find a winning lottery ticket.

We shouldn't put any of them on a pedestal anyway. The empires of each of them was built on locking people into their closed systems at any cost. Jobs did some remarkable things, Zuck seems like he's going to do positive things with his fortune, and Gates might eliminate the world of a lot of disease, but you only have to look at past cutthroat titans like Andrew Carnegie to know that redemptive actions like that leave behind at best an ambiguous legacy.

There was a great piece on Grantland this week talking about Silicon Valley, Halt and Catch Fire, and Microserfs. It had this bit which I thought was really interesting:

"The moment reads two ways. Either it’s about Zuckerberg, for all his billions, being Just Like Us — a slave to the same digital toy to which we’ve subcontracted management of our memories, our personal interactions, and our sense of self-worth — or it’s about how Facebook, born of Zuckerberg’s sense of exclusion, has made us just like him. The notion that the tech visionaries whose inventions colonize our daily lives are actually uploading the virus of their personalities to the global unconscious is an idea that floats through The Social Network; it’s the engine that powers Alex Gibney’s forthcoming documentary Steve Jobs: The Man in the Machine, which stops just short of postulating that every iPhone includes a fragment of poor angry uncharitable emotionally stunted Steve’s immortal soul, like a sliver of wormwood from Mordor."

http://grantland.com/hollywood-prospectus/silicon-valley-hal...

If you want to be like Zuck, what is it about you that you are going to infect the world with?

And most of the founders are older guys who are also boring, they don't do exciting movie star CEO stuff. They just found a super specific niche and are filling it really well and making great money doing it.

So when I go and talk to these same investors, it is clear that most of them don't really care about moonshots or big market shifting ventures, despite what they say. They want cash flow positive companies who might have a reasonable chance of being acquired and don't really need that much work - because make no doubt, all of those unicorns or whatever you want to call them take a ton of work from the investors.

But I have no desire to be a part of one of those companies. They generally aren't particularly ambitious at global scale - they aren't trying to change human behavior or move markets like airbnb, uber, facebook, google etc...they just want to make good money, and are generally pretty good at it. Might as well go work for Oracle at that point.

I wonder if many of these programmers don't really like coding, just like the waiter doesn't really like serving people.

Maybe it is just because I live outside the Silicon Valley filter bubble but I have no plans to ever start my own company. I like to architect really nice systems, and turn them into well coded implementations. I don't ever want to have to touch a pitch deck, or deal with investors, or worry about making money, or burn rate.

I'll be the first engineer at someone else's startup, but I don't want to be a businessman first, engineer second. To me coding is not a gateway to getting rich by creating the next Facebook/Snapchat clone. Instead coding is the destination, and I'm at that destination, doing the thing that I enjoy every day.

“I can be tricked by anyone who looks like Mark Zuckerberg,” Paul Graham, co-founder of the seed investor Y Combinator, once said.

Graham already pointed out that this is a joke and that the quote was taken out of context: http://www.paulgraham.com/tricked.html

If we want to predict what the next MZ is going to be like, shouldn't we restrict our training data to the set of successful unicorn founders instead of the set of all founders?

Most startups fail, so it seems this data is most useful for predicting who won't be the next MZ.

The deep irony of the consumer-focused venture industry—for all its talk of technology and transparency—is that it still operates with the shrouded obscurantism of a medieval guild.

People find more likely what they they recall more easily and vividly, so this is like adjusting a linear model only considering the outliers (top or under stellar), so the fitted model don't represent the general population of founders whose life is more mundane and less interesting that a film star.

Edited: To clarify.

Illicit drug industry sounds like one. Am I taking this quote out of context, are startup founders customers of VC's? Did not realize that was the hierarchy if that is the case.

<< Yet if someone like that came to a top venture capitalist’s office, he or she could very well be turned away. Start-up investors often accept pitches only from people they know, and rely heavily on gut feelings, intuition and what’s worked before. “I can be tricked by anyone who looks like Mark Zuckerberg,” Paul Graham, co-founder of the seed investor Y Combinator, once said. >>

Paul was right - this meme just won't die.

<< People will probably still repeat that quote, but now if someone does it will be proof that either (a) they didn't do their research or (b) they have an ideological axe to grind. >>

http://www.paulgraham.com/tricked.html

First, the article says

Many people think they know what the founder of a tech start-up looks like: a 20-something man who spent his childhood playing on computers in his basement and who later dropped out of college to become a billionaire entrepreneur.

And then refute this with the more "diverse" reality that

The average founder is 38, with a master’s degree and 16 years of work experience.

Note no mention of the gender of the average founder. Next, they consider the study by Bloomberg Beta that the article implies refutes the "myth in our heads of what the prototypical start-up founder is, and that myth is an early- to mid-20s white male who studied computer science at an elite school and dropped out"

But the methodology of the study is to create a machine learning model that uses the "backgrounds" of people (which seem to include which companies a person worked for, and I would guess also which school) to predict which people will become a founder. They then define the people who the machine learning model scores highly, as "potential founders" and examine the demographics of potential founders. The results quoted are:

While only 12 percent of current founders are women, when they searched for potential founders based on the other characteristics of successful founders, 20 percent of the people they found were women.

“If you look at just the professional histories of the people who got funded, then it suggests people who share those histories are much more diverse than the people who get funded,” Mr. Bahat said.

Only 53 percent of founders have technology backgrounds, indicating that a computer science degree is not a requirement. Of potential founders, 8 percent of those with technology backgrounds are women.

So in the strange logic of the author, the machine learning model's validity is not called into question even when the difference in gender bias between "potential founders" and actual founders is mediated by highly relevant information, such as a technology background. To be specific about the problem with this logic, if the machine learning model selects a completely different looking population (ignoring gender) from existing founders, that is evidence that the VCs are biased, but it is also evidence that the machine learning model is wrong. The high level problem with this approach is that the same variable they are trying to show is biased (being an actual founder) is the one they are using to train the model which selects "potential founders" in order to prove the bias.