“One of the most common pieces of advice given to young people is to get a university degree. It’s true that in many cases a degree from a good university helps you get your foot in the door. But in most cases it doesn’t tell you what to do next. You have to learn that at the job. The curriculum for the knowledge economy should be more like the one for the Industrial Revolution. People need to learn specific skills, not just general ability.”

There is something else tech companies can do to help. They can make it easier for people who don’t have college degrees to get the jobs currently reserved for people with them.

This would be good for two reasons. First, it would fill those jobs with people who really want them, not just people who settled for them because they couldn’t get anything better. Second, the other path from high school to a well paying job would create a more varied labor market, where people without college degrees could work at tech companies and do things that require more training and cleverness than driving a truck or waiting on tables, but not as much as programming—for example, working in tech support, or as a systems administrator (most students of mine who now have such jobs are doing very well), or designing websites.

What I am imagining is an intermediate level of work that doesn’t require a college degree today but might if we created it. Most programmers I know are getting paid between $60k and $120k per year (in Silicon Valley); somebody doing one of the above jobs could be paid between say $30k and $60k. If we could create enough of these jobs to employ half the kids who now go to college (and thus would be employed at

It is not a coincidence that the companies founded by college dropouts have tended to make software. The closest thing we have to a formula for starting successful tech companies is this: take the riskiest components of your plan, and make them even riskier. The riskiest component in most plans is starting with no users. So forget that, and make it your plan.

But while dropping out of college may be necessary to start most tech companies, it’s not sufficient. It’s only one element in the recipe. College dropouts have a good track record at making software because making good software is so complicated that it takes years to learn how, and you can’t do it in school. This is true even if you’re a graduate student at a leading computer science department.

There are two reasons you can’t learn much about software by taking classes or getting a degree: one is that there aren’t any classes or degrees; the other is that if there were they wouldn’t help much anyway.

If you want to know whether someone knows how to program, ask them to program. It’s the only reliable test I know of. But if it’s so easy to tell who knows how to program, why are there degrees in computer science? Because you can’t tell

A lot of people, even most people, want to work for big companies. One reason is that it’s a safe career move. You don’t have to worry about whether the company will still be there in a year or two. The bigger the company you work for, the less likely it is to disappear overnight.

We can trace this trend back to the rise of the nation-state, but it took a great leap forward with the Industrial Revolution. There was suddenly a much larger middle class, and everyone wanted jobs with big companies.

The downside of this trend was a new kind of risk: not the risk that comes from being on your own, but the risk that comes from being part of something huge. A new employee of General Motors was safe, but if General Motors went bankrupt he was sunk along with it–and indeed it did go bankrupt in 2009.

A few years ago I noticed an odd fact about these companies: their names all end in “Inc.” Even companies that did something completely different from each other were all ending their names in “Inc.” Why?

I think what’s going on here is an analogy. People want to work for big stable companies because they imagine them as being like big stable countries. And when you’re talking about

It’s not just the tech companies. In 1995 I was running a small company, and we hired someone who had dropped out of high school to work for us. We didn’t do this to make a point; we did it because he was smart and got things done. But even back then it was getting harder to find people like him. It’s no coincidence that the personal computer revolution happened in California, where you could legally hire teenagers.

In my experience there are two main paths to wealth in Silicon Valley: starting a company or joining one early on–preferably both. And neither is easy. The only way to start a company is to get together with other people and do it. To increase your odds of finding good cofounders, you want to be near as many other potential founders as possible. And you want to be near the kind of people who will turn into users of whatever it is you’re going to make.

The only way to join a startup early on is for the founders to want you on the team. For this, too, you want to be proximate–close enough so they can meet you and get a sense of whether they want you around all day.

Silicon Valley has two advantages: lots of founders, and

A friend of a friend is a director at one of these companies. She told me that she likes to hire two kinds of people: specialists and generalists. The specialists are the ones with really good technical skills, particularly the “10x” programmers you hear about. They are so much better than everyone else that it doesn’t matter what else they’re like.

The rest she hires as generalists, because in every other way they are so much better than the specialists that it doesn’t matter what their technical skills are like.

You can’t tell if a startup will succeed, but you can tell if it has the right kind of users. That’s what investors look for: not just growth, but growth of the right kind. This is why investors like to see a startup get traction by growing virally. When a product takes off on its own, it means users love it. If all your growth comes from sales and marketing, that means users aren’t eager to tell their friends about you.

What are the right kind of users? The best kind is other startups, especially in the same industry or market segment as you. (If they’re not in the same segment, they may be able to help you pivot.) You want them using your product before you have any salespeople, because then getting them will be evidence that your product is good. Startups have an incentive to try out new things that big companies don’t: they’re desperate for an edge.

Tech companies should therefore look for a following among startups long before they try to make sales calls on big companies. They should also let other startups use their product for free, or at least at a steep discount. It won’t take long to make up for this in sales to these startups once you have paying customers; meanwhile you