Most work is akin to construction jobs: you work until the house is built. You just have to keep the servers running day after day. You keep writing code day after day. You teach another class. The only risk you take is that they may not be work for you next year. The work is not, in itself, risky. Showing up day after day, and doing your best, is often enough.
Other work is more like gold prospecting. Almost all freely available open source is never used. Almost all research papers are useless. Most books find very few readers. Most blog posts are written in vain.
With gold prospecting, the results are nearly as certain as with construction jobs. Programmers know that their current project won’t compete against Linux. Mathematicians know that their research paper will be read, at most, by handful of people. Writers know that their next book won’t sell more than a few hundred copies.
However, gold prospectors rely on probabilities. Even if your chance of success is only 1%, you are guaranteed (within 1%) to succeed if you try 500 times. If your chance of success is a meager 0.1%, then try 5000 times.
That is why so many open-source programmers, scientists and writers thrive despite the odds: they simply keep on trying. It is only a risky business if you cannot try multiple times. You might be tempted to object that writers don’t try to market their books 500 times. But some do! J. A. Konrath collected nearly 500 rejections from publishers. In a recent blog post, he provides a list of comments by major publishers who all rejected his best-seller The List. Though no publisher would take his book, he is now making $5000 a day from the sales of this self-published book alone. He won because kept trying long enough. I am convinced that most gold prospectors eventually succeed, if they try long enough.
Shortly after completing my Ph.D. I decided to become an entrepreneur. I did not have contracts lined up. I did not have a business idea. I did not have investors. I did not have business experience. Yet I never ran out of money or work despite the apparent risks. Some projects were failures and I did lose time and money. Other projects were rewarding beyond my expectations. It averages out over time.
Most prospectors just put their best work out and hope for the best. Sometimes they are lucky and the work is appreciated. Yet it is very difficult to predict success ahead of time. Some of the major research papers, including some work that lead to Nobel prizes, were initially rejected. In fact, the number of Nobel-winning research papers that were initially rejected is astounding (Campanario).
Though my arguments appear reasonable enough, consider how hard it is to apply the gold prospector model in real life. Young researchers are often judged by the work they did during the Ph.D.: you only have a few short years that make or break the rest of your career. We are quick to judge a writer based on the market success of his books: it is difficult for us to imagine that some could go from a nobody to a wealthy writer like J. A. Konrath in a couple of years.
We tend to be overconfident in our assessment of gold prospectors. A much easier, and more meaningful, distinction is between gold prospectors and construction workers. Some people are simply not interested in finding gold. They would rather just build houses. Among college professors, you have those who are eager to start yet another research projects (gold prospectors) and those who do research when it is a requirement to keep their job (construction workers). Among programmers, you have those who start their own projects, and those who will never touch code again if they can get a management job. Both types can excel at their jobs, but the spirit is different. Prospectors accept failures as part of their business, whereas construction workers would rather avoid failures entirely.
Reference: Juan Miguel Campanario, Rejecting Nobel class articles and resisting Nobel class discoveries.