Tens of thousands of articles have been written about how to start a startup once you have an idea. However, very few of these talk about how to live a startup career that will generate good ideas, even though this is really the question most aspiring entrepreneurs need to answer.Here’s some thoughts on how to stack the deck, drawn from my own experience doing speech recognition research at Stanford, running a YC-backed startup over a year in the space, and observing many of my YC and Stanford colleagues do similar things.
Most well-known entrepreneurs had a track record of small successes years before they broke out. Andrew Ng and Daphne `Koller were already successful researchers well before they started Coursera. While it’s impossible to guarantee success, if you’re not practicing entrepreneurship, it’s highly unlikely that you’ll get there.
The purpose of practice is not necessarily to start a company at that time, but to get good enough in your role and your market that you can credibly convert your ideas - eg. free high-quality education online - into opportunities - eg. a high-quality education startup founded by well-known Stanford professors. For some ideas, there’s a low expertise bar required to turn them into opportunities and you can do so in the process of running a startup - Segment learned about the need for a universal customer API during YC. For others, though, the bar is much higher - Coursera couldn’t have been started by students .
Practicing entrepreneurship boils down to making something people want and getting it in front of them. Like ice skating, this requires first choosing a market and then making a commitment to work in that space until a product is shipped, at least a six-month process, but sometimes years depending on the field.
These are two distinct periods, and behaviors in one phase aren’t appropriate for the other. Choosing a market should feel close to a BFS, and shipping a product should feel close to a DFS. In the first phase, you’ll be heads-up - cautious, thoughtful and broadly-read. In the second phase, you’ll be heads-down - impatient, willing to compromise and purposely ignoring competitors, no matter what challenges or distractions you may encounter.
Both phases are influential to your success. Your market choice will be the single largest factor in your project’s overall impact, but you won’t be good enough or motivated enough to generate opportunities in the market until you’re an expert in it, and that takes time.
Most entrepreneurs are exposed to markets unconsciously based on personal biography, and don’t actually do a BFS. Tim Westergren was working on matching musical scores to films well before he even thought about turning his process into Pandora. However, sticking only to your current interests only gets you a random chance they will happen to coincide with a rapidly changing market.
Instead, engaged entrepreneurs and researchers pick their markets consciously and then invest to become experts in them, regardless of whether they have an initial personal connection or experience. This two-step dance is much harder - you need to be both a broadly-learned and well-connected investor and then an aggressive, user-focused product expert, but it gives you the best chance of finding really big ideas within your lifetime, rather than just having a small chance of stumbling on them.
When asked this question, a surprising number of VCs will shrug and say “no one knows.” While it’s impossible to predict every successful startup, running a startup career is similar to running an academic career, and there are plenty of researchers who get good at this.
 In the past, it used to be that most companies required large teams and had to be started by experienced leaders, but computers changed many of these rules. Both types of companies exist today and a surprising number of people in the Valley still get them confused.