Unicorn Jobs

Once I had built up enough savings, I jumped out and tried (and failed) to start my own company before joining up with a TechStars team. Worked out pretty well in the end: if I'd been in a "good enough" tech job, I probably wouldn't have taken that leap and wound up in an awesome job with a great team.

I've worked in enough companies and done enough consulting and advising to have some thoughts about this.

First, stop obsessing over binary distinctions that don't exist like "unicorn job". It's like "being rich". You won't get there by obsessing about it. There are shades of gray on this one.

Second, it's better to ask for forgiveness than permission. If Guido only spends 50% of his time on Python, that's probably because he doesn't want to spend more than 50% of his time on Python. (I doubt he explicitly asked for permission to spend 50 vs. 75 or 25 percent of his time on it.) As with the creative arts, you'll make better stuff if you're part of the world, rather than completely isolated. Work for other people-- and, except when its interests contradict your own, for the benefit of your firm-- but on your own terms.

Third, "political forces" always exist but are not always irresistible. Yes, sometimes you have to work on things that wouldn't be your first pick if you were unconstrained, but that build your credibility or advance your career. Accept that. Everyone has to do some selling and alliance-building. Do that work, do it well, and make allies. It's important to stop hating that aspect of work; it's part of the game, too. It's actually fun once you're halfway good at it, and if someone like me can develop those skills then anyone can.

Fourth, make sure to find people, companies, and managers whose passions match your own. Don't try to sell a Java shop on Python. Find a company that uses Python, if that's what you want. It's much easier in the long run to keep looking until you find a company that agrees with your ideology, than it is to try to change everything.

Fifth, never work on stuff that isn't interesting to you or good for your career. If you get assigned work that hurts your career and bores you, work quickly to find someone powerful who can use you for something else. (You might get fired in this process; accept that risk. It's better than being taken advantage of for 5+ years.)

Sixth, don't use the words "personal project". It may be something you initiated, but it's probably useful to someone. Find some way (possibly an abstract one) that your project is useful to someone else, and preferably your company as a whole. Again, make allies. Ask others for advice (this is a big one!) and let them have input into your work (but don't compromise on the vision in a major way). Treat people you work with as first customers, not as obstacles. Don't make it obvious that you're trying to take control of the show. People will ask you to join them long before they'll be ready to be led by you.

There's a lot more that I can say, but I hope I've made my points clear. You don't get this "unicorn" job or environment by becoming a brand-name engineer (it works in reverse; the great engineers get the autonomy and use it to become really good at something). Rather, you get creative at selling the work you want to do, make allies, and find companies and managers who are going in the same direction. I don't want to make it sound easy, but you don't have to be a celebrity to pull it off. You need to know what you want and work hard to get it. You also need to learn the difference between asking for permission (never do it) and consulting others (often do it). Don't ask "Can I do this?" (Counterintuitive fact: most bosses dislike being explicitly asked for permission. You're not doing him a favor, but asking him to take on risk-- for mostly your benefit.) Say, "This is what I want to do; how would you recommend that I go about it and make it maximally useful to those around me?"

Of course, nothing is black and white, and rationalization is just as strong an intoxicant as overreaching corporate policies are.