Why a touch of secrecy can help creative work

Though I am a long-time blogger and I spend most of my day talking or writing to other people… I am also quite secretive about the research that I am doing.

There are reasons to be secretive that are bogus. The primary one is that you are afraid others might steal your ideas. That’s overwhelming bogus because ideas are cheap. I can write out long lists of good ideas that are almost sure to work, but that I will never pursue for lack of time, energy, competence, and interest. Coming up with the idea is easy, executing is much harder. If you are running so low on new ideas that you need to take other people’s projects and compete with them, then you are doing something wrong.

If you have fierce competitors that might quickly adopt your ideas and run with them… then maybe you are working on the wrong project in the first place.

No. I am sometimes secretive for other reasons…

  • Sharing your project ideas is rewarding in itself. But it is not a good policy to reward yourself for work that has not been completed. Look around you at people who never finish anything… they are often quite open about all the marvelous projects they have going. It is so much easier to start something than to complete. Don’t reward yourself early.
  • Most things I work on end up not working the way I expect. So I don’t want overcommit to a narrative that might need changing. Once you have begun explaining what you are doing to others, it gets harder to change the project. You have to go back and say “you know what? I was wrong.” Having to admit you are wrong is fine, but it creates friction in your mind.
  • There is communication fatigue. If you keep announcing something that is not yet there, you risk eroding the interest of people who are expecting your work.
  • Early work often looks and sounds naive. It takes time to find the right way to present your work. You do not want to ruin your first impression… and the first impression does matter.
  • In the public eye, we tend to think like others… human beings are gregarious and we have a strong tendency to want to “fit in”. Isolating yourself from the state-of-the-art can be the best way to “ignore” the state-of-the-art, and that’s sometimes just what is needed. You don’t want people to tell you that what you are going after “has been done before”. You want to see what you can come up as independently as possible… otherwise, you risk just being stuck thinking like everyone else.

Of course, I am not always secretive… far from it… I am extremely transparent… you can literally follow up almost everything I do programming-wise on GitHub. I often post my research papers on arXiv months before I announce it.

Secrecy comes at a cost…

  • It is harder to recruit the right people to your cause if your work is done in secret. If you need more ressources, it is pays to drop secrecy.
  • Secrecy introduces friction of its own that might slow you down.
  • Early feedback, even critical feedback, can be extremely valuable.

So there is no hard rule about when to be secretive. You have to make judgment calls on a case-by-case basis. I tend to be secretive when…

  • I’m starting something that is “new” for me. Maybe it is a problem I am unfamiliar with.
  • Having more people would not obviously help. Maybe it is the case that I already have great collaborators.
  • I expect to be working for a long time still on the project.

I tend to be more transparent regarding work on problems that I understand well.

Leave a Reply

Your email address will not be published. Required fields are marked *

To create code blocks or other preformatted text, indent by four spaces:

    This will be displayed in a monospaced font. The first four 
    spaces will be stripped off, but all other whitespace
    will be preserved.
    
    Markdown is turned off in code blocks:
     [This is not a link](http://example.com)

To create not a block, but an inline code span, use backticks:

Here is some inline `code`.

For more help see http://daringfireball.net/projects/markdown/syntax