A number of people have asked me in the past week why Buttondown isn't open-source, given:

  1. the love and overt financial commitment we have to open source
  2. the increasing number of "open source startups" (Cal, Maybe, Lago coming to mind)

I preface this answer with the fact that this is coming from my personal blog and not Buttondown's blog, and maybe a year from now either my philosophy or my math has changed on the matter, but the answer right now is something like the following:

  1. To be open source in a meaningful sense, the process of working on Buttondown would have to change pretty drastically.
  2. The number of users who benefit from Buttondown being open source is fairly small.
  3. The theoretical increased rate at which the product could improve is not particularly high. [1]
  4. I think running a company (even an idiosyncratic one) is a fairly distinct skillset and bag of incentives from open source maintainership, and as bad as I am at the former I think I'd be even worse at the latter.

I write the above in pencil, not pen — I think points one and three could very well change in the future. But right now I think the vast majority of startups who "go open source" do so more as a marketing tactic than as a genuine belief in open source, and I think the best way for Buttondown to "be open source" is to open source components (e.g.) and contribute financially to the software on which it relies.


  1. I find it instructive to look at Posthog's commit history. ↩︎

Lightning bolt
Subscribe to my newsletter

I publish monthly roundups of everything I've written, plus pictures of my corgi.
© 2024 Justin Duke · All rights reserved · have a nice day.