Cobind Retrospective: Twenty Years Later


In the summer of 2003, Bryan Mills and I were working on a variety of projects together when we got a call from a friend about working on an early stage tech stack that his startup was putting together.

That ideation process evolved into the startup we would cofound a year later called Cobind.

The inflection point occurred when I had a dream one night, got up, and said, "we enable anyone to create and download a custom linux distro from the web browser".

At that point, we really just had a set of build scripts that took some parameters on one side and built a custom linux CD- ROM on the other.

A year later, we'd have a browser-based build system capable of slicing and dicing an application, its dependencies, and its branding into a complete product stack built into a CD-ROM downloadable ISO from a build farm.

Cobind DIY Linux Build Tools

And so we set out to build the Cobind DIY Linux build tools: a web-based application that front-ended a build farm for building and downloading a custom CD-ROM containing your application, your branding, and your dependencies all baked into the CD-ROM, which could then be burnt to a CD-ROM drive.

This should sound familiar to anyone working in cloud today, but then it seemed impossible until the demo burnt a CD-ROM.

What Looks Good Might Break The Company

I realized we needed a test case for the build tools The idea of the Cobind Desktop materialized as a minimalist linux distro to solve the testing problem.

Cobind Desktop would almost kill Cobind because the desktop got so popular that we couldn't afford our bandwidth after a single night in the tens-of-thousands of downloads because it was free.

We had a fellowship and in a whirlwind of activity in under a year, we'd given everything we had in terms of time, money, and spirit. We hadn't run out of time or spirit, but Cobind had run out of money.

Failure is the greatest teacher.

Today, Bryan Mills works for Google Pittsburgh while I've ricocheted from startups to R&D labs like an air hockey puck.

The Blameless Post Mortem

At the turn of the century, it was all about kiosks and branding - the physical computing experience.

Novel ideas languish until the ecosystem envelops them and makes them seem early, or as we like to say, only obvious in retrospect. We realized this while entertaining conversations with progeny, rPath, and timesys.

Today, it's bittersweet to realize how linux spins or flavors or derivatives are just a few clicks away when we typed fast and furious to produce the demo redolent of any startup twenty years ago.


We both agree that the following things would have made our startup journey easier:

  1. A coworking community. Emphasis on community, not money.
  2. A salesperson. Beyond the demo we needed to "always be selling" but neither of us had the sales gene.
  3. Funding alternatives. The first money is the most expensive money.

We both are sentimental about what might have been, but share deeply in the realization that nothing could have opened as many doors in our future than the learning that comes from cofounding a startup. We have Fred Gohh and Idea Foundry to thank for that.