Building A MLOps Startup in Public
Why we chose not to be a stealth startup
Have you ever seen a great talk or read a wonderfully insightful article about a new, emerging technology and excitedly went to Twitter/LinkedIn to check it out further, only to be confronted with the inevitable ‘Stealth Startup’ title to block you from accessing further information.
Having been (somewhat) part of the startup world for a few years now, I’ve recently wondered why founders are so hesitant to share their ideas with the world, no matter how premature the actual execution/development of their idea is. This seems to be particularly a Silicon Valley thing—perhaps the relative abundance of VC funding makes founders feel less pressure to build a relevant audience around their ideas early on. Or perhaps there is so much competition that it feeds into the fear that their ideas might be stolen. In any case, it seems to be almost an egregious oversight to gather feedback as soon as possible.
Now of course, I’m probably not accounting for all possible reasons that a startup might go into stealth. Having said that, I’m sure there are many founders out there (we’ve made the same mistake) who fall into the trap of trying to create the ‘perfect’ product before launch and needlessly delay kick-starting that feedback cycle.
Contrast this to the Indie Hacker community: Most people there understand there products are going to start out small, and they need eyeballs on it way before its key-ready. There are numerous advantages to build an interested audience early, and get hands-on feedback as soon as possible:
You gain an early audience that can spread the word.
Your user persona gets more refined and nuanced as more relevant people give you feedback.
You can course correct quicker to avoid the most low-hanging, simple mistakes you would have inevitably made without any feedback.
You build up good practices early to not only record feedback, but incorporate it in development cycles through smart internal processes.
Some people might criticize you, or laugh you off (I’m looking at you, HackerNews comment thread), but that only adds to the chapter of the book you’re going to write about your success eventually.
It holds you and your team accountable for their own efforts.
In light of this, the ZenML team have decided that we won’t make the same mistake. We’re building our tool out in the open. What does that mean? Well, we’re open-source in the first place, so not much changes, as pretty much everything is available publicly.
However, we’ve decided to put an extra emphasis to record this journey as it continues. We’re going to use this newsletter to publish regular updates, not just on a product level, but on a company level—the sort of stuff you’d normally communicate only to investors. We feel that level of transparency will give us accountability, motivate us to deliver faster results, and hopefully result in more people paying attention and giving feedback, ultimately the most important thing at the seed stage.
We plan to ramp this up as we go. Hopefully starting with a monthly update, and if more people are interested at more frequent intervals. This post is the first in (hopefully) a long journey. The ending of this journey is unknown, which makes it exciting. It might end in brutal failure, or blinding success. Either way, the goal is to be honest, so that its beneficial to you, dear reader, if you one day decide to go down a similar route. Subscribe if you’re interested in finding out how an opensource, tech startup is built out in public, or if if you’re just curious.
If you are interested in what we are building, and not just in the journey, then do take a look at this newsletter’s accompanying blog post, that goes into the details of the challenges we are taking on with ZenML.
Thank you and stay safe. See you soon!
H