Build log
I Built the Wrong Thing First
Most founders build the product. I built the database. It took three years to understand why that was the right mistake.
I built the wrong thing first.
Not the wrong company. Not the wrong idea. The wrong layer. I spent the first year building products: a private equity intelligence platform, a content suite, a CRM. And I kept hitting the same wall. The data model could not hold what I needed it to hold. Relationships between entities were not relationships. They were foreign keys pretending to be. Context collapsed the moment you needed it to travel.
So I stopped building products and built a database.
The Decision Nobody Understood
That sentence sounds clean now. It was not. It felt like failure for a long time. You do not raise a seed round by telling investors you have decided to build infrastructure instead of the thing you told them you were building. You do not explain to people in your life why you are three years in and still do not have a product they can touch.
What I had instead was a graph. A Rust-native, on-premises, three-lane truth architecture that could hold deterministic facts, probabilistic inferences, and hybrid states simultaneously. A database that knew not just what was true, but how true, and when, and why the system believed it.
I called it CongDB. Congregation. Because context gathers.
The Feedback Loop Nobody Warns You About
Here is the thing nobody tells you about building infrastructure: you do not know it is the right decision until much later. Products give you feedback immediately. Users complain, metrics move, something breaks in production and you fix it. Infrastructure gives you silence for a long time and then, one day, you try to build something on top of it and it just works. The feedback loop is years, not days.
That silence is the hardest part. Not the technical problems. The silence.
Building With a Brain That Works Differently
I have a neurodivergent brain. I am not always sure if that is an asset or a liability in this context. On the good days, I can hold the entire stack in my head simultaneously: the graph layer, the entity resolution engine, the client-facing platform, the vertical products, the SPV structure. It all lives in working memory as a single coherent system. On the hard days, the same thing that lets me see the whole system makes it almost impossible to explain any single part of it to anyone else.
Building alone makes that worse. And better. Both.
LLMs are the first cognitive tool I have found that can match the pace. That is not a small thing to say.
Why This Exists
I started Rustcrate because I needed somewhere to put the journey that was not a pitch deck and was not a press release. The Panamorphix newsroom exists for the company. This is for the building: the actual week-by-week experience of constructing something technically hard, commercially unconventional, and personally costly, with a brain that works differently and no co-founder to sense-check the 11pm decisions.
If you are building something similar: infrastructure, not product; long cycle, not short; regulated sector, not consumer: some of this might be useful to you. If you are not, it might still be interesting. I do not know yet. I am going to write it anyway.
That is the other thing about building the wrong thing first.
Eventually you find out it was right.