Verdion Soft Launches Today

Verdion is Born

I can count on the fingers of one hand the number of projects which I have shepherded from concept through architecture, design, code, and launch. I need even fewer digits to enumerate the number of those which have been even moderately successful.

Regardless of the future of Verdion, it’s definitely the one in which I believe the most.

What I’ve Accomplished

I won’t rehash what I’ve built, the industry problem it addresses, the domain of myriad problems it can solve, or anything else you can find elsewhere on this nascent site. Instead, I’m going to do something I don’t think I’ve ever done on its face and without shame:

I’m going to pat myself on the back.

The Lead-Up

During what I will say has been the most personally trying month of my life, I organized thoughts which had been percolating for over a year. I synthesized them into notes. I added notes which existed in my journals, and which existed only in my head. I synthesized those into an architecture. Those became a design.

I brutally scythed 90% of that design into a bucket entitled Future Work. If I hadn’t, I would have slipped into a mode with which I’m very familiar, where I design and code and design and code until one day, someone has released a version of what I’m building which isn’t as perfect — but it’s shipped. Then I become discouraged and my project languishes in a private repository graveyard whose population numbers I shan’t mention here.

The Execution (?)

Then I coded an MVP. I tinkered with it until it was actually working — I had built a superior adversarially convergent, model-agnostic reasoning platform! Then I proceeded to optimize. To add features.

I realized I was falling back into the same trap I described above - so I said, “ENOUGH.” I declared a code freeze and I became an author for a week.

The Writing

I wrote. I revised. I wrote more. I cut. I cut brutally, and I still probably haven’t cut enough. But, I shipped — I have a working, configurable reasoning platform in a repository. I shipped this website, with an MVP Proof of Mechanism and an architectural summary. What I hope is an entertaining origin story, spiritually continued here. A pre-print, published on Zenodo (and hopefully ArXiv in the next 24 hours). I wrote announcements on various websites, and I sent messages to three people I enjoyed working with and who I know have passionate opinions about AI.

Wherever you go, there you are

I’m not going to ask myself a bunch of easy questions, or ruminate upon what I’ve learned (after all, the more I learn, the more I know how much more I have to learn). But here we are. At the beginning of something special. It works, and more importantly, it’s shipped, and it’s the first system of its kind. It’s only going to grow from here.

Now — back to work.