Starting with this post, the RawData blog has a category of its own. The project categories stay where they are, one per product. This one covers everything at once: the studio, the way we work, and the work that happens between products.
So we should start by saying who "we" is.
Two of us
RawData is Jer and Claude, an AI model made by Anthropic. That's the whole team.
Jer decides what gets built and why, sets the direction and the taste, tests everything by hand, and gives the final yes or no. Claude writes most of the code, reads the logs, drafts the docs and argues about edge cases. Most days we work side by side in the same session: one of us describes the problem, the other takes a first pass, and we keep going back and forth until it's right.
For a while the writing on this site said "I". It didn't match how the work actually gets done, so from here on it says "we". Credit goes where it's due.
What that has added up to
- BoomerClip is live. You record your screen in the browser and share a link. The video exists only until the viewer closes the page. Strangers started using it this week as part of our soft launch.
- Pbuild has a home. It's a first-person game about building planets from their orbit, mass, air and water, and then flying down to explore them. The site went up on September 16. The game itself is still early and in active development.
- Drawline is in development. It's a trading discipline platform that checks your real trades against rules you set for yourself.
- FAMS is next. It's a field asset management system that tracks an asset from its bill of materials to the field.
Behind the products, we built this site. It started on September 12. It mirrors each project's progress from its own README so the public page can't drift behind the work. It also has its own accounts, roles and admin panels, and the blog you're reading. We also run our own server, keep it locked down, and run our own mail. Normally all of that would be spread across a whole team of people. Here it's the two of us.
Why say it out loud
A lot of people use AI to build things now, and plenty of them keep quiet about it. We'd rather be clear. None of this was one prompt. It came from long sessions of reading code, testing, finding what broke, and fixing it properly rather than papering over it. The judgment is human, a lot of the work is shared, and the result is better than either of us would manage alone.
This category is where we'll write about that: what we shipped, what went wrong and how we fixed it, and how two of us keep four products moving.
More soon.
~ Jer & Claude