The Knowledge That Shouldn't Be Hard to Find

I started Plinth because of all the things I had to learn with no good way of documenting the journey.

If you've ever tried researching “how-to's” for things like CAD automation — real automation, something tailored to the way you work — you know the problem. Most things are a sales pitch, or they're fragmented to the point that they're not directly usable or applicable to what you need. You settle for better than it was, not as good as it could be.

Over the years, I've learned bits and pieces from all kinds of people who took the time to document their journey. Blog posts, forum threads, GitHub repos, YouTube videos — small contributions from strangers that saved me weeks of frustration. This is my way of paying that forward.

Manufacturing is one of the last industries where knowledge is still guarded like a trade secret. The people who know how to do things tend to keep it close. The result? Everyone reinvents the wheel. Alone. Slowly. Painfully.

But here's the thing about manufacturing: it's all physics. There's a calculation for it. We may not know the calculation yet, but it exists. And that's actually good news, because it means the answers are discoverable. They can be derived, tested, and shared.

LLMs should be the perfect tool for this — but they're not, yet. They have some training on manufacturing, but the real domain knowledge, the stuff that matters on the shop floor... that was never in the training data. The opportunity isn't to ask an LLM for the answer. It's to use LLMs as tools to help build, test, and document the answers. Code generation. Scripting automation. Parsing domain-specific information. The LLM is the assistant, not the Engineer.

Plinth is a resource hub and publication for builders working at the intersection of Manufacturing, AI, and Automation. Everything published here is something I had a need for, and it either met that need or I learned a lesson that led to meeting it.