About the name: Forge is temporary. When I first wrote the specification I asked an AI for working-name suggestions and kept one because the folder needed to be called something. There was no serious naming process behind it. If this becomes a real product, I would expect it to have another name.
Why I have complicated feelings about this one
I wrote the original specification months before several ideas in it became normal parts of AI coding tools: agents working around a shared project, visual workspaces, browser-like inspection, delegated implementation, review loops, and a layer above individual coding agents that understands the work as a whole.
Then I designed far too much of it.
The specification accumulated dependencies on systems that barely existed. Some of those systems were apparently also going to be my responsibility to build. It became increasingly coherent on paper and increasingly difficult to begin in reality, so I froze it.
Watching parts of the general direction appear in other tools later was painful. I do not mean that as “I invented all of this first.” I did not. Plenty of assumptions in the old spec were wrong, and other people solved similar problems in ways I would not have predicted.
What bothered me was more personal. Having the product instinct early did not matter much if I never turned it into something people could use.
That has happened to me often enough that I am trying to stop treating follow-through as a problem for future me.
What I still care about
Individual coding agents have become much more capable. The part I still find interesting is the layer above them: how a piece of work is understood, split up, delegated, reviewed, remembered, and presented to the person directing it.
I do not want another methodology that replaces coding with ceremony. I want capable agents to be easier to direct while keeping engineering judgment with the human who is responsible for the result.
I am returning to the idea with much less appetite for empire building.
The old specification does not need to survive intact. I need to find the smallest useful interaction that proves the idea is worth continuing.
I do not know exactly what that is yet. That is fine for now.