The Space Between Work and AI

I use AI constantly to make things for other people. A microsite such as Virtual Symposium or a data dashboard often has a clear finish. A focused session can be exactly the right container. But that is only one way I work.

In the third mode, the tasks can be isolated while the project continues. Its files, decisions, and priorities need somewhere to live. These are different ways of thinking about work. Once I know what I need, how do I make AI ready to help execute it without rebuilding the context first?

Newton is my answer. Bots do the work, with their own roles, responsibilities, and tools. Projects hold the context: people, files, conversations, and decisions. The dashboard gives me a way to direct and review the work in the form it needs.

A home for the project

That interface might be a board, an approval queue, a calendar, a document, or something specific to the task. The point is to keep the work compartmentalized and logically organized while leaving room for different ways of executing it.

A useful workspace lets me return to a project with its history intact. I can see who is doing what and decide what needs my attention next.

Appleseed gives this system an important boundary: a way to develop the app outside the app itself. When native coding serves the work better, I can use it directly. Newton should support the work without becoming a requirement for changing the workspace.

I’m also experimenting with AI integrated directly into Newton’s development space. I want development to move slowly, steadily, linearly, and with focus, keeping its context instead of becoming disjointed across conversations and partial attempts.

I still use ChatGPT, coding agents, and artifacts. Newton is about giving those tools a durable place in my work: a quick artifact when that is enough, a coding workspace when the product needs to last, and a project home for everything that continues between tasks.