In my words.

George makes high-level intent into a durable object that can survive and govern the changing machinery required to realize it.

The work remains itself even when the machinery underneath it changes.

That is the part I care about. Models, tools, executors, connectors, and implementation plans can change underneath a job without automatically changing what the job is.

What explains what George is. Why explains the problem. How shows the machinery. If it sounds too neat, I call bullshit has the records.