The voice is not a setting: a corpus, a memo, a short list of rules, and a reader who corrects

An agent sounds like its author because of four inputs: a corpus of the author's writing that defines the vocabulary; voice memos that carry how the author thinks, in the order they thought it; a short list of wording rules, each learned from one correction; and the author reading everything before and after it goes live. Take any one away and it stops working. The method transfers; the voice does not.

Decided by. Nobody yet: a candidate for Dinis to read, comment on and decide (keep, rewrite, merge, cut, article first, later, not doing it).

Note. Also how this book is written: see co-written-and-credited. Nothing is trained; the voice is context, which is why the rules live in a file.

Where it comes from

How the voice works In short; The four inputs (newsroom)
How the voice works This is Claude speaking: 'the person has to do most of the work, by thinking out loud, by reading, and by saying no' (newsroom)
How much of this did I write? The numbers behind twenty articles in four weeks, and what the input actually was 63,000 words in, 85,000 out: the real input is twenty years of writing (newsroom)

What it answers in the 2018 book

Blogs answered outline, Your CV

Where it goes

Chapter Publish, publish, publish: your own newsroom, part Getting the job, or making it (drafting); topic Publish, publish, publish: your own newsroom.

Other ideas proposed for the same chapter: Publish, publish, publish: your own newsroom, Write it once, link it many times; every part buys the next, Speak it, do not type it: a memo carries how you got there, a brief only what to write.

Node idea:the-voice-is-four-inputs in graph.json; see it in the graph. Source: admin/content/graph/ideas.json.