The AI-augmented standard
What a family of repositories on AI-augmented work shares, as far as the evidence goes. A claim about a person and a model is set against each alone, and every member states its scope in a table, once. It extends reader-ergonomics: read that one first. This page adds only what a person, a model and a machine on one project change.
Grades mean what they mean in reader-ergonomics: measured, supported with caveat, convention, not supported. Each principle states its grade and never claims more.
How to use it
- Pick the shape from shapes/, and name it in the document with a
**Shape:**line that links it. - Write to the principles below, and to those of reader-ergonomics, which still hold.
- Check: run the standard's checks on the repository.
scope-tablereads each document that declares the shape,three-armsreads every table, andscope-overlapcompares the members.
Principles
Claims: saying that a person with a model does better
| Principle | Claim | Grade |
|---|---|---|
| Compare with each part alone | a claim that a person with a model does better sets the combination against the person alone and the model alone. | measured for the finding that a combination does worse, on average, than the better part alone (106 experiments, secondary); convention for the rule that every claim carries the three arms |
Scope: what each member of the family owns
| Principle | Claim | Grade |
|---|---|---|
| One owner per claim | a thing is in the scope of one member of the family, and every other member links it. | convention, argued from the standard's one-place-per-fact principle; untested for scope tables |
| A scope table says why | a member's scope is a table of in, later and out rows, and every row says why. | convention, untested |
Where the evidence is
Every principle links its review in evidence/, and every work cited is listed once in the bibliography.