Feedback
Think something here is wrong?
Every row is published with its source, so you can check it. Three kinds of disagreement move this record: a statement that is wrongly recorded, a measure or a line in the singularity test that should be different, and an error on a page. Each leaves a trace that the change can cite.
The three forms
A statement is miscoded, misdated or misquoted
A row in the statements table that says something its source does not, carries the wrong date, or is coded wrongly: who it endorses AGI for, the year it names, which question it raises. Give the row's id or the person and the date, what is wrong, and the source where the words are.
A measure, a line or a reading should be different
One of the four parts of the singularity test rests on a measure and a line fixed in advance, and on readings of that measure. If the measure is the wrong one, the line is in the wrong place, or a reading is wrong, say what you would change it to and point at the published source that supports it; a line moves on a source and a dated reason, never on a feeling.
An error on a page
A number, a date, a name or a claim on any page that is wrong. Say which page, what it says, and what it should say, with a source where there is one.
Each form opens an issue in the public repository with the right questions already in it. Issues are used because they are durable and quotable, and folding one into the record is a change that can cite it. Or write a plain issue: the page or the row, what you would change, and why is exactly enough.
What happens to it
I rule on it. Some disagreements change the record; some are answered and left as they are. Either way the issue records the ruling and the reason.
A corrected statement gets a dated note in its row that cites the issue; the search that found it stays in the search log. A changed measure, line or rule is a dated amendment with its reason, never a quiet edit, because the test's lines were fixed before the numbers were looked at. A new reading is a row added with its source. The expert panel is fixed, so that it cannot drift toward the story; adding or removing a person is an amendment too.
The repository is at github.com/jfredson/agi-zeitgeist. The data is free to reuse with credit (CC BY 4.0) and the scripts are under the MIT licence; how to cite is on the data page.