Substack is constructed as a competitive environment — competitive for attention (likes, restacks), for sales (leaderboards, rising buzzboards), and for ideas (interest, conversation, endorsement.)

Authors build reputation, form creative connection and communities of interest, collaborate, solve problems, develop skills, share knowledge, earn revenue, engage in policy discourse, negotiate business deals, and earn a living from this environment.

False positives and false negatives from an endorsed authentication tool have social impact beyond just an article. Damage to an author’s reputation with no timely recourse affects every reason to use Substack, as does false endorsement of work that may not even have seen original thought.

The treatment of different classes of authors (neurodiverse, ESL, different dialects, specific subjects) also matters for access, equity and diversity of perspective.

On Substack the tool evaluation must be on authenticated Substack prose, with respect to the nuances of authorship (viz. developing an original idea vs machine translating/editing it) and against community impacts.

In the early days of its deployment here, talking about the tool in the abstract — potentially for some other function, using other data — misses the point, and is misleading.