Reply Rules?
A questionable idea from Substack
Instead of writing earlier in the day, I seem to be writing even later….
Thanks for all of the congratulations and kind replies to yesterday’s anniversary post.
Also, although I know many of you also read it and have therefore already seen the news, I can’t resist putting in a quick plug for my primary newsletter, From My Bookshelf, which today turned two years old (and therefore has a celebratory half-price subscription sale…). Worth checking out, if I do say so myself.
In other news, Substack today released information about a new feature called “Reply Rules,” which apparently lets writers supply guidance to their readers about the kinds of comments and notes they accept. The e-mail announcement described this in rather glowing terms as another way of helping writers build their own personalized communities, helping “even more people to be fully themselves and at home on Substack.”
I take a somewhat dimmer view of the matter.
Here is an excerpt from Substack’s description of the new feature:
Anyone can now set their own Reply Rules, which will appear to anyone replying to them…. There is now a system that learns every time you hide a reply and preemptively hides replies it thinks you would hide yourself (you can always see these and un-hide them, which the system will also learn from). Soon, this system will read your Reply Rules when you set them and use them as guidance for its actions, in addition to what it sees you do.
They call this a way of using “contemporary technologies” to improve the Substack experience.
Two words, however, are conspicuously absent from the announcement: artificial intelligence.
This new system is obviously a way of using AI to screen and block replies, based on the “prompt” you supply, and adapting its actions by drawing upon your responses. Why don’t they describe this as an innovative use of new AI technology?
Presumably—or so I suspect—because they know that giving it the AI label would cause many users to react skeptically.
My own view is that if a new, AI-based feature is introduced, writers should be told clearly that it relies upon artificial intelligence. For some, that knowledge could affect their willingness to use it. Substack probably has a higher than average percentage of users who are reluctant to promote the further use of AI because they object to the ways in which its development has drawn upon the words of writers without their consent.
As for me, I don’t intend to use “reply rules.” I’m a moderate AI skeptic, but my real reason is much simpler: if you take the time to read what I’ve written and reply to it, and if you exercise ordinary civility and courtesy, then I want to see what you wrote and respond to it personally. My “reply rule” would therefore in any case be no more complicated than, “Be nice.”
Which should not really need saying.
I really enjoy Substack, the pros massively outweigh the cons and I especially love the banter, interactions, etc. But my goodness, with every new 'innovation' my head's about to explode, I can barely process it all, it's almost like they're trying to weed out Luddites like me or those without the patience and tolerance to put up with this nonsense. Whatever next?