I stopped posting my writing on Substack a while ago — for many of the reasons listed here — but I never bothered to unsubscribe from their product update emails, partly out of laziness and partly because I take a morbid interest in the direction the company is headed. (I’ve been collecting string for what may eventually turn into a video essay about the cultural position of Substack and the disconnect between the stories company executives like to tell and the reality of what the platform actually is.)

The latest update is a new content moderation tool. Or at least that’s how they’re pitching it. And if you know anything about the reasons why a lot of good writers have fled the platform in recent years, you’ll recognize how funny this is.

The details are extremely vague, but the “tool” as such seems to consist of (1) the ability to write house rules about what is and isn’t allowed in the comments section of your publication and (2) a “system” that will supposedly adapt to your moderation decisions and automatically hide posts that violate your rules. All of this is pitched as a way to “create the best possible conditions for cultures of many varieties to flourish online” and put “power in the hands of individuals and communities, rather than platform owners.”

Let’s break that down:

Number (1) is extremely funny because there was absolutely nothing keeping Substack’s users from creating moderation rules before. It’s just that now those rules will show up next to the comment box underneath every post. (Innovation!)

Number (2) is more interesting because it feels like an admission on Substack’s part that moderation is an ugly, thankless task that no one wants to actually be responsible for. This is from their Notes post announcing this whole thing:

And you don’t have to shoulder the moderation burden alone. There’s now a system that learns from what you do, and preemptively hides replies it thinks you’d hide yourself.

Presumably, this means Substack is rolling out some sort of machine learning/AI program that has some discretion when it comes to flagging comments. But the company is very careful not to call this an AI feature. All the references are to “the system” or “a system” that handles these decisions. Substack’s leadership probably recognizes that many of their hardcore users gravitated to the platform precisely because of their branding as the place where content moderation goes to die — an ideological commitment that extends to a refusal to demonetize newsletters operated by actual, swastika-wearing Nazis. As such, Substack has put themselves in an impossible position: They recognize that big-name writers on the platform need help with moderation, but any steps they take to do that risks triggering a backlash from the aggressive free speech warriors they’ve oriented their business around. Any sign of an automated system for content moderation — no matter how tame or limited in scope — will look like censorship to that crowd!

Also, it’s just very funny to me that Substack frames this offloading of content moderation onto users as a grandiose philosophical position BUT simultaneously turns off comments on their official posts. It’s almost like they recognize how awful and exhausting moderation is and don’t want to bother with it. Funny how that works!

P.S. A little update: Sometimes I have ephemeral thoughts that don’t deserve an entire blog post, but also don’t really belong on social media. So I spun-up a new page on the website for those. Check it out if you want! I’ll remind you that it exists now and then.

P.P.S. My All The President’s Men project is still in the works, I promise. It’s been taking far longer than I planned because (gestures vaguely at everything and nothing)