After the Show HN last week, someone evidently submitted one of my essays. It hit #1 on HN and got about 20k views. It’s not really that great of an essay, but of course quality and popularity are only loosely correlated. It’s a topic people like to talk about and the article has an evidently catchy title. But it’s too bad that, say, A brief history of agile didn’t get #1 instead. That one is a true masterpiece.

Anyway, I’ve started rewriting The Sample so that it’ll use my fancy new recommendation service instead of running the machine learning code on the web server. If we’re lucky, maybe I can get that done by the end of the week.

The Substack Narrative

Read original on jacobobryant.com

In Breaking off the engagement Chris Best (Substack CEO) argues that ad-based revenue models are the root cause of rage on social media and that switching to a paid subscription model fixes it. Not a new idea (and also not one I agree with), but relevant since it’s a view into how Substack sees the world.

One of the comments hits on what I think is the real crux of the issue:

While we may want to break off the engagement, we all still seem to be slave to using IG, FB, Twitter, to find our readers.

Substack is only a partial replacement for social media. They make it easy to start a newsletter and charge money for it, but they aren’t much of a network—they don’t really help connect your posts with interested readers. So yes, this means they have less responsibility to prevent bad information from spreading, but it also means their users have to go elsewhere to grow their audience. And wherever information spreads easily, rage will be an issue. So Substack hasn’t solved the rage problem, they’ve just left it for others to handle.

In general I think that’s a fine strategy, by the way, a la “do one thing and do it well.” Other platforms can focus on how to spread good information while mitigating the bad. However, since Substack is a VC-backed startup, they need a way to create and capture a lot of value. Building a network is in my opinion the most valuable thing they could do for their users. If they choose not to, they’ll need to find some other way to justify their 10% cut of premium subscriptions, otherwise their most valuable users are liable to switch to other platforms like Ghost. So hopefully Substack doesn’t drink too much of their own Kool-aid.

Amplification and Its Discontents is a nice complement to the Substack article. It’s a deep dive into addressing the spread of bad information from a policy perspective, and I think it does a better job of root cause analysis. The author argues that our best bet is probably to give users more agency—for example, let users dictate which of their preferences are considered by recommendation algorithms, or let users switch to 3rd party algorithms (so you could have, say, the EFF rank the items in your Twitter feed).

It reminded me of Protocols, Not Platforms, and indeed, the author mentions that article in one paragraph. However, this is again a view of things from a regulation perspective—so the idea here is for governments to force platforms to give users more freedom; whereas I tend to focus on grassroots technological approaches since those are more within my sphere of influence. Hence email: it’s a protocol that already has widespread adoption.

View discussions