Against Substack

The one thing blogging didn't need: centralization.

Why Substack Works

It pains me to write this, but it’s necessary. Blogging is dead. There’s really no way around it. The Golden Era of people having their own information spaces is gone and newsletters are in. It’s clear to anyone that these days almost no one has a blog and almost everyone has a Substack. This is the new model.

It took me a long time to admit it. Not that blogging was dead (that’s been obvious for about two or three years now), but that a change to Substack would be necessary if anyone was to read what I have to say. With Substack up and blogs down, I suspect the next victim is easily going to be emags. This is despite the fact that some attentive observers pointed out that the one thing Substack lacks, the ability to pay for a bundle of peoples’ opinions as one instead of paying each commentator separately, is present in magazines and newspapers, but not present on Substack. Of course, most emags don’t operate on that model, instead providing free content in one centralized place to readers and payment and guaranteed readership to potential writers. What will kill them will be the same thing as blogging: convenience.

In the Substack model, convenience is offered in the form of centralization. As a content aggregator, Substack’s primary asset is its ability to consolidate the blogosphere into one place. This was going to be inevitable. YouTube did this with internet videos, Valve did it with video games, and Spotify did it with music. Consolidation isn’t necessarily a bad thing.

Many wonder why it didn’t happen earlier. Readers have wanted this for a long time. The rise of RSS feeds (which are similarly, now dead) as the OG content aggregators was in part due to the inability of readers to manage the number of updates their favorite blogs were having online. Wordpress, Medium, and others allowed you to make an account to follow your favorite blogs, but then you had to have a separate account for each. Readers were never the problem, writers were, and the reason why writers never coordinated to go to one place until Substack is relatively straightforward: it’s actually rather difficult to monetize a blog and actually rather easy to monetize a Substack. Before Substack, the model was set up something through Paypal, do the pain-in-the-ass Medium subscription thing, or (later) have people subscribe to your Patreon. From the perspective to the blogging platforms, there was never an incentive for writers to switch their content from one blogging space to another due to the fact that none of the platforms incentivized you to do so.

Substack does. Through subscriptions.

---

This new form of blogging is ideal for most people on the individual level. It’s convenient for both readers who no longer want to have to track down all of the different threads of information they’re trying to follow and for writers who finally want to sit in one stable position where they can get paid (the popularity of the platform also means that for once, people are reading blogs again). In a sense, Substack’s model will succeed for the two simple reasons that it’s accessible to everyone and it’s convenient.

But personally, I’ve had some hang-ups. McDonald’s is accessible and convenient, too; and while I wouldn’t go so far as to say it’s going to be as much of a disaster for society as McDonald’s has been, I do think some of the issues that centralization in any sort of information structure presents should seriously be considered. So, in brief, I’ll outline my main hangups and you can tell me, maybe on Twitter, what you think.

Centralization at the Cost of Novelty

The point I’d like to make here is broad and will take more convincing, most likely in the form of another blogpost I have already written and am planning to release the second the project I have been working on with my PhD advisor is turned into a pre-print. The main dynamic to outline here (and the dynamic identified in the model presented in my project) is that while centralization allows for information to be found, identified, and aggregated, it often leads to a loss of novelty. Why? Broadly speaking, centralization creates consensus. We see consensus being driven on Twitter, Reddit, Facebook and other websites at the cost of genuine clustering. On the one hand you might object here and tell me that the people on Twitter or Facebook have basically done everything except achieve consensus, and you would be right, but only if by consensus you quite literally meant agreement. To support your point, you could even point to polarization as evidence that consensus has not been achieved, and yet again, you would be right. If you meant agreement. But when I say consensus, I don’t mean agreement on what is right or wrong, so much as I mean agreement on what is relevant for discussion.

This is the sort of phenomenon that has been popularly discussed as “Seeing the Smoke,” a throwback to a psychology experiment from the 1960s which placed groups of students in a lab under the premise that they would be filling out a questionnaire. As the students filled out the questionnaire, researchers pumped smoke into the room. In one version of the experiment, students were placed in the room with two confederates who were in on the experiment. These confederates were instructed to ignore the smoke and continue filling out the questionnaire. Imagine for yourself what you would do if you went into a strange room on campus and smoke started piling in the door. You would obviously go and check it out, and most people did - except when the confederates simply ignored the smoke. They could have very many disagreements about what was going on in the questionnaire, but so long as they’re talking about the questionnaire, they’re missing the forest for the trees.

In networks, this class of phenomenon manifests when individuals come to a consensus together. Under centralized or more connected networks, consensus is achieved more rapidly than in non-connected or decentralized networks. In real life, this is the same reason we sequester our juries or basically expect that they lived as outsiders from public opinion (surely the jurists in the Kyle Rittenhouse case had to have been from the Moon in order to have no awareness of the context in which their case arose). We might say that consensus might not be so bad if everyone is thinking the right thing or so long as there is a bit of diversity, but a stickier problem is that once consensus is achieved in centralized systems, it’s maintained much longer and is much harder to get rid of than in decentralized systems. This is to say that once the network has made up its mind, it can be very difficult to introduce novelty into the population.

But is it really the case that Substack can drive consensus in this manner and how are they doing it? I think so, through several mechanisms. The first mechanism by which consensus is introduced is that people on Substack are certainly going to copy what others on Substack are doing. Like Twitter, and there is perhaps too much overlap between here and Twitter, everyone can sort of see what everyone else is talking about here (aside from the writers who lock their pages to unpaid subscribers, many of whom I bet have probably already commented on Ukraine at a minimum). I will say that I don’t expect this to necessarily be as bad of an issue as it could be, but I think it will be more difficult for specific sorts of concept-driven blogs to cluster together in isolated topic areas than it was for them to cluster together during the days of blogs.

The second mechanism is that the scope of what constitutes an audience has greatly grown. There’s the general expectation all writers require a niche, but that they don’t necessarily want to be irrelevant. In the days of blogging, your main audience was often the other people online who were into the things you were into. Once again, we had clusters, but in the days of Substack, your main audience is everyone. This is a bigger issue than the one presented above. You might think that bloggers altering their writing to fit more and more readers instead of fewer and fewer arcane knowledge holders is a good idea, but this is again something that comes at a cost to novelty. There’s a reason difficult to describe specialist concepts and terms exist, and it certainly isn’t the case that it’s because they’re useful to everyone. While blogs back in the day could cater to the small clusters of people who understood what they were talking about (and to other bloggers talking about the same thing) specifically because there was no expectation other people would care about these things, such a model will be disastrous in the Substack ecosystem. Dumbing things down is an easy way to kill novelty.

The third mechanism by which Substack will drive consensus is one that I worry about the most, and for that reason I’ve given it its own section below.

Patronage and the Expectations of Meeting Again

As a writer, I hope you all enjoy this. In fact, I hope you subscribe to my newsletter after reading this. It would really be nice. In fact, it would be even nicer if you threw me a couple of dollars - that might motivate me to write even more. In fact, that would motivate me to write even more and it would motivate me to write more of the things that you want to see. It’s here we have two issues at the heart of what makes Substack a success: the expectation that I keep writing on here and the expectation that you pay me to do so.

On the blogosphere, it was always nice to get paid. Certainly no one would turn money down. But money was inconsistent and so was the expectation that you would blog. On Substack, the primacy of monetization creates a treadmill effect. The expectation that you might get paid creates the expectation that you should provide content to people. This creates two separate problems.

The first problem is that our expectation of writing requires that we do so, consistently, not in a punctuated manner. This doesn’t seem like a problem until you consider that when you’re a specialist, you might not actually have that much to say in a consistent manner. Even for a generalist, the expectation of a daily or weekly newsletter means that slow weeks and fast weeks are gone, your rate must always be the same. Writers in this case have a few options: they can speed up their writing (at the cost of quality), they can shorten their posts (at the cost of saying things they’d find meaningful), or they can switch from specialization to generalization (to the loss of their expertise). This last option, where writers stop talking about arcane, specialized subjects and start talking about general topics or events in the news worries me the most. You are seeing it a lot already - people talking about things they know nothing about and people believing them because of prestige bias (people biasing towards others because they’re prestigious, with this bias extending towards areas which are unrelated to their success). There is already a problem on Twitter and other websites of every specialist believing they’re a specialist in everything and every generalist believing the same. Without sufficient clustering online, we lose the ability to actually specialize.

The second problem is a bit different. In 2018 at the beginning of my blog, a frequent reader of mine sent me several hundred dollars through PayPal as a donation in appreciation for my writing. I was obviously humbled by this generous donation, and one thing I recall him saying was that I should keep in mind that the donation should not mean in any way that there’s an “expectation to write anything special.” Thinking on this donation from time to time has made me consider more seriously the different forms of patronage that others provide. This donation, one lump sum, was a bit different from Patreon, which is what was normal at the time. In this case, the instantaneous one-time donation meant that even if my donor wanted any kind of special content, there was no guarantee he would ever get it. Compare this to the Subscription model, and the implications are different: the expectation I will see you again next month carries with it the weight that I continue to write next month (and not only that I continue to write next month, but I continue to write about things you want me to write about). Substack is technically not moderated, but this is certainly the moderation of others’ work by direct payments. This might not be an issue for blogs which remain small. But for anyone who wishes to grow, growth means generalization, and this comes at a literal cost.

A Problem in Collective Computation

At the end of the day, nearly every downside I have presented here can be seen as an upside. People who write on more arcane subjects will speak to more and more people who will, maybe, understand these more arcane subjects a little bit more. In return, Substack makes it easy for these people to get paid for their knowledge. More people are going to make money off of Substack than they ever made off of blogs. I won’t disagree that it’s a good thing for them and for their readers. But what it could be bad for is the information ecosystem. I worry about what the ideal information ecosystem looks like and current rates of information loss.

Nonetheless, I’m here now. Why? This is what blogging is now. In 1989, W. Brian Arthur introduced the concept of technological “lock-in” for competing technologies. His theory was that as a larger proportion of people begin to adopt a specific technology, fewer users are likely to switch to others (that is to say, certain technologies outcompete others simply because everyone else is using it - this holds true even for technologies which are inefficient, such as the QWERTY keyboard I am using right now). I suspect while few people would read most of my other blogs if I were to update them, many of those same people who wouldn’t check out my larger Wordpress blog would be happy with receiving an email from this one.

Of course, blogging wasn’t perfect, but decentralization meant that things were small, loose, informal, disconnected, disjointed, impenetrable, not meant for me, not meant for you, and not meant for normies. The disappearance of the specialist forum full of online snobs is a disaster for the exploration of novel information (to this day, my most obscure Google searches find their answers in pre-2015 forums). These forums were replaced by Reddit, which itself is itslf moderated, centralized, and algorithmized. The same thing has happen now with blogs, as we’ve found our settling point for where our long form ideas are supposed to be.