So Long, Farewell, Auf Wiedersehen, Good Night
In which I decamp from Substack to Buttondown, and announce a new format for the newsletter.
This will probably be my last newsletter on Substack.
Subscribers will not notice much of a difference as far as deliveries are concerned, but the Web archive will pick up—where I had long ago planned to move this newsletter anyway—on Buttondown.
Buttondown is also where I host my serialized book project, The Nature of Software, which I can get back to now that I have gotten this monstrous schwack of other writing out of my system.
This decision is tangentially related, as some of you might expect, to the recent defense by Substack brass of their decision to continue to host Nazi newsletters, and, by extension, broker payments to the people who write them. While I had been considering emigrating for a while, this move on their part has certainly made the decision easier.
As I wrote in that very first newsletter (which is also available on Buttondown), I am lukewarm at best about newsletters as a format. I find them sharply limiting as an author, and as a reader I find them at best to be a characteristically bulky notification mechanism that tells me it’s time to go check somebody’s website. When blogging was new, I found the strictures on the format to be limiting too. Newsletters are just blogs—an already highly-constrained conceptualization of the Web—that have been further constrained, so they can squeeze through the narrow conduit that is e-mail. As I remarked a little over three years ago, this appears to be the way we’re communicating now, and due to the facts on the ground, to send newsletters—unlike blogs—you need and intermediary.
Since I intend, reluctantly, to continue communicating by newsletter, and to do so I require an intermediary, I would prefer one that I can trust. In the three-plus years I’ve used them, Substack has engaged in all sorts of slimy, growth-hacky behaviour. There’s stuff I can ignore, like fact that they spent millions seeding their stable with a hand-picked coterie of authors and pundits, or how they fudge their bestseller badges by orders of magnitude, or the mushrooming accoutrements of Yet Another Social Network. There’s stuff that I have definitely benefited from, like their referral infrastructure, which was the main reason why I stuck around. But most acutely, every time I send a newsletter through Substack, I have to spend about three hours reformatting it, because of the deliberate decision on their part that the only sensible way to publish a newsletter, in the year of our lord 2024, is to type it into your Web browser by hand. The only way I can understand this is that they want to control the authoring process, up to and including retaining the master copy of your writing. I resist this, and I pay for it—to the tune of three otherwise-billable hours per newsletter.
Substack has clearly spent a lot of money on their text editor, which does accept certain Markdown-inspired keyboard sequences, but in response to numerous customer requests to just process Markdown (a sort of folk notation that enables you to encode text embellishments without your fingers leaving the keyboard), their response is “nah, we’re cool”. For me, this means taking another day to release, because painstaking manual reformatting is an extra task on top of writing the newsletter itself. For Buttondown, I just pipe my master document through
pandocto turn it into Markdown and paste it into the browser in one shot. I could also automate this process completely, since Buttondown has an API, while Substack, by design, does not.Another behaviour of Substack is that it nags you to plug sign-up forms into your outgoing newsletter. Here it either requests that you place two by hand or otherwise offers to do it for you. Only in the former option do you putatively choose what it says. I have found, however—not sure if this is a bug or a design decision—that anything I write in the call to action gets overridden by their awful saccharine boilerplate that is an arrangement of words I would never use. This, I suppose, is the author-side complement of the other tacky, growth-hacky behaviour of throwing up a nag screen to the reader, every time you navigate off the page and back again.
So, after three years as a Substack user, I have a mounting list of complaints, all of which can be traced to deliberate strategy on the part of their executive team. Now let’s talk about Nazis.
Every company based in the US that resembles an “interactive computer service” has an incentive to tell the story that they are as close as possible to what in the business is called a “dumb pipe”. This is because of a law that says that they are immune from liability for harms caused by whatever third parties put through said pipe. You may have heard of it, it’s called Section 230. Now, the fear on the part of these companies is that if they intervene too much editorially over the content that flows through the pipe, it will be considered their speech rather than third-party, and thus they will be liable.
That’s my understanding, at least. I am neither a lawyer nor an American, I have just been osmosing the offgas from this debate for the last several years.
Intermixed with this entirely rational apprehension, however, are sophomoric libertarian pronouncements about free speech or whatever. Here is where Substack has proven itself manifestly disingenuous. Their chief communicator asserted no less that the best way to deal with Nazis is to debate them in an open forum (and far be it from them to get in between that), as if those ideas hadn’t already been repudiated, in blood, on the literal battlefield. Substack, of course, as Dave Karpf and others note, would take its 10% cut of the subscription revenue from both sides.
What makes this disingenuous, as many have pointed out over the longitude of Substack’s operation, is that there absolutely is material that they will not tolerate, and predictably—as with just about every other social media platform—it orbits around the nucleus of porn and other sex-related content. A reasonable conclusion to draw from these facts, therefore, is that Substack is not concerned about hosting—and funding—Nazis. Karpf (again) hypothesizes that this position is merely amygdala-driven, to the extent that a number of their major revenue-generators—including certain authors they pay directly—say casually racist shit all the time. The thinking goes that if they got rid of actual Nazis, there would be pressure campaigns to get rid of them too, and this, unlike Stanley Stormer’s Führer Körner, would actually ding their bottom line. But then, the other possibility is they actually like Nazis?
The thing is, I don’t want to have to investigate. I remember back in the pandemic before-times, somebody I knew was convinced Jack Dorsey was a white supremacist✱, because of Twitter’s infamously lopsided content moderation enforcement tendencies. This question teetered on the edge of plausible deniability, at least for me, because it could also be explained by indolence or cupidity. Content moderation, after all, is always a money-losing prospect in the first order, so it’s entirely possible that one’s half-assed infrastructure catches people from marginalized communities expressing their frustrations about being, you know, marginalized, while skipping the CHUD contingent for using coded threats and slurs, scrupulously sculpted over many iterations to skirt enforcement. Applying Hanlon’s Razor would conclude that it just wasn’t enough of a burr in Dorsey’s backside to ever do anything about it. Whether there was any more to it would mean a level of muckraking I’m simply not set up to do. Making a positive determination that the people helming these companies truly have fascism in their hearts is an expensive proposition, because fascism is something fascists tend to conceal—until there is a high enough concentration of them, that is. Indeed, plausible deniability is what affords fascism to hide in plain sight.
✱ I should note that this person’s stern conviction was nevertheless not enough to get them off Twitter—but Elon Musk sure was.
How you get to counterparties you can trust, without the burden of thoroughly investigating the otherwise plausibly deniable, is to look for the ones who are willing to crisply articulate their principles, like Justin did over at Buttondown. In so doing he advanced a benchmark by which he was willing to be held publicly accountable. Importantly, he didn’t guarantee that his platform would be always and forever spotlessly Nazi-free, but rather the implication is that if he becomes aware of any sign of Nazi content, he will act swiftly to dispatch it. The important thing is he has made a pledge and we can hold him to it, while Substack—who, since I started writing this document has made the perfunctory gesture of deleting five Nazi newsletters (out of a sample of dozens handed to them)—continues to weasel. I don’t want to have to spend the effort determining if I’m materially supporting a radioactive ideology—and enticing others to support it—on top of the effort it already takes to participate. As such, this is an easy decision for me.
On the positive side, I’m thinking of tinkering with the format of my main (free) newsletter in a way that would be easy to turn out on Buttondown but less so on Substack. What I want is to send much shorter emails that serve as a jumping-off point to longer, or at least less-hamstrung articles on the Web. This way I can rig up an internal queue that gets flushed, say, once every week or two, in the form of a newsletter, that serves as a contemporaneous snapshot of whatever I’m thinking about or working on.
What I’ve been up to all December
I expected I would ship Intertwingler before the end of 2023, but I got diverted by a request to do something that ultimately turned out to be just as important. In the spirit of the Programmer’s Credo—“we do these things not because they are easy, but because we thought they would be easy”—it amounted to an overhaul of an earlier project (and precursor to Intertwingler), and came out the other side as a chunk of code that can layer on top of Intertwingler—once I finally ship it—and become its first app.
The current status of Intertwingler is that I haven’t touched it since the beginning of December. What remains to be written is its transformation function code, which is part of the “spine”, and the last critical piece. By providing an interface to programmatically derive content from other content (changing file formats, resizing images, caching, compression, and much more), the transformation function infrastructure results in Web-based systems that are more compatible and more efficient, and saves developers a whackload of time making their systems more compatible and more efficient.
So now, after a month of overhauling the IBIS tool—two weeks of which done in dribs and drabs while visiting family over the holidays—and my first cold (still never had covid!) in years, and this impressive 7000-word bolus of backlogged newsletter writing (which I mercifully chopped into three pieces), I can hopefully get back to it.
Why We Care About Plagiarism
Because I was busy all December, I missed all the Ivy League drama, so I asked the internet “what did Claudine Gay actually do?” That is, what precisely was the extent of the transgression that cost her the presidency of Harvard? These are my remarks on the matter.
I Read a Novel For Once
Because I read for a living, and then plough the bulk of my extracurricular reading time into learning things, I am terrible at consuming fiction. If, however, you shove a novel in my face and say this is good and you have to read it, I will (eventually) read it. In this case, I shoved Joanne McNeil’s début novel Wrong Way into my own face, because I had read and loved her (non-fiction) book Lurking.
The book is about a 48-year-old woman named Teresa who gets a below-the-API job at a sprawling tech company. The book is meditative, contemplative. It absolutely nails the change-the-world demeanour of the marketing rhetoric and the saviour-CEO. It also has a lot of subtle things to say about class, work, and aging. Pick it up.