An algorithm for bullshit is in fact not terribly needed in the presence of Big Tech, because Big Tech’s founders and top management generate it with such natural facility that they could outpace even the best generative AI tasked with that objective.
Substack’s CEO Christ Best let us know this week that he believes the problem with “legacy social media”, e.g., the major commercial platforms that were established after 2005, is algorithmic design that caters to the “basest version of ourselves”, the readers and creators of social media. Substack’s algorithm, he says, serves “the best version of you”.
Why can’t I serve the best version of me on my own? Why am I presumed right now to be lacking the algorithmic prosthetic supplied by Substack, to not be reading with agency?
The answer, says Best, is that “culture is also about learning why you want what you want, refining your tastes, challenging your convictions, and promoting humanity and agency over the process.” That I cannot learn this by myself, of myself, in myself. I need Substack and its proprietary, embedded, non-described algorithm to do it for me.
Let’s think back for a minute on print media in the 20th Century, where the kind of writing that Substack hosts would have been found. Print media had a different kind of necessary overhead, so no publisher could have offered to publish hundreds of thousands of newsletters curated entirely by their authors, and no one could have afforded buying them if they were published. (Nor would anyone have had the space to receive them and store them.) So print media had to have editors hired by the publishers to make judicious decisions about what to publish.
Those editors in turn generally targeted a particular kind of imagined reader. As the 20th Century wore on, they often collected more and more data via market research about those readers, in part to help solicit advertisements. But they also cultivated a readership by engaging in a kind of dialogue with them about what the reader believed about themselves and their interests. The editors often disclosed some aspect of their view about what represented appropriate material for the audience, and the audience often communicated with the editors about their reactions to what was published.
In many cases, the editors believed they were bringing information and perspective to the readers that the readers did not already possess. But they also constrained what they brought to the readers based on the identity that the publication had developed. You didn’t expect to pick up Armor and read articles about the latest in women’s fashion. You didn’t expect to pick up TV Guide and receive a run-down of the latest trends in British poetry. You didn’t buy the New Yorker and wonder where the articles about the bat-headed boy were, or the National Enquirer only to be shocked by 4,000 pages of the Pentagon Papers printed verbatim.
Nothing about the identity of the publication or the general selective principles guiding editors was a secret. The only thing readers might wonder is “why did they choose this essay? commission that article? publish this author so much?”. Readers also did howl bloody murder from time to time when they didn’t care for what they’d read, and sometimes that led to backtracking or acknowledgement of error. If it was that kind of publication, that is.
In that world, for all its manifest flaws, I would have understood what an editor meant if they said they wanted to publish at a certain level of quality, and I would have understood what they intended if they said they wanted to expose their readers to material that the readers would find enlightening but also new, interesting, uplifting, transforming, surprising.
When Best says that he thinks Substack’s algorithms are necessary to refine its readers’ tastes, challenge their convictions and promote their agency, on the other hand, what does that even mean? There’s the hilarity right at the gate of touting your platform as promoting the agency of readers via an algorithm whose weighting and functioning is concealed. My choices get more developed via an unaccountable, invisible, unadjustable infrastructure that makes choices for me?
If William Shawn was making choices about what New Yorker readers should read, well, Shawn was a knowable figure within the circles than many of its readers traversed. He was quiet and quirky but he was a human being. A reader—or a writer—could envision how and why he made decisions. And a reader—or a writer—could decide if Shawn’s decisions were appreciated or not.
Substack’s owners want both to be involved in discernment, to make better culture, and to not bear any human responsibility for what is discerned, or to own the propositions and ideas behind what they will uplift and what they will marginalize. The most you’re going to get from Best is that they want to “challenge their convictions”, which I take to mean “our algorithm is going to show you stuff you hate politically and culturally, because that’s good for you” without having to actually say that’s what they mean. And I don’t buy that they think that’s a value; I think they think that’s what creates more traffic and more subscriptions.
I will tell you what it is also: barely different from Twitter, Facebook and the other “legacy platforms” deciding to shove some content down my throat because it mines my attention or creates a datatrail that can be resold. It’s not much different from those platforms constantly urging me to use their algorithmic infrastructure to flog my own content in various ways. The business model’s a bit different here, prioritizing engagement in how it leads to subscriptions, but it amount to the same kind of pushing the reader into an autonomous car, locking the door and taking away the map while telling them that wherever the car goes, it’s for their own good.
I know all the ways Substack’s algorithm might work and none of them justify the claim that the algorithm knows better than I do what I need, or what I ought to read. The days where anyone believes recommender engines actually work like that are way way back when Netflix held a contest and everybody actually thought they’d build something real and better from it.
The kind of algorithm that really would do what Best claims to want would be one that was 100% transparent to readers and 100% under their individual customizable control. And it would be 0% infused with a claim by a platform CEO that he has a way to know what I value without having to engage me human-to-human.
Speaking of quality and humanity and how they might intersect with making money, I’ve been taken with this article from Philadelphia Magazine, “The Rise and Fall of Korshak Bagels”.
I’ve never had a Korshak bagel, though I’ve planned to on at least four occasions. The reason I haven’t is that when I’ve been down in that neighborhood, around East Passyunk (full of good foodie places), the lines outside of Korshak have been unendurably long. I love a great bagel but I don’t love one that much. If I’m down in that neighborhood or in the Italian Market, hey, I got lamb tacos and Stargazy pies waiting for me, so off I go.
But what really grabs me in the article is on one hand Korshak’s admirable insistence on doing right by his workers and on doing right by his own love for making bagels by hand in the way he deems satisfying and real. That’s how long-standing artisanal businesses function when they do: they value labor, they keep things at a small and sustainable human scale, they rise from the passions of a creator-owner, they don’t chase trends or do market surveys or have an MBA-approved business plan.
But when they stay alive, they don’t do one thing Korshak apparently refused to do, which is to charge what that’s worth. His bagels were expensive, mind you. (I looked that up each time I walked past and saw the line.) But if you’re that into “pay people well” and “make by hand” and folks are then in turn that into “I want to eat what you’re making”, well, charge more. And then charge more again. Stop when people stop waiting in line. $6 a bagel? $8 a bagel? A $18 bagel sandwich? Hey, if there’s folks who will pay, then that’s how you pay those workers a good wage and how you stay sanely in business. A great artisanal bagel is not an act of equitable justice in its own right.
Mind you, I think great food is justice. But I think there is and more importantly can be great food that is also affordable. Much as I think there can be safe and comfortable shelter for all, affordable quality education for everybody, and so on. But some of that, at scale, does take finding efficiencies in the production process, especially barring a complete socioeconomic transformation that produces economic justice in all of what we make and need and consume. Even without that, there can be taco joints and fried chicken shacks and roadside barbecue and great bread that just about everybody with two dimes to rub together can afford.
For that reason, I don’t think a guy with Korshak’s aesthetic and economic vision needs to feel any guilt at all for selling bagels at caviar prices, if there are people who will pay those prices. But also, plainly, I am not Phillip Korshak. So I also kind of appreciate it when any person goes full-on Don Quixote against their windmill. However much you might do a Sancho Panza schtick off to the side, you’re never really going to get through to them, even if they’re never going to win either. And everybody needs a windmill now and again, whether it’s a bagel or a Best.
Rather a tremendous amount of effort is being poured into an effort to keep the Big Tech growth machine going, and yet the returns seem to keep diminishing. What is the last big tech innovation to really disrupt how we have lived? It seems like a tie between the iPhone and Facebook/Twitter--although I am old, and so perhaps the answer is TikTok or the streaming services that were familiar already by ca. 2010/2011. But everything since then has been in service of refinement, not disruption--AWS is a giant business but it is not really a disruptive one in terms of the daily lived experience (as opposed to the infrastructure of the world). I think that SV folks are chasing the dragon to try to find that next giant hit, rather than focusing on enabling and having sustainable profits, and in doing so they'll wreck a lot of their own options.