Biologists use the term endling to refer to the last remaining member of a particular species. As of May 2019, there is only one remaining Blockbuster video store left, located in Bend, Oregon. Digital streaming services have become so ubiquitous and well stocked since Netflix’s streaming debut in 2007 that the brick-and-mortar business model for video stores now seems ludicrous. Whereas independent bookstores and vinyl record shops have witnessed a remarkable resurgence in the last five years, video stores are functionally extinct; even the lone Blockbuster lacks a breeding partner with which to carry on the species.
Journalists offer many reasons for the resurrection of certain “dead” media formats in small, independently managed stores. Like me, many of them favor the human element of selection. The staff in these stores thins the herd so that the customer doesn’t have to, only stocking items that they believe to be worthy of the discerning customer’s time. When too much access threatens to be just as disheartening as too little, the curation of a perfect collection becomes a sincere form of critical practice.
Thinking about the last vestiges of video stores makes me feel particularly nostalgic because my future was sealed while I was in one. It was February 2008; I took a call on my clamshell dumb phone and received the news that I had been accepted into a doctoral program in English literature and cultural studies. The voice at the other end of the line told me that an accredited institution of higher learning wanted to offer me five to seven years of income to teach freshmen to write and to spend time thinking about the ideas that mattered most to me. It was a big day, and I’ve never forgotten that I received the call while browsing in an independent video store in the back of a coffee shop in Pittsburgh, Pennsylvania. The store was called Dreaming Ant, a name whose meaning is still obscure to me.
One of my favorite things about Dreaming Ant was how chaotically its films were organized, like a great used-book store. The owner cared more about the size of his collection than its display, which means that it was both quite difficult and all too easy to find great films. He arranged his collection more like a museum exhibition than like a Starbucks snack-food display. There were DVD boxes nestled in every available nook and cranny on the wire shelving and stashed away in countless thick white vinyl binders.
It was clear that the owner believed that someone, a real customer, would care to watch each and every one of these flicks. What’s more, he didn’t even go to the trouble of renting his own storefront. Instead, his personal collection occupied the back of a bustling coffee shop. The place had a great vibe. The collection seemed like it was chosen by a geek for geeks: bizarre, obtuse, foreign, nonsensical, abstract—all the things an aspiring pedant like me could want. And every time I found the right film, it felt like I had accomplished something, like I had discovered something real.
I’ve tried to fill the gap the closing of that store left in my life with many digital streaming platforms. Ironically, the app that filled it best never really worked. The experience of using the streaming service FilmStruck felt more like finding a movie in a video store than like finding an option on any competing app; that was simultaneously its best and worst feature. The service offered more fine films than I knew what to do with, but I had to shoulder the responsibility of choosing which to watch. The quality of the content mattered more than a breezy user experience. Picking a film, for me, was about finding something worth watching, not just finding something to watch.
FilmStruck’s interface felt like the digital equivalent of being sneered at by a snooty video store clerk—if you didn’t find the right film, that was on you. Like its brick-and-mortar predecessors, the app eventually went forever dark, on November 29, 2018.
Netflix, Amazon, Hulu, and Spotify now actively work against the kind of discovery process promoted by video stores and FilmStruck. Every time we rank a movie, review a purchase, place an item in our virtual shopping cart, fill out a questionnaire to see which Game of Thrones character we are, or even just click a link, our algorithmic self is computed, processed, and delivered back to us. Sure, algorithmic selection has also pushed me to discover great new films, television series, books, and songs, but I can’t help feeling that it was the black box of the algorithm—not me or even people like me (critics, historians, snobs, and buffs)—that made the selection. This is all a way of saying that these apps work so well that it’s hard to remember what seeking out culture in the wild was like.
What troubles me is that the algorithmic selection of content based on our established preferences and behaviors also effectively cordons off content that may challenge our tastes, excluding stories and characters that might change the ways we see the world. Literary critic Viktor Shklovsky crafted his theory of literary art around the concept of defamiliarization. This theory applies just as well, if not better, to film, despite the fact that Shklovsky was writing in 1917, when the medium was still young.
In his essay “Art as Technique,” Shklovsky writes: “Habitualization devours works, clothes, furniture, one’s wife, and the fear of war. … The technique of art is to make objects ‘unfamiliar,’ to make forms difficult, to increase the difficulty and length of perception because the process of perception is an aesthetic end in itself.” In light of this, I want to reconsider the aesthetic ends of discovery and selection. Instead of nostalgia for dead media and times gone by, we need a fuller sense of the experiential components of art.
Permit me an oversimplification for the sake of argument. The critical difference between a film and a movie is all in the manner of watching. Movies are supposed to be consumed as straightforward entertainment; films, on the other hand, are aesthetic objects that invite rewinding and repeating, so that the viewer can fully investigate their meaning, or lack thereof. Movies are often watched again and again, but usually in one go, as a whole. Even the casual study of film requires more rewinding than do commercial movies. Although neither disc-based film delivery nor streaming video files involves physical tape or filmstrip that must be rewound to move backward in the narrative, the term rewind persists, and is even used by viewers who have never held a VHS cassette. Nostalgia works in funny ways.
By breaking the old business model, Netflix has also changed the way we watch. Maybe FilmStruck’s flaws appear so grievous now because film watching is so much more causal, so much easier: the idea of “Blockbuster and chill” feels laughable, if not totally inconceivable. Another way of thinking about this comes from considering how apps like Netflix have ascribed a new importance to genres.
Scroll through your recommendations on Netflix and you’ll see hyper-specific “genres” created around your tastes; at the time of writing, my home screen included “Critically Acclaimed Movies Set in New York City” and “Cyber Paranoia.” No creator or critic has ever codified these genres, but there they are on your screen for you to swipe through. Netflix knows me very well, but FilmStruck didn’t even bother to treat me as an individual, and thus never had a chance of wowing me with algorithmic legerdemain. FilmStruck protected the identity of its collection rather than uncovering the identity of its user.
In fact, FilmStruck’s design systematically stood in the way of its user: leave your film paused too long and the app forgot your place, the search function did not auto-complete based on previous searches and was quite picky. If you were a poor speller or typist, or if you didn’t know exactly what you were looking for, you’d fail to find films and thus erroneously assume that you didn’t have access to them. For instance, if you searched for “Wells,” you got no results; search for “Welles,” on the other hand, and you got nine films starring or directed by Orson Welles. Netflix has no trouble finding Orson Welles’s The Stranger (1946) with either search term and it even offers alternatives. Netflix delivers a search result no matter what you type into the field.
Some of FilmStruck’s other design decisions were so odd that they must have been deliberate. For one, there was no way to review films. You could add movies to a watch list, but there was no way to offer feedback of the numeric, star- or thumb-based, let alone written, variety. The app’s “recommended for you” feature was downright austere compared to the cascading lists of recommendations provided by all the other streaming services. Netflix even goes so far as to automatically play the trailer for a new item it predicts you will like. This recalls an embarrassing question that Netflix has no doubt asked many of us mid-binge: are you still watching? Yes, we are.
The algorithmic selection of content based on our established preferences and behaviors effectively cordons off content that may challenge our tastes.
A generous reading of FilmStruck’s defects might point to the auteur theory of cinema, which privileges the director’s vision and, consequently, visionary directors. If you pulled up Seven Samurai (1954) in FilmStruck, you would see 10 “related films” directed by Akira Kurosawa—some samurai pictures, some thrillers—but none of the two dozen Zatoichi samurai flicks held in the service’s library. FilmStruck’s interface design seemed to assume that the user cares only about seeing films from the master Kurosawa, a director any film buff should already know about, rather than accept the possibility that he was just in the mood for rad sword fights.
Moreover, since there was no way to review or rank the films you had watched, recommendations were not personalized in any way; in fact, my version of the app didn’t even correctly register the films I had completed. Thus, a film I watched for 10 minutes impacted the app’s recommendation system just as much as one I had watched, rewound, and analyzed repeatedly. It was on me to remember which films I had seen and liked—something that’s pretty hard to do these days given the sheer volume of content we consume, and that is often outsourced to the app itself.
The real cause of FilmStruck’s poor design—and eventual discontinuation—was financial. Turner and WarnerMedia (subsidiaries of AT&T) saw the potential audience for films as too limited, too niche, to invest in app design. They would rather enrich their rights library. WarnerMedia and rival media conglomerates, like Disney, know that they will require vast catalogs of blockbuster content to have a chance of dethroning Netflix.
According to my definition, film needs to be challenging. We shouldn’t forget that the history of film delivery has always been rarefied and secluded. Because of the niche audience, foreign and independent films were cloistered in special events (festivals) and special places (indie stores and college libraries). Big-tent content organization does not fit within the history of art film. Netflix’s library seems endless, whereas FilmStruck tried to focus—it is thus not surprising that the latter couldn’t flourish in this new media ecology.
FilmStruck’s failure makes sense. Compare the FilmStruck app’s omissions to the argument that Netflix’s business model is as much about surveilling and predicting user behavior as it is about delivering content. In What Algorithms Want: Imagination in the Age of Computing (2017), media scholar Ed Finn writes that “Netflix can track precisely how their customers watch particular shows, how long they hesitate between options, and perhaps even how much pausing, fast-forwarding, or rewinding goes on.” With this analysis, Finn tries to explain how Netflix works its magic. And when magic is the expected state, how can a competing app deliver a remarkable new experience? Maybe the only way to deliver a novel experience was to design a user-unfriendly app.
FilmStruck’s interface relied solely on written blurbs to classify films, so that even if you managed to find the director you were looking for, your search results were sorted by release date, rather than according to the critical or popular reception of the film. Compare this to the way you regularly use Netflix or Amazon Prime and you’ll see that FilmStruck lacked functions that many now expect—features that impact what you actually watch more than professional reviews, customer reviews, or even friends’ recommendations.
Netflix tracks and calculates everything we do with and around the app—when we watch; how long we watch, down to the second; even where we watch—so that their interface can gain knowledge and authority, as Finn argues, by presenting themselves as “computational curators of human content [that] actually function as much more complex collaborative systems coordinating millions of users, corporate or collective objectives, and elaborately abstracted structures of knowledge.”
Put another way, FilmStruck was stuck in Web 1.0 thinking when compared with Netflix’s pathbreaking metrics and surveillance model. Nevertheless, selecting and discovering content through FilmStruck evoked the education in film I got from picking through the Dreaming Ant catalog. For both experiences, I was forced to spend supplementary time researching films and directors I would then seek out. I read academic essays, Wikipedia entries, subreddits, and best-of lists, all compiled by human reviewers. FilmStruck and Dreaming Ant both made me do my homework. Netflix feels slicks and psychic while FilmStruck felt messy and conceited—all too human, really.
FilmStruck protected the identity of its collection rather than uncovering the identity of its user.
FilmStruck’s drawbacks echo the financial failures of my beloved Dreaming Ant. Both privileged art appreciation above commercial viability, and inefficient sifting through strangely ordered content over a streamlined, calculated experience. I, for one, counted this as an asset, or at least a possible asset. The human–like flaws of FilmStruck seemed to work for this app, because it was intended to host difficult films.
I am not wistful for the days of being sneered at by a video clerk, but I think that denigrating the purpose they served costs us something. FilmStruck, Netflix, and perhaps even the newly debuted Criterion Channel have spirited away people like this clerk and replaced them with persuasive technologies that try to actively shape our decisions about the culture we consume. Content algorithms work their alchemy by making each of us the arbiter of our own good taste, and we lose much from this narcissistic self-fulfillment. FilmStruck’s difficult interface asked the viewer to make his or her own decision about a smaller collection of highly esteemed films, rather than offering the omniscient invisible hand and the seemingly infinite supplies of Netflix et al. Difficult things are not inherently good, but easy things are not good enough.
While researching for this essay, I discovered that Dreaming Ant’s website is still online, despite the fact that the store closed in 2013. The experience was strangely affecting. Have you ever gone back to a website that you used to frequent to find it exactly the same, although you have changed profoundly? I moved out of Pittsburgh in 2014, but I remember now that I used to frequent this site to avoid writing my dissertation. I would daydream about which film to rent for the weekend, and how I might impress my new girlfriend, also a doctoral student in English.
Usually, I would seek out a film mentioned in an essay by a colleague, or by a professor. More often than not, I found it listed in Dreaming Ant’s digital database. I would then schlep over to the store to pick it up. Sometimes the disc would be checked out by the time I got there. Sometimes discs went missing, never to be found again. Sometimes its case would be misplaced, and I would need to ask the clerk for help. Sometimes he would give me attitude about my choice. Sometimes he would praise it.
I admit that I dozed before the end of more than a few of these movies. I can’t count how many I didn’t even manage to begin. I probably still owe some late fees. But I would still gladly pay twice as much as I did for FilmStruck to have that store, or one like it, within walking distance of my home.
Returning to Dreaming Ant’s website reminded me of something that I had willfully forgotten: I, too, have found a way to abstract away the video clerk, to supplant him with the creation of my own refined algorithmic self. Returning to this website, I was struck with nostalgia for a previous life. When I clicked on “End of Days,” the entry the store had posted on December 4, 2013, describing its impending closure, I was trying to go back in time. After the page had loaded, it displayed only the message: “Error establishing a database connection.” I guess that the past is past.
This article was commissioned by Ben Platt.