What used to be known, literally, as “the cutting room floor,” now exists as a digital bin, an assortment of deleted scenes, unused (and in today’s mode of industrial documentary production, perhaps even unviewed) footage — material that, through its absence, haunts any finished audiovisual work. Often when this material is revealed, on a Blu-ray supplemental features disc, for example, the director’s elisions affirm the strength of their initial creative editorial decisions. Other times, particularly in biographical documentaries, the unused material becomes a kind of lacuna, suggesting not only paths unexplored but a failure to engage with all that’s messy, contradictory or simply overflowing within a richly lived life.

When Gary Hustwit (Helvetica, Objectified) began making a documentary on British composer, musician, artist and activist Brian Eno, whose work spans the glam rock of early Roxy Music, ferociously inventive singer/songwriter solo albums, ambient music, and behind-the-console producing (Talking Heads, U2, Devo), the creative polymath expressed his distaste of the genre. “I didn’t fancy the linearity of conventional biographies,” Eno told Alissa Wilkinson of the New York Times. “Lives don’t run in straight lines, and every time we think about them in retrospect (i.e., every time we start remembering) we actually rethink them. Our lives are stories we write and rewrite. There is no single reliable narrative of a life.” 

So then, to convince the reluctant subject, Hustwit devised a filmmaking approach that would drawn from another of Eno’s pioneering pursuits, that of “generative music.” From the tape loop experiments of works like Discreet Music and Music for Airports to later music-making apps like Bloom, Eno, inspired by the work of composers such as John Cage, Cornelius Cardew and Karlheinz Stockhausen, introduced indeterminacy, chance operations and the variable sonic outputs of different technological processes to create what he called “generative music.” And with Hustwit’s new documentary, ENO, just as the Bloom software produces an ever changing composition, Brain One (an anagram of Brian Eno), the software Hustwit created with British artist Brendan Dawes, produces an ever-changing — with each output — film. In addition to cinematically mirroring one of Eno’s most storied compositional techniques, the doc answers, conceptually, at least, that thorny question about the archive. Only a small percentage of the more than 500 hours Hustwit accumulated could be in any version of a Brian Eno doc; with ENO, opening today at New York’s Film Forum following a Sundance premiere, that decision of what to include, on a version-by-version basis, is finally done by computer. For the viewer, what’s on-screen is as important as the knowledge that whatever’s missing still exists and could appear in a future viewing, or be taken in by another viewer halfway across the world.

Or, to say all of this more simply, the intelligent and quite charming ENO rousingly answers a question Hustwit said he asked himself — “Can you watch a documentary about Brian Eno that doesn’t mention Roxy Music?” — in the affirmative. (Although the band, along with a fantastic clip of lead singer Bryan Ferry presenting a “Best Producer” MTV Music Video Award to Eno, were both in the cut I saw at a recent Film Forum press screening.)

The overall anxiety around generative AI’s escalating ability to create synthetic images and to supplant human labor has colored some’s reaction to ENO, even though the film is warm, witty and with a human pulse throughout. In fact, ENO is less a product of this NVDIA-driven historical moment than a continuation of experiments that directors have been embarking on for decades. “My own work is in the pull of a pair of forces that defined the late twentieth century—the Cinema and Cybernetics, the Projector and the Computer,” wrote the video artist Grahame Weinbren in 1995. In works like The Earl King, a collaboration with Roberta Friedman, Weinbren presciently addressed “questions that arise… about how cinema changes when its apparatus is linked to a computer—just as one can investigate changes in the structure of cinematic communication when recorded sound is added to the moving image.” 

A couple of weeks before today’s Film Forum opening, Hustwit and I connected over Zoom to discuss the archive, Brain One, ENO, Eno, and Weinbren’s question as it applies to a film about one of the most philosophically-inclined artists of the 20th and 21st centuries.

Filmmaker: I’d like to start by diving a little bit deeper into the process of making a unique version of ENO. As I understand from other interviews, to generate a DCP for a screening, such as the one I recently attended at Film Forum, you fire up the program, key in a command and the computer spits out a unique version of the film, or at least as unique a version that can be given the existing variables. Where, then, if anywhere, does the human factor enter? Do you sometimes watch the DCP back and wish that the version was different, or that a particular scene was included, and make changes? And for the Film Forum run, I presume there are multiple DCPs —

Hustwit: Yeah, one for every day. And, yes, I can click “generate” and a version of the film comes out of the engine with its custom 5.1 surround audio mix pretty much ready to go onto a screen. But it took thousands of human hours and filmmaking creativity to get to that point. The reason it works as a documentary and a cinematic story is because we’ve programmed these algorithms with our sort of “filmmaker intelligence,” not artificial intelligence, and because we have amazing editors who know how to craft scenes but also how to craft materials that the engine can then retextualize, or remix, for each version. So, there’s a lot happening under the hood of this project. We wanted some of that be part of the visual aesthetic of the film, so when you see [on screen] all the file names, that’s in real time the engine looking through the data set at all the things it could play next. Some of those file names you’re seeing you might see later in that version of the film, but most of [that footage] you’re not going to see, so it’s sort of a way to let the process also be part of the viewing experience too. There are versions of the film that come out that I’m like, “That’s not my favorite scene,” but that’s part of the point.

Filmmaker: So, do you surrender to the machine at that point?

Hustwit: Definitely. I have to step back and look at it from a bigger perspective, which is that I’m not creating each individual version of the film, I’m creating this bigger structure in which it can create itself. For every output there are scenes that I love and that I root for and hope will show up. But if they don’t, that’s alright. You gotta be okay with that, and my perfect version of the film is different than yours or somebody else’s. It takes a bit of the subjectivity, or a little bit of the authorship, [out of] each version, but, of course, I curated the whole set of ingredients that go into this thing. And if the ingredients are good, you can recombine them in a lot of different ways and still have it turn out well.

Filmmaker:  What are some of structural elements that carry over from version to version that you presumably spent a lot of time with your editors devising? I understand the beginning and the ending of the film are more or less the same across all the versions.

Hustwit: Yes, the beginning of Brian in his studio in the backyard of his country house and then the ending scene of him in the rose garden talking about why we need art —those are always in this generation. And then there are a handful of scenes that come up either all the time or more often than others that provide a little bit of an editorial skeleton. If you have those tentpoles in there the rest of the thing can be wildly varied and still feel like a structured story with an arc. And, you know, even if in our programming we haven’t suggested connections between different scenes and themes, your brain still kind of does it anyway. And each [viewer’s] brain does it in a slightly different way, which is the fascinating thing. People come up and say, “I love that you went from the scene of Brian talking about being curious to the next scene where he’s looking at bugs in the garden.” And I’m like, “That’s great, but we didn’t know there was a connection there, or not one that I was aware of, but now that you say it…” Our brains want to solve puzzles. We want to make connections between things. So that’s happening here too.

[The current Film Forum] generation is different than even the Sundance [generation]. We continue to change and tweak it in terms of how it’s constructing the story. I’ve watched the film now 32 times with audiences, every minute of it. Normally, I can watch [a film of mine] five or six times and then I can’t look at again. That’s one thing that’s different with this project and part of the reason that I wanted to do it — I wanted to be surprised when I watch my film.

Filmmaker: Thinking of the film as broken into acts, does the third act have, say, a predominant set of scenes to draw from, or is the film randomized all the way through?

Hustwit: It’s variable all the way through though the third act will probably pull in certain scenes that are more thematically linked to things that [Eno] is talking about in the third act. So, yeah, there is an underlying thematic structure to each area of the film, which is another thing that gives it that sense of an engaging narrative arc. But then there are scenes that could be anywhere in the film, these kind of little moments in time, interstitials. But there’s no rigid chronology as to how the scenes come out. It’s more about ideas and the kind of subject matter of what he’s talking about that dictates how scenes will be grouped.

Filmmaker: You say you have watched it 32 times with audiences. What sort of conclusions did you draw from those viewings that then made you change the parameters of the program?

Hustwit: Do certain scenes work back to back? Is there a connection [between them], or not enough of a connection, or too much of a connection? There are definitely rules, like, “If this [scene] plays then this one doesn’t.” Certain scenes will unlock other scenes and vice versa because they’re too similar and would be redundant if they were both in the same version. And then there are features that we haven’t even unveiled yet. Brendan is still making new generative scene functionality that will integrate into versions of the film later this summer, although maybe some will appear in in some of the Film Forum screenings too. I just added some [footage] to the data set yesterday, and I’m going to keep adding until the Film Forum [run begins]. So, it’ll have some new tricks [as compared to] even what you saw at the press screening. That’s why I love this approach. It’s not like, “We made this film, and now it’s done.” It can keep evolving and changing.

Filmmaker: Historically, ENO seems to represent this interesting moment in the middle of this whole AI journey in that the human presence, with you as the creator, is very much involved all the way through. One can imagine a version in the future that’s completely automatic, from DCP creation to theatrical delivery. But you’re constantly curating the archival, deciding personally what’s part of the data set, which is very different from a generative AI model that’s using whatever has been scraped off the internet. You’re dealing individually with distributors, supervising the DCP creation. You’re kind of the Wizard of Oz figure behind the curtain and, as you say, redefining the role of the director in the release process.

Hustwit: It’s definitely more performative. My background is in music —independent labels and bands — and when I first got involved in film it was through music documentary, like the Wilco film [I Am Trying to Save Your Life]. I’ve always wanted the film exhibition process to be a little more like music — you know, that idea that the band can change the song at any time. Just because the song has been recorded, does that mean they have to play the exact version note for note at a live show? This idea that the film is a fixed medium that can never change is something I want to reject. I’m not saying that linear film is dead and that you can’t make linear films anymore. But we can experiment with this, and we have better technology to experiment with these kind of formal changes to a film. I’m trying to encourage more thinking around that.

Filmmaker: What sort of dialogue do you see the film having with film criticism? Most of the press has been positive, but there was an Indiewire review that seemed really aggrieved by the concept of a director relinquishing so much control over the final edit of their film.

Hustwit: Which I think is strange. Just to come back to the music parallel, it’s like, if you saw a band in Boston on such and such a date, and this is what the setlist was that night, it might be the same as the one they played in L.A, or maybe they changed it. [Similarly], you judge [ENO] on what each version of the film does, and the [different versions] do different things. People ask, “Are there good outputs and bad outputs?” I’m like, “They are all good, but they just have a different feel to them.” I guess they’re all funny, because Brian has a great sense of humor, but some are more introspective, some have more music and others it’s more talking about ideas. They have these different personalities from version to version.

[With ENO] I wasn’t trying to disrupt film criticism, I just wanted to make a documentary like I would normally make but have it be different every time. That was all I was trying to do. And now we’re releasing it in England with Picturehouse in 17 theaters and a different version every day for a week of screenings is 70 DCPs to 17 different theaters. Navigating all the technical [issues] of showing this film theatrically has been a challenge.

Filmmaker: I imagine that these demands are taxing the back offices of distributors like a Picturehouse.

Hustwit: Exactly. A lot of cinema chains and streaming networks are not ready for this yet, but we’re trying to work within the legacy systems as much as possible and also just level up everyone’s technology to be able to do this and to [work with this technology] with other films in the future. It’s been amazing to get feedback from other filmmakers and creators who have ideas for this technology that are different from how I’m using it. We’re trying to work with studios and other filmmakers to bring it to fiction films. You could have a Marvel movie that changed every time you saw it. It doesn’t have to be about different endings but different paths, because there are a million different ways to tell the same story.

Filmmaker: There have been other experiments along these lines, the sort of “choose your own adventure” form of storytelling. Is there a proprietary “secret sauce” to the technology you’ve created?

Hustwit: Yes, it’s a proprietary system, patent pending. It’s initially a very basic platform in a lot of ways, and we’re continuing to evolve it. We started working on this five years ago, and the overall technical capability of the software has increased exponentially. There’s so much more than we haven’t have it done, but to go back to your earlier question I’m not sure if it will be completely automatic in the future. [In the future] there will definitely be segments of filmmaking or commercials that can be completely automated, but, you know, I’m a documentary filmmaker, and I like real footage of real things in our world.

Filmmaker: I know you as both a great filmmaker but also as a very smart and canny business person who finds new ways to market independent films. With Helvetica you created a film that reached far beyond the filmmaking community and because there was such a large addressable audience of design fans you could do your own DIY distribution. Obviously with ENO there’s a huge built-in fan base, but, also, that there are multiple versions means new ways to monetize the film. I notice, for example, on your website that for $7,500 you can screen an entirely unique version for a private group.

Hustwit: That’s with me there. It’s not that much if I’m not there. That was just in response to needing to limit the amount of those “come to our college” screenings. I just thought, “Okay, I’m going to do five this year.”

Filmmaker: I guess my question is, how much did the unique distribution opportunities of this approach inform the development of the overall project itself?

Hustwit: Awareness of or the exploitation of those things was an afterthought. It wasn’t like I thought, “If I make a movie that’s a million versions of Brian Eno’s story then fans will have to go watch it a million times.”

Filmmaker: I actually did think that had to be at least a partial motivation. I’m a huge fan, and I plan to watch the movie more than once. I saw it for free at a press screening, but I’ll be buying a ticket at Film Forum.

Hustwit: That’s great. You’re part of the experiment. But for me, that’s secondary. There have been times when I have wished this was a conventional film that we could have sold at Sundance and then I wouldn’t have to be educating cinemas and streamers about the technology and how it can work with their systems. Until I got to Sundance I didn’t really think, “Oh, it’s not going to work with Netflix.” Or, that everyone is going to want a director’s cut. I could do a dynamite directors cut, because I know where all the best stuff is, but that would defeat the whole purpose of the project. It’s not about “my version,” it’s about “everybody’s version.”

Filmmaker: Is there going to be a home video version?

Hustwit: Definitely, but it might be different every day. It might be a film that’s always on, just changing and evolving 24 hours a day. You could watch it for an hour or five hours. Or maybe it’s quiet and ambient overnight and ramps up in the morning. Or it’s different for different seasons and different parts of the world. Or I could stream a unique version to everyone who wants to see it. The problem now is that I can’t do that for a million people, or 100,000 people, or even 10,000 people at the same time, but those are the systems we’re working on creating right now.

Filmmaker: So, there will not be a physical media version?

Hustwit: I don’t think so. The project doesn’t lend itself to a fixed medium. But we can make an app. There are all kinds of different ways to do it. Over the next several months we’re doing live events where we’re creating [the film] in real time on stage. And then these theater runs where there’s a different version every day we’ll continue doing until the end of the year.

Filmmaker: What’s your favorite Brian Eno album?

Hustwit: Oh my god.

Filmmaker: Or your favorite today?

Hustwit: I’ve always loved Music for Airports. It’s definitely one of my favorite albums. It’s not like I listen to it all the time, but it has some super deep meaning for me personally. I love Another Green World. And as kind of weird and flawed as it sometimes is, I like Here Come the Warm Jets. There’s stuff on those early records, like Taking Tiger Mountain by Strategy, which is post-punk before post-punk.

Filmmaker: The song “Third Uncle.”

Hustwit: That song is incredible. One of my first exposures to [Eno’s] solo stuff was hearing Bauhaus’s cover of “Third Uncle” when I was in college. And, of course, I knew he produced Bowie and everything else. I really like the collaboration he did with Tom Rogerson, Finding Shore. I like a lot of the collaborations, like the one with [David] Byrne, My Life in the Bush of Ghosts. And, you know, he is one of the most prolific artists. If you just look at what he did in the seven-year span from 1973 – 1980, it’s mind boggling. And he’s still in his studio eight hours a day making music. That’s all he wants to be doing. He doesn’t want to be talking to journalists or participating in documentaries, he just wants to in the studio, sonically exploring and making these worlds.

Filmmaker: In the spirit of this project, I thought I would use Eno’s and Peter Schmidt’s Oblique Strategies to form the last question. I just found a website that has them randomized, and I’ve called one up that I’d like you to address within the context of making your film. Unfortunately, it gave me one that’s kind of a layup for you, I feel. I thought of refreshing the page and coming up with a better one, but doing so felt like it would be against the spirit of the project, which is to live with what the computer generates. The card I drew was, “Magnify the most difficult details.” How did you possibly do that in the making of your film?

Hustwit: Interesting. That does kind of describe some of the visual aspects of the project. The programming and all the work that Brendan has done is, I think, the real story, and highlighting the work throughout the film [by showing the computer code and source files] is one way we’ve magnified it. It’s funny, the whole idea of having an Oblique Strategy card in the movie was something we went back and forth with. I was like, “Is this too corny to have David Byrne or Laurie Anderson picking a card?” But it is a fun device, and it skews each version of the film in a different way. Depending on what comes up, it unlocks certain scenes. It’s an organic use of Brian’s ideas about creativity as a kind of action point in the movie. Now that I’ve seen a bunch of [screenings], I’m glad we did it.