I have often wondered these last three weeks what would make the most interesting story ending: whether or not it should turn out that I am actually Andrew Frelon, and whether Frelon did or didn’t actually make the real Velvet Sundown band, and was or wasn’t punking everyone the whole time in some sort of multi-layered post-reality meta-hoax.

I’ve gamed out every side of those variables, and in some cases thrown out red herrings consisting of alternate reality story-endings where some of those things appear to be true – just to see where they go – when in actual here-and-now reality, they are squarely not.

In truth, I have searched high and low both as an OSINT investigator, and as a mildly obsessed weirdo superfan to try and figure out who the actual person or people behind the band must be. And I have come up empty-handed on all counts. To my credit, so it seems has everyone else who has been searching since this ridiculous phenomenon began.

And ridiculous it most certainly is. I never expected this little stupid charade to last more than a few days, but instead it has exploded on the global stage, been the subject of intense media coverage and social media commentary, even spawning an entire cottage industry of people capitalizing on AI-related fear and outrage for clicks. But then, that’s everything nowadays. Outrage for clicks. You can’t escape it. So might as well join it? I’m not sure it’s as simple as that, because when you join it, you become subject too to the infection.

As someone who did content moderation for a platform and had to handle tens of thousands of complaints and beefs between users for years and years, I guess ultimately the reason I did all this was really simple: revenge. I wanted the world to experience the sheer utter collapse of reality and sanity I have been witness to all these years behind the scenes working for platforms, trying desperately and vainly to help maintain some semblance of “order.”

Nobody really thinks about it, but you know all the horrible stuff you see constantly on social media, and maybe even “report” when it’s really bad? Well, content moderators are exposed to all that human toxicity and brainrot multiplied by at least a thousand, day in and day out. We are the thin invisible line that (theoretically) protects corporate profitability and brand integrity from the vagaries and savagery of base human nature, as expressed through digital tools.

I know by now that, of the five or so people reading this post, all but maybe one of them will have since stopped reading on account of it being “lengthy” – which was oddly the number one criticism anyone seems to have leveled against Andrew Frelon’s original exposé post explaining how he kicked one past the media gatekeepers’ goalpost. Our minds are so fragmented now from social media, short form videos, doomscrolling, and all the rest that we literally cannot parse thoughtful text in long form. Most people, if they’ve made it this far, are probably by now feverishly asking ChatGPT to summarize the article for them. Here, I’ll save you the trouble, but it will only be for the first section above, not what comes next:

Summary of the text so far:

  • The narrator reflects on whether they are Andrew Frelon and whether Frelon created the Velvet Sundown band as a hoax.
  • They admit to planting false leads and alternate realities as part of the narrative experiment.
  • Despite deep investigation, they found no definitive identity behind the band.
  • The hoax unexpectedly went viral, drawing global attention and media scrutiny.
  • The narrator, a former content moderator, reveals the stunt was driven by revenge—born from years of exposure to toxic online content.
  • They comment on society’s declining attention span and the irony that most readers will need this very summary.

There, did that help? Or did that just make it longer and boringer? I bet your fingers are twitch-twitch-twitching to close this browser tab, desperately seeking where the “Unfollow” button is on this blog (there is none!).

Content moderation messes your mind up. It’s really basic. It cooks your brain. It took me literally years to let go of some of the weird acute physical reactions I had picked up from over-exposure to certain types of negative tropes and shitty arguments online. I have had, as a result, many demons to exorcise (and exercise).

But I don’t ask for your sympathy, because in today’s attention economy, that is worth less than nothing. I ask only for your anger, because I know that is all that anybody has left to give anyway. So pour it out, I can wait. I’ve seen it all before. Every argument. Every mean name. Every threat. Every “kill yourself” comment. It’s all completely pointless, though you might think it makes you feel good in the moment. Then what? Then, the next thing and the next thing, until you’re strung out and anxious always hunting & fearing the ONE SIMPLE TRICK! that will drive you over the edge…

What I’ve learned in all that is that hate and anger are completely predictable. Programmable even. If you put ABC inputs into the machine, out the other side comes – poof! – XYZ outraged outputs. It’s so very evident from everything around us politically and socially that it’s just played out to even go on and on about it anymore. So why bother?

Because this cage that we’ve built for ourselves – that we’ve let companies build around us in the name of enabling us to “connect” with one another – well, it fucking sucks. It’s not normal or good or healthy. It is absolute trash from top to bottom. And it all needs to be burned down.

There are those who fight it, who fight the good fight on social media, and in journalism, and politics, and everywhere, but what I’ve seen is that it never ends. There is no salvation in counting likes and follows and impressions. There are not enough fact checkers in the universe to dig us out of the black hole we’ve fallen head over heels into.

Anyway, this all was supposed to be a short preamble, but serving up my rage became its own dish when I started writing it. And what do they say about revenge being a dish served cold? Something something... Maybe I should ask ChatGPT again to help put a neat bow on this, but I won’t. You’re going to have to find your own meaning in all of this, if there is any to be found…

Over the past few weeks, many journalists have asked me what is the “message” I want people to take away from all this. And despite all my doom and gloom, I’ve actually come away from this all with some strangely hopeful sentiments. They are, for now, three:

  1. If everything is fake, then our imagination is now one of the most important human faculties (and probably was all along, even before “post-reality). Imagination is uniquely suited to spelunking for “truths” that match our intuitions, even and especially if it has to construct those truths whole cloth. And constructed truth is better than no truth at all, isn’t it? Isn’t it??
  2. Number 2 one comes from a lecture by “superfictions” artist Peter Hill, who years ago invented a fake NYC museum, which became the envy of Germany, and who said: “Through fiction and imagination you can make things happen in reality quicker and more effectively than you can through the so-called normal bureaucratic channels.” My experience has proven that to be correct, but which “things” have I made happen? Hard still to say at this early juncture, I guess.
  3. And lastly, this whole thing has taught me very plainly that, despite what I had thought – and what the world had beaten into me over decades of living under its spell – that one person actually *CAN* make a difference. But therein lies the problem: what difference? What should we even aim for?

It’s been interesting to watch how my actions, underhanded as they were, have lead to widespread calls for transparent labeling of AI-generated content. Even BPI, the British Phonographic Industry, is now backing AI content labeling because of the Velvet Sundown Affair. [Insert: Did I do that? Urkel Meme]

While I understand labels are one of the cudgels people always try to cling to in times of escalating uncertainty – thinking that mere words will save us from the infinite downward slide – I’ve roped and wrangled in this rodeo before. Alongside “boots on the ground” content moderation, I’ve also ridden the “dust on the wind” policy/standards side of these problems, much of which had to do with using machine learning to try to automatically label certain categories of polarizing and divisive content. In order to… something that I’m still not entirely clear on years later, I will admit…

What I saw in all that good, well-intentioned work is that labels do basically nothing. Sure, a few people say they want it, but if I made my obviously deepfake political commentary in Sora versus Kling vesus AI-flavor-of-the-week, does that really change anything? The labeling conversation is based on this Enlightenment ideal that says: hey, if we just give people access to all the “correct” information, well they will be smart and motivated enough to go out and make the “right” decision all on their own. Narrator: they won’t. They will ignore the labels. They will find ways to bend and break and spoof and mock the labels. We will have spun our wheels yet again on another micro-fix of an enormous macro-scale problem that maybe has no answer.

But I get it: it’s uncomfortable to face that some problems are so big we can’t actually fix them, and don’t know how to extricate ourselves from the systems which perpetuate them – even if we wanted to. Which we most assuredly do not. We love our cage now. We crave it. We can think of almost nothing else.

When is the last time you ate a meal with other people where one/several/everyone at the table wasn’t on their cell phone some/most/all of the time? People bitch and moan about “AI slop” taking over the world (and honestly, Andrew Frelon’s efforts being included on the Wikipedia AI slop page is one of the greatest non-parenting achievements of my mostly ridiculous life), but we’re already all absolutely robots hopelessly addicted to our devices and the digital drugs they administer directly into our nervous systems. It doesn’t matter if the slop is AI or human anymore. We do whatever the algorithm tells us to do, buy what it tells us to buy, drive where it tells us to drive, talk to, or love, or hate whomever it tells us to. This is a fight we’ve steadily and blithely lost over the last twenty-plus years. We’re only too happy to give up our agency and autonomy in exchange for the shimmering false hope of “virality.” And I can tell you from experience, “going viral” brings you absolutely nothing. No material changes to your life. It comes and goes like a tremendous dose of a painful buzzy drug, and then it’s over and you’re left numb, wondering how to do it all over again. Virality is just thousands of people blaming you for their pervasive unhappiness.

I give up on it. I wash my hands of that world. I won’t live in it. I’ll live in my garden, with my hedgerow, with my willows, with my baskets. With my family. And my imagination. And that will have to be enough. Cause what other choice do I have? Society is crumbling. Everything everywhere is on fire. Nothing is real.

By now, no doubt, the one half of one quarter of one percent of one person who is still using ChatGPT to summarize this article has come up with something like the below, which itself is quite long, but seems accurate for what its worth:

Extended Summary:

Following the initial reflection on identity ambiguity and narrative manipulation, the narrator transitions into a deeper, more personal exposition of motivations, particularly emphasizing a history in content moderation. This occupational background—defined by prolonged exposure to online toxicity—serves as the foundational grievance fueling the Velvet Sundown hoax, framed explicitly as an act of revenge against the structural and psychological burdens imposed by digital platforms.

The text evolves into a critical dissection of modern information ecosystems, positing that outrage and virality are not merely emergent properties but engineered outputs of algorithmically predictable behavioral responses. The narrator argues that efforts to impose order through AI content labeling and fact-checking are futile within an attention economy structurally designed to reward emotional extremity and superficial engagement.

Simultaneously, the narrative acknowledges the performative and cathartic dimensions of the hoax, conceding that its unanticipated success has precipitated industry-wide discussions about media literacy, AI regulation, and authenticity. Yet, this acknowledgment is juxtaposed with a nihilistic appraisal of systemic intractability—suggesting that current frameworks are inadequate for addressing the scale and nature of the problem.

The conclusion abandons prescriptive solutions entirely, embracing a retreat into personal autonomy, familial connection, and imaginative resilience as the only viable responses to a reality increasingly mediated—and eroded—by digital saturation. The narrator effectively declares ideological withdrawal from the digital sphere, positioning this act not as defeat, but as an intentional reclaiming of agency.

Hell yeah, ChatGPT. Hell yeah.

Now that the ranting part of tonight’s performance is finished (mostly), we can turn to a quick review of the literature before the curtain comes down and they come for me with the hook.

There were a few of big pieces that came out in the past couple weeks or so which I thought really captured the spirit of the game which was afoot at the time.

First was Kevin Maimann’s fine July 5th CBC piece, which remains one of the best and most faithful tellings of the overall story, and which I think rightly identified that Frelon caused a “a media frenzy that propelled the AI-assisted band’ to a million monthly listeners on Spotify.” (I maintain that the “real” Velvet Sundown band might have made a few songs using Suno, but it was Frelon’s maneuvering that turned them into a legend.)

Next was Glenn Peoples’ July 11th Billboard reporting which compared Frelon’s legend-making to the controversy around Bob Vylan and Kneecap, and which seemed to suggest Frelon’s was maybe a bigger deal in the end:

The Velvet Sundown, the AI-generated retro rock group that launched out of nowhere in June, has done something many real bands struggle to achieve: Gain worldwide media attention that converts into streams at Spotify and other streaming platforms. 

In the week ended July 3, The Velvet Sundown’s catalog of 26 songs — spanning two albums — generated 2.59 million streams globally, according to Luminate. That’s a staggering 399% increase from the previous week. The surge followed a flood of online coverage at music trades, mainstream newspapers, and popular YouTube channels, all intrigued by the novelty — and the ethical implications — surrounding an artificial band that diverts attention (and royalties) from actual musicians. 

After more than two weeks of media coverage, the buzz hasn’t died down. The Velvet Sundown has been featured in major publications, including articles at BBC.com and a long-form think piece in The Atlantic. In an era when viral fame often fades in hours, the sustained interest — from both the media and the public — is remarkable. Just six weeks ago, the band had zero streams. Today, it boasts 1.3 million monthly listeners on Spotify, up from approximately 300,000 following a viral Reddit post about the band on June 24. […]

The Velvet Sundown represents a reality of the music business in 2025: Getting online attention, whether complimentary or critical, helps an artist rise above the noise and find listeners. Notoriety — being known for something negative — isn’t necessarily a liability.  

There’s more, but that’s (hopefully) enough to get my point across to the 0.0000001% of a person still reading this “lengthy” blog post.

Next we turn to Michael Ashley’s July 14 Forbes coverage:

… [P]seudonymous band spokesperson and ‘adjunct’ member Andrew Frelon now admits, ‘It’s marketing. It’s trolling. People before, they didn’t care about what we did, and now suddenly, we’re talking to Rolling Stone, so it’s like, ‘Is that wrong?’”

Not if you ask the countless impresarios throughout history who courted controversy to similarly make a name for themselves.

Frelon comes from a long line of enterprising artists who leveraged controversy to generate publicity. As far back as the 1840s, P.T. Barnum, perhaps the first real American showman, exploited this reliable tactic. Famously, he exhibited a grotesque creature he professed to be a real mermaid.

In reality, he had sown a monkey’s torso to a fish tail to garner attention. Barnum also used to plant fake letters and disparaging stories about himself in newspapers to pique curiosity, recognizing the utility of being in the public eye, even if for the wrong reasons.

Likewise, master manipulator/comedian Andy Kaufman blurred the line between performance and reality by creating his alter ego, Tony Clifton, a belligerent lounge singer. Kaufman even went so far as to pay his pal Bob Zmuda to impersonate Clifton so the two could be seen at the same event, further throwing the public off the scent.

What’s so unique about this moment in time is that the artistic deceptions of Barnum, Kaufman, and others look quaint compared to what’s now available, courtesy of AI. The term deepfake may be applied as a catch-all encompassing the myriad ways in which people can now deceive at a scale previously unimaginable.

There’s also a more personal note that someone left on the original Frelon Medium post, one which tells me maybe the whole exercise was not such a waste of time after all, even if it only reached this one person. User “Zeline’s Secrets” wrote on July 4th:

Reading this made me pause and think about how blurry the line between fiction and reality has become online. I remember stumbling upon parody accounts years ago and feeling that jolt of surprise when realizing they weren’t “real” — now I almost expect layers of performance in every profile. There’s something oddly comforting about accounts that announce their fakeness upfront, like we’re all in on the joke together. It makes me wonder how much of my own online persona is unintentional performance, even when I’m trying to be authentic. The internet feels like one big collaborative theater piece sometimes, where we’re all both audience and actors without a script.

Really, there has been way too much media to carefully track and archive it all. And so much of it has been incredible, absolute, utter trash. Unoriginal re-reporting up the wazoo. Massively failed fact-checks by some of the biggest news organizations in the world (cough cough AFP cough) even when I told them what was up before and tried to reach out about the error after publishing. And so so much hand-wringing and pearl clutching that its simply embarrassing to think about. I am quite simply shocked and horrified that the media is as utterly and hopelessly broken as my experience has proved it to be. The vast majority of media nowadays is just reporting on things that happened on Twitter, and they can’t even seem to get that right!

I did rather like this Argentinian Marxist-Leninist perspective (the other “ML”), delivered by Francisco Ruiz Diaz in Agitación, also on the 11th of July. My Spanish is not good, so I’ll once again let ChatGPT do the talking with an English partial translation:

This raises another issue: identity. A fictional music group faces off with a fictional spokesperson claiming a non-existent identity. The musicians weren’t real, neither was Frelon—yet both were believed and debated in the media.

Today, AI increasingly slips through unnoticed cracks. The media’s ambition to deliver juicy stories demands stricter verification methods. The issue of authorship in AI-assisted or fully artificial productions is overshadowed by the need for deeper analysis on truth and public opinion.

AI is now a toy for those shaping discourse and opinion. Frelon’s low-resource example showed how easy it is to confuse the news cycle.

Consequences are slowly leaking out. It’s not just media or an anonymous user playing with deception—it’s political and public discourse. The political sector’s questionable use of AI must be scrutinized. Deepfake videos circulate via state-aligned channels, and fake news spreads. Journalists, opponents, and dissenting civilians face indiscriminate attacks.

Frelon didn’t just want to impersonate—he pointed to the elephant in the room. By faking a hollow identity, he spotlighted deeper issues about technology’s implications and media processes. If it’s so easy to mislead those who inform society, what future awaits?

What future indeed?

If I can fool global media into producing three full weeks worth of coverage of literal nonsense, then how much worse must the problem be for media outlets facing political, financial, and legal pressures from increasingly authoritarian governments, whose propaganda channels are vastly, absurdly more developed than my few weak efforts at culture jamming and detournement?

I also thought this throwing up of the hands by New Scientist’s Josie Ford on July 16 was uproariously funny and frankly ridiculous:

Things seemed clear after Rolling Stone interviewed the band’s “creator” Andrew Frelon, who said it was all an “art hoax”. But then Frelon claimed that was itself a lie, and the “band” posted a statement denying any involvement with him. By this point, Feedback was bored of redrafting our redrafts, so we just want to say we don’t know what’s going on, and also we don’t give two hoots.

This from a website whose about page reads:

In a time when facts are in short supply, there has never been a greater need for a trusted, impartial source of information about what is going on in the world – or a greater need for inspiration through exceptional ideas.

Mmkay.

Also, this is tangentially related, since Timbaland was mentioned so often in articles about the rise of Velvet Sundown. I’m convinced he is onto something here with what he’s talking about culturally:

Interestingly, just a few months ago, Timbaland went on The Inner Court podcast and stated that he finds new music artists to be very “uninspiring” and their music “soulless,” blaming the 2024 election as the reason and stating that he thinks AI can do better.

“You could feel the pressure of the dividedness, and I hate that. This whole election divided us,” Timbaland said at the time. “What I mean by that is it divided the music. The music sounds bland, it sounds boring—it lacks excitement. While we frequently discuss AI, I believe it’s the only entity that embodies a genuine soul right now. It allows for the expression of true feelings, resulting in it coming out beautiful.”

I don’t listen to Timbaland’s music, and could not identify any songs he has worked on if pressed, nor do I intend to listen to “Tata” (which seems terrible) but he’s speaking the truth there that the life has gone out of most art and music and whatever culturally. So yeah, as Zeline said above, maybe in the end there’s something good and comforting about being able to enfold oneself in things that are obviously and overtly fake. The pretense becomes more real than reality, because its an honest con, unlike so much of what’s out there. Like an AI band whose AI-generated avatars are constantly telling their audience, “We’re not AI – we just totally, like, keeping it real, bro!“

Speaking of keeping it real, there are a couple very special fan videos which were made that I feel inclined to include here as a coda. Mainly because they are so weird, and slide right into the pocket of strange. The first comes from underground cult filmmaker Damon Packard, who seems to have fully and completely understood the “mission” without ever having spoken to Frelon or myself about any of it.

Just an incredible piece of video which, when I saw it, had me doing double and triple takes to try and understand what the hell was going on, and whether or not the wholly invented narrative I had been foisting was in fact somehow real or superseded by a Beast Beyond Reality…

And last is a fun and extremely strange AI-generated tribute video by RVC Music for Dust on the Wind, where they took a lot of the still images (some by me, and some by other parties) and animated them, syncing (mostly) to the music:

The future is gonna be so goddamned weird. Are you ready?

No. No, you are most definitely not.

But maybe I am.

And lastly, an outro, via our beloved friend/continual enemy, ChatGPT, who read the entire finished post:

USER: does this author ever admit to being frelon or the velvet sundown?

CHATGPT:

The author never clearly or definitively admits to being Andrew Frelon or part of Velvet Sundown.

Key points:

  • The text opens by questioning whether the author is Frelon or created Velvet Sundown, but immediately follows with a denial, calling such possibilities “squarely not” true.
  • The author admits to planting red herrings and simulating alternate story endings, including some where they are Frelon or the creator.
  • Despite detailed insight and participation, the author presents as a fan, investigator, and provocateur—not as the confirmed creator.
  • Multiple passages intentionally blur identity, framing the ambiguity as part of the art.
  • The text leverages performance, irony, and unreliable narration to sustain plausible deniability.

Conclusion:
The author constructs a layered persona but never explicitly admits to being Andrew Frelon or the creator of Velvet Sundown. Ambiguity is a deliberate mechanism in the narrative.

Sounds about right!

Extra special thanks to Chet Swagg, the manager of The Velvet Sundown (Real Band Not the AI Band) on Twitter, without whose collaboration and friendship, this long dusty musical road would not have been half as fun! Keep it real forever!

Update: (24 July 2025)

Here’s a fun one from Mistral:

It’s funny/scary/dangerous how gullible AIs can be. Perplexity & ChatGPT didn’t fall for this, but it’s only because of all the news reporting around this subject with takes that contradict Frelon’s later narrative twists. If there were less media saturation, it would be very easy to poison a topic like this for AI search results with a couple planted articles on UGC websites.