Velvet Sundown, or, How Scared Should We Be of AI Music?
23:41 Share

Thank you.

Welcome back, friends. Quick headlines before we dive in. First of all, thank you to today's sponsors, KPMG, Blitzy, Plum, and Superintelligent. To get an ad-free version of the show, go to patreon.com slash ai daily brief. And if you are interested in sponsoring the show, shoot me a note at nlw at breakdown.network. We are starting to get a little tight on fall inventory. So if you are interested, definitely reach out. Today's episode is an interesting cultural exploration. So without any further ado, let's dive in.

Welcome back to the AI Daily Brief Headlines Edition, all the daily AI news you need in around five minutes. Today's first story could easily be a full main episode because it is not just a story, it is a major political realignment, but let's talk about it a little bit.

The short of it is that the Senate has stripped out a curb on AI regulations from Trump's tax bill. Tuesday night's voterama on the so-called Big Beautiful Bill sifted through hundreds of amendments to the omnibus funding legislation. Among them was an amendment that removed a 10-year ban on states passing AI regulations. The provision was hugely controversial heading into the vote. It received heavy lobbying support from the tech industry, with Andreessen Horowitz, Palantir, and Anderil among the most outspoken supporters.

The argument from AI companies was that the maze of state regulations was impossible to navigate for the fast-growing technology sector.

And indeed, various states have already passed more than a dozen different AI laws with over a thousand more proposed. AI czar David Sachs and Commerce Secretary Howard Lutnick were also in favor, arguing that any constraint to AI growth is a national security risk. The president, for his part, had stayed silent on the moratorium. Concerned citizen groups lined up on the other side, arguing that consumer and copyright protections should be paramount. Many made the point that without a federal law in place, AI would be unregulated for the time being.

Senator Marsha Blackburn of Tennessee led the drive to scrap the moratorium, acting on behalf of the Nashville music industry. Tennessee recently passed the Elvis Act, which prohibits non-consensual use of AI to mimic musicians' voices, which would of course have been one of the laws that would have been blocked. Ultimately, the moratorium on state AI regulations was removed in a 99-to-1 vote, with support in the Senate collapsing entirely. Even Senator Ted Cruz, who championed the moratorium, ended up voting to remove it after negotiations on a compromise collapsed.

Retiring Senator Tom Tillis was the lone vote against the amendment. Now, Tillis has become a bit of a chaos agent since he announced that he won't seek re-election on Sunday. He opposed the larger bill itself and voted against every single amendment, so there's pretty much no signal there.

Wired reports that a compromise was basically in the bag over the weekend. It would have cut the moratorium down to five years and include a string of carve-outs for consumer and copyright protection. But the reports are that Steve Bannon essentially scuttled the deal from his podcast studio, arguing that, quote, in the first five years, they'll get all of their dirty work done. After the votes were cast, Bannon posted, massive win for MAGA over tech pro oligarchs.

Celebrating the victory, Brad Carson, the president of AI Safety Group Americans for Responsible Innovation, said, All right, so let's talk about this for a second.

First of all, there are many reasonable debates to be had around the complication of the state and federal system when it comes to emergent technologies. It is absolutely the case that companies having to deal with different regulations in 50 different state jurisdictions can be a huge drag. We've seen this in financial services. We've seen this in lots of other areas before.

Point being, you don't have to believe that the technology companies are just trying to get away with never being regulated at all to understand why they would be in favor of this amendment.

At the same time, there's probably a pretty reasonable position that the better response than a ban on states regulating is superseding federal regulations that actually create understandable rules of the road. But none of that is what I actually think is interesting about this story. One of the major new groups in Trump's coalition this time was Silicon Valley. And now, post-breakup with Elon, Silicon Valley is once again a complete political orphan.

MAGA has completely broken with big tech. Representative Marjorie Taylor Greene, for example, posted, banning states from regulating AI for 10 years is a gift to big tech and a disaster for American workers and states' rights. And so now we have maybe the weirdest coalition in American political history of AI safetyists on the one hand, MAGA on the other. And if that doesn't say just how chaotic AI is going to make everything, I don't know what does.

Speaking of that Elon Trump breakup, or at least specifically Elon, XAI has closed $10 billion in new funding, but investor enthusiasm seems to be at least a little bit faded.

This deal has a troubled history. And the trouble is, of course, directly related to that breakup. At the beginning of June, the Wall Street Journal reported about the pitch meeting where Morgan Stanley was marketing $5 billion in XAI debt. Investors had a split screen reading through some very sparse slides that lacked financial details while watching Elon Musk set his political career on fire live on X. One detail that was included in the financials is that XAI is apparently now burning a billion dollars a month.

Now, we know that AI is capital intensive, but there's still a dramatic gap between XAI's expected $500 million in revenue and that significant burn rate. Now, in response to the reporting, Elon Musk said Bloomberg is talking nonsense, but it's hard to tell what's real anymore. Bloomberg is now reporting that three additional investment banks were brought into the deal before it closed. Not super unusual, but implying to some that Morgan Stanley perhaps did not have enough takers within their own client base.

Now, this could all be nothing. Morgan Stanley said, this transaction, which was oversubscribed and included prominent global debt investors, reflects confidence in XAI's vision to accelerate scientific discovery and advance humanity's collective understanding of the universe. Still, the debt was priced at around 12% interest, close to the current yield on CCC-rated junk bonds.

Lastly, alongside the $5 billion in debt funding, XAI has also closed $5 billion in a strategic transaction. We don't know who that strategic partner is or how the company was valued in the deal, which only adds further questions.

The only reason that all of this is notable is that XAI and Elon in general have had absolutely no trouble whatsoever raising funds in the past. Their last round in December raised $6 billion and included a murderer's row of 14 of the largest investors in the world, including Andreessen Horowitz, BlackRock, Fidelity, Kingdom Holdings, Lightspeed, MGX. You get the picture. The only returning name for this deal, at least that we know, was Morgan Stanley. And their only participation was marketing the debt for a fee.

And yet, despite all the warning signs, you kind of have to revert to the axiom of never betting against Elon. This is still a gigantic pile of cash that will fund the expansion of the Colossus supercluster and probably see the company through to the end of the year. What's more, fading the intensity of the XAI team does not seem like a winning proposition right now. Pictures are circulating of a tent city inside the company's office as engineers work around the clock to put the finishing touches on Grok 4.

Lastly today, speaking of the big labs, Anthropic has hit $4 billion in ARR, but the competition for AI coding is getting more intense. The release of Cloud Code has helped Anthropic quadruple their annualized revenue since the beginning of the year. However, their symbiotic relationship with Cursor seems to be turning a little sour. The information reports that two key leaders of the Cloud Code project have been poached to join Cursor as head of engineering and head of product.

They wrote,

McKay Wrigley wrote, insane that Cursor saw the rise of Cloud Code and went, yeah, let's just hire both Cloud Code's lead dev and project manager away from one of the world's biggest AI labs. Absolute monster hires. And a reminder, you can just do things. Talent world continues to be crazy, but for now, that is going to do it for today's AI Daily Brief Headlines Edition. Next up, the main episode.

Today's episode is brought to you by KPMG. In today's fiercely competitive market, unlocking AI's potential could help give you a competitive edge, foster growth, and drive new value. But here's the key. You don't need an AI strategy. You need to embed AI into your overall business strategy to truly power it up.

KPMG can show you how to integrate AI and AI agents into your business strategy in a way that truly works and is built on trusted AI principles and platforms. Check out real stories from KPMG to hear how AI is driving success with its clients at www.kpmg.us slash AI. Again, that's www.kpmg.us slash AI.

This episode is brought to you by Blitzy. If you're a technology leader, here's something that probably sounds familiar. Your organization's competitive edge is buried in legacy code that desperately needs modernization, but the resources required feel out of reach. That was the case for a global investment analysis firm. They needed to migrate 70,000 lines of complex MATLAB financial algorithms to Python. Algorithms that drive investment decisions for trillions in assets. Their estimate? Months of high-cost specialized engineering work.

Instead, they partnered with Blitze. Blitze's autonomous AI preserved mathematical precision and generated over 80% of the new codebase, completing the migration with just five days of engineering time. They cut the timeline by 95% and saved 880 engineering hours. If your organization is facing similar modernization challenges, visit Blitze.com to schedule a consultation and discover how AI-powered development can transform your technical capabilities. Today's episode is brought to you by Plum.

You put in the hours, testing the prompts, refining JSON, and wrangling nodes on the canvas. Now it's time to get paid for it.

Plum is the only platform designed for technical creators who want to productize their AI workflows. With Plum, you can build, share, and monetize your flows without giving away your prompts or configuration. When you're ready to make improvements, you can push updates to your subscribers with a single click. Launch your first paid workflow at useplum.com. That's Plum with a B, and start scaling your impact.

Today's episode is brought to you by Superintelligent, specifically agent readiness audits. Everyone is trying to figure out what agent use cases are going to be most impactful for their business, and the agent readiness audit is the fastest and best way to do that.

We use voice agents to interview your leadership and team and process all of that information to provide an agent readiness score, a set of insights around that score, and a set of highly actionable recommendations on both organizational gaps and high-value agent use cases that you should pursue. Once you've figured out the right use cases, you can use our marketplace to find the right vendors and partners. And what it all adds up to is a faster, better agent strategy.

Check it out at besuper.ai or email agents at besuper.ai to learn more. Last month, a new psychedelic rock band called The Velvet Sundown started appearing on people's playlists on Spotify. They dropped two albums and quickly racked up a couple of hundred thousand listeners on Spotify. The only problem is they probably don't actually exist. A Reddit user on the anti-AI subreddit posted,

Behold, a completely fake band on Spotify with 300,000 monthly listeners. Not a shred of evidence on the internet that this band has ever existed. AI-generated artist photo and album covers. Description reads like ChachiBT. A woman on TikTok said it was the first song in her Discover Weekly. I'm 100% sure this content is pushed by Spotify itself to further minimize the amount of money they pay artists.

The potentially generated description, in case you're curious, reads, There's something quietly spellbinding about the Velvet Sundown. You don't just listen to them, you drift into them. Their music doesn't shout for your attention, it seeps in slowly, like a scent that suddenly takes you back somewhere you didn't expect. Their sound mixes textures of 70s psychedelic alt-rock and folk rock, yet it blends effortlessly with modern alt-pop and indie structures. Shimmering tremolos, warm tape reverbs, and the gentle swirl of organs give everything a sense of history without it ever feeling forced.

The Velvet Sundown aren't trying to revive the past. They're rewriting it. They sound like the memory of a time that never actually happened, but somehow they make it feel real. Now, while there might not be any em dashes in this, it certainly does sound like the kind of schlock that you might get from ChatGPT. So is this band an AI construct? And if they are, does it matter? In the past week or so, the Velvet Sundown has become a very hot topic of conversation in music and internet cultural media.

Consequently, we have multiple deep dive investigations trying to figure out whether in fact the band is AI generated. The first and perhaps most obvious indication is the pictures. I think for anyone who has spent any time with AI, they certainly have that feel of AI generation.

The album covers are almost certainly AI-generated, even if the rest of the band is not. But at this point, it should be pointed out that having AI-generated covers and AI-generated descriptions doesn't make the band fake. It just makes them like pretty much everyone else at this point who's using AI for generating images and generating descriptions.

A bigger tell is that none of the band members seem to actually exist. They don't have any social media accounts, haven't done interviews, and in fact have zero internet presence whatsoever. In fact, the band's social media presence only spawned recently, once they started getting media attention. Their Instagram is filled with very obvious AI-generated photos, and the song's credits are also suspicious with no producers or extra writers outside of the band listed. Then there are the songs themselves. Music Radar wrote,

The band's country-tinged roots rock bears the unmistakable lo-fi veneer of a Suno creation, but is convincing enough to pass by undetected if sandwiched in a playlist between two authentic songs.

And indeed, while most of their listens now are curiosity after people have read articles like this one, the initial couple hundred thousand listeners came largely from Spotify's generated playlists like Discover Weekly. A huge amount of Spotify listenership is in the form of personalized playlists that are generated by Spotify's recommendation algorithm. In a post on X, the band denied the rumors.

With a bio that reads, yes, we are a real band and we never use AI, hashtag never AI, they wrote a thread, absolutely crazy that so-called journalists keep pushing the lazy baseless theory that the Velvet Sundown is AI generated with zero evidence. Not a single one of these quote unquote writers has reached out, visited a show, or listened beyond the Spotify algorithm.

This is not a joke. This is our music written in long sweaty nights in a cramped bungalow in California with real instruments, real minds, and real soul. Every chord, every lyric, every mistake, human. Just because we don't do TikTok dances or live stream our process doesn't mean we're fake. The fact that some blog editors would rather pretend we're a bunch of machines than admit an unknown band is out here grinding and made something people enjoy is insulting. We've had to lock down our personal accounts due to harassment, all because some writer wanted clicks and couldn't imagine people like us existing outside their sanitized indie media echo chamber.

Shame on every outlet amplifying this narrative. We are real. Think next time before you erase people. Forgive me for piling on, but methinks the lady doth protest too much. Indeed, when it comes to whether the Velvet Sundown is a real band or an AI-generated art piece, if I were a betting man, I would put basically all of my chips on it being some sort of very intentional experiment.

that's meant to provoke exactly the conversation it's provoking. Now, what I don't know is to what extent it's from a pro-AI or anti-AI source, but it's definitely trying to provoke a conversation. And so let's have that conversation.

One of the things that's really interesting about this is that in other social media channels, think TikTok and YouTube, there is an incredible amount of AI content that is self-professed and unabashed. There's the Bigfoot and Yeti videos, vlog warts, all of these things that I profiled recently in another episode. Music hasn't yet seen that level of infiltration.

Yes, we have had a couple of breakout moments in AI music. King Walonius' BBL Drizzy being used during the whole Drake-Kendrick feud was a really notable one. And again, on Spotify, there are accounts like Beats by AI Official, which use AI to create humorous songs that have racked up just an incredible number of views and 10 million likes, but which again are very, very clear.

Beats by AI Official got its start with a series of posts called Asking AI to Make a Hit Country Song. It did it every day for months, eventually built an audience, and turned that into some very humorous and very not-safe-for-work songs. And yet, broadly speaking, there is definitely a stronger reaction against AI music than AI and other creative genres. A piece in Music Business Worldwide recently was called The AI Music Problem on Spotify is Worse Than You Think.

They point to outlaw country artist Aventus, which has over a million listeners each month, and which is for sure and known to be AI generated. Now, in the case of Aventus, it does appear to be a little bit more of an intentional product from a musician. When someone a couple of months ago asked in a YouTube comment what role AI played in the artist's music, the anonymous owner of the channel replied, the voice and images created with the help of AI, the lyrics are written by me.

And if you go to the credits on Spotify, the written by and produced by is someone named David Vieira.

The article then goes on to point out a couple of other examples, coming of course to the Velvet Sundown. And if you go check out Twitter slash X, there are people like Anju Online who write, not enough music people are actually paying attention to this Velvet Sundown situation. The future of streaming services is about to be so bleak. Coen Rad Sheepers writes, AI now births entire music acts. I came across the Velvet Sundown, which has lyrics, vocals, and visuals all generated by AI. This again just highlights how its creative output is outpacing our ability to regulate it.

Scary part is that I would 100% listen to this over many quote-unquote real artists. Mossimo, at Rainmaker1973, writes, The Velvet Sundown apparently has only existed for two weeks and has over 400,000 monthly listeners. It's a totally generated AI band. Easily, people who primarily interface with music via algorithmically generated playlists will soon only listen to AI-generated music.

Signal writes, There is a lot to unpack here.

First of all, while I agree entirely that whoever is behind the Velvet Sundown is creating an entire experience that understands narrative and is not just about the music, I don't think it's fair to claim that there's actual cultural velocity here. These are not hit songs. These are not diehard fans. These are curiosity listeners, at least half of which have come pretty directly from the articles about the curiosity that is this band.

One way to tell is the disparity between the monthly listeners and the followers. There have been 634,000 monthly listeners, but only 11,000 people have actually decided to follow the band. Now, what's super clear is that there is going to be more of this type of experimentation.

Sometimes people will anonymously use AI. Sometimes musicians will explicitly and clearly use AI. But like any other technology tool, it is going to impact the music that gets created. There are inevitably going to be AI-generated hits. It is just absolutely inevitable. What's more, there is an incredibly large market for background music in everything from videos to games that will likely find its way to AI generation as a cheaper solve.

The question is, how problematic is this? And from a creativity standpoint, I don't think that the introduction of AI somehow threatens human creativity without AI. That's basically what my entire interview with Rick Rubin was about. Where it gets more dicey and where I think people are reacting is to what extent AI-generated music competes with human-generated music

for limited attention. And the irony here is that the AI that I think people are actually angry at

is not, in fact, AI-generated music. It is instead the tyranny of the algorithm. None of this is new, so forgive me for trying to make it sound profound. The reality is that we live in an algorithmically mediated world. The social networks where we consume content, media, news, music, videos, all of it, is carefully curated for us based on micro-expressions of our interests and intents,

all which get sucked into a big AI Borg and spit back out at us things that it thinks we will like. The Velvet Sundown sounds like a very particular type of band. It's not surprising that the Spotify algorithm is picking up that if you like other types of bands in this genre, you might respond positively to them as well.

And I think in some ways what people are reacting to is the idea of being forced to consume something which feels artificial. What makes this whole moment so interesting to me is that we are talking about the confluence of two types of AI, new AI-generated content and the algorithm-run channels where that content gets discovered.

As I discussed in that show about AI video, it's very clear that AI is taking over some of those algorithm channels already. But it wouldn't be all that surprising to me if people have a different type of reaction in music and want a different approach. For

Forever ago, literally more than two years ago, Product Hunt founder Ryan Hoover wrote free startup idea that will likely get you sued. AI Spotify. How it works. AI Spotify hosts AI generated music from your favorite artists. Anyone can submit music and the best songs surface based on listens and likes. Music with the most listens earns a pro rata share of subscription revenue reserved for the original artists.

For example, Drake could claim money generated from his likeness on the platform. Artists that do not want to participate can opt out entirely, banning any music that uses their likeness or individually allow songs they endorse. Of course, there are many ethical and legal issues with this model, especially with labels, but maybe this is a germ of a shower thought that has potential. Now at the time, what we were focused on was how people were making songs that imitated the voices of prominent artists. Obviously now with things like the Velvet Sundown, we're talking about artists that are generated from scratch.

And so perhaps AI Spotify looks a little bit different. But it still wouldn't surprise me if we start to see some number of channels which reach escape velocity, which are based on hosting new types of AI generations separate from other types of human creative experiences. Maybe not, and maybe it all gets integrated into one spot, but it feels like new types of content when they emerge tend to bring with them ultimately their own new networks as well.

In fact, I think it would be surprising if we didn't see something similar with AI. For now, like I said, I'm pretty sure the Velvet Sundown is trying to provoke conversation. So go check them out and then share what you think about it on your algorithmically mediated social network of choice. That's going to do it for today's AI Daily Brief. Until next time, peace.