That's really interesting - although when I read that story I thought the point was the broken recommendation system.

Happens to me a lot on Youtube too - you watch just one thing - and now your recommendations are full of things like that one thing. Or Facebook/Youtube, somehow: I watched a single video on Youtube about Viking sword fighting and suddenly my people recommendations on Facebook were at least 20% militaria fans, always visible from the profile picture already. It finally stopped but it took well over half a year.

That's why I thought this story was about recommendation systems recommending either only narrowly what you already know, and finding something new is not really well supported or not better than random, or that one outlier can skew your recommendations for a long time. Worst is there is no way to tell the system "stop recommending me this kind of stuff" at least for the second problem, no manual way to make adjustments for the user.

> Happens to me a lot on Youtube too - you watch just one thing - and now your recommendations are full of things like that one thing.

Spotify's recommendation system is much better than youtube's (at least for me). I frequently "joke" that it is one of the few applications of ML that actively make my life as a consumer better.

Having been following Spotify's "Discover Weekly" for several years now, I'm actually really impressed how it manages to blend my long-term taste with recent moods. If I've been listening to one type of music for 1 or 2 weeks, there will be a noticeable uptick of it in the recommendations, while still mixing in less recent tastes.

As a father of a young child, Spotify's algo's are completely useless to me. If their ML is so smart it should be able to determine that Row row row your boat doesn't mix well with Anthrax. Wish I'd be able to toggle 'don't recommend kids music' somewhere.

I used Spotify to play music at a children's party. I am still getting recommendations for "Happy Birthday" months later. According to Spotify, it goes very well with Sheryl Crow and R.E.M.

Almost every monday when the "discover weekly" is rebuilt i get new songs which are not in english.

I have no non-english songs in any of my playlists and when they show up on discover weekly i flag them.

I would think that recommending me Spanish/Russian music when my entire collection is english should flag a problem with the system?

Yet week after week at least one song on discover weekly is in some foreign language.

I would argue for the vast majority of people language is not the primary indicator of what they consider good or bad music. I would assume it doesn't go into the algorithm at all and I certainly would not want it too. I have discovered some really cool music that I'm not even sure what language they are in this way.

I would wish you great success trying to frame an argument that the "vast majority" of people are not primary concerned with language.

It is a bit hard to listen to/enjoy music when you cant understand a single word.

Sure they had "gangnam style" but that was really a more a "one off".

> I certainly would not want it too.

I'd love to know if the majority feel this way.

given there are countless articles such as this: https://community.spotify.com/t5/Desktop-Windows/How-to-get-...

i would think most find this annoying?

Well i think the statistics are very strongly in my favour because English pop music is very popular all over the world and there are plenty of countries where the majority of people don't speak English (and still listen to English music).

When I was young we (in all the kids my age) listened to plenty of of English music even though we didn't understand anything. And let's not even talk about the plenty of examples of songs in English where even native speakers don't understand the majority of the lyrics.

Are you American by chance? Feel free not to answer, but living in the US I was surprised to learn friends in other countries commonly listen to music they can’t understand (most commonly English, sometimes German).

Put things in perspective for me - we really live in a bubble here. I guess Spotify could help reinforce that if that’s what the market wants.

I think it's likely you're right, although I haven't looked for any hard numbers.

But anecdotally, I only speak English and I listen to lots of Finnish, Swedish, German, and Norwegian music.

I don't know if Spotify just started recommending this stuff based on the few metal bands I knew from those places, but it started recommending other genres too and I really like lots of it. Not sure how I'd find this music otherwise, actually.

Why would you only listen to English songs?

i would think it would be obvious, but here goes - because i only speak/understand English?

That's exactly why I enjoy listening to foreign pop music: I cannot hear how vapid the lyrics are.

But besides that, great music does not need to spell it out to convey meaning and emotion.

If the appreciation of a foreign song keeps growing, I will eventually look up a few translations. Their mixed interpretation usually only enhances an already enjoyable experience.

For me the lyrics of songs are mostly irrelevant and it doesn't matter if I understand it or not. I usually don't pay attention to to it anyway.

Can be helpful to turn off watch history during such sessions, or to remove the videos from your history afterwards. Agreed that there should be an automated solution, but you can avoid it today will a small amount of work.

Are we talking about the same Spotify? Because the only reference I found to removing songs from my recommendations is a thread from 2020 that says it is not possible.

https://community.spotify.com/t5/Accounts/How-do-I-delete-my...

You can start a private session next time you have a birthday party, to prevent even more birthday songs being recommended.

https://community.spotify.com/t5/FAQs/What-is-Private-Sessio...

It sounds like a nightmare. Much like my YouTube suggestions, which now are mostly my kids' planet videos, and my wife's workout ones.

I set up the browser session we use as a TV with profiles entirely to prevent any more of my spouse's yoga videos from leaking into my carefully curated feed of machine shop footage and construction videos.

Spotify does have an incognito mode to prevent this, but you have to carefully enable it beforehand. Why you can't just delete stuff from your history is baffling.

I agree. There are some playlists and songs I only listen to in the gym, and similar songs constantly float to the top of Spotify's recommendations for me due to gym being 3x week.

I did consider creating another "gym-only" user on our Family plan, but Spotify should really have a way to create named contexts for the one user. e.g. commute, gym, run, working from home, on a plane, etc.

Hell, Spotify should use GPS tagging on your plays and determine "oh look, these songs are only played here, and always skipped elsewhere, hmmmm."

But that is too complicated!

I actually watched someone give a presentation at spotify on exactly this when I worked there. This was probably about 5 years or so ago. I have no idea what happened to that project. Probably the biggest issue would be getting people to give you location permissions but they also talked about working around that. It's a big org though they kill off ideas all the time.

It seems easy enough to "sell" it to people listening. Location, speed based track selection.

Pandora does this with 'stations'.

There's a lot I like about Pandora and a lot I don't. The "stations" you mention can be pretty good. The problem I have is with the app itself. It's buggy and learning to use it well can be frustrating. A simple example would be when using the back arrow to return to where I was in the app. If I try to go back to a previous screen, it shrinks the app and when I resize it, Pandora restarts from scratch. This seems to take forever and it happens just about every time I use it. The only reason I haven't deleted it and switched to another app, is that I can't be bothered with all that while I'm working or driving (the only time I use it). Of course, there's other, more common issues like searching for a favourite song and only finding live sessions or worse, finding out they've stopped carrying a certain artist on my playlist.

Seems to be an unsolved problem to train the algorithms for recognizing different situations. One recommendation for all roles the user has. Though, thinking about, it's probably unsolvable as long as the interfaces remain simple and focused on satisfying only the one user, not the different roles, which would complicated the interfaces.

I think this is fairly well solved from a mathematics point of view (high-school level k-means clustering). The unsolved bit is simply how we get the Spotify et al. product managers to care.

All they need to do is filter out content with genre=kids from recs. They could create a separate kids recommendation item if people actually want that.

This wont happen whilst they have family and kids’ account types. Charge parents more for the premium option of not filling their carefully curated genre recommendations with nursery rhymes and Ed Sheeran.

It's fairly trivial to keep the recommendations consistent depending on the latest request, not just the logged user. Pandora does a good job at this. We have it on our Alexa and at dinner, we take turns with the kids requesting songs. If we stop requesting, it keeps playing an internally consistent series based on the latest song we asked for. So if it's the kids' choice, it's a never ending row of Kids Learning Tube :-)

Spotify is missing user profiles, in the Netflix fashion.

From a technical point yes, but profiles are too rough and hard to use for this. What I mean is more some way to automatically maintain a kind of sub-profile, for each different aspect of a user, but exposed user-friendly and effortless. Most people won't maintain a separate user profile just for recommendations, as it's too much work for too little benefit.

spotify is usually on people's personal device like a phone not on communal device at home like a TV.

Speak for yourself. I use Spotify on my TV at home. Multiple people use it.

rude response.

i said 'usually' .

I would consider Alexa a communal device. Spotify is heavily used there in households. Source: worked on Alexa.

It should be noted that in a Tesla, changing the Driver Profile (stores settings like seat and mirror positions) also changes the logged in Spotify account.

For young kids with no personal devices of their own, they would only get Spotify from their parents' devices.

People use it to play white noise for their babies which makes all the future song recommendations useless

Both Spotify and Apple Music algos have no idea about music mood vs time of the day vs activity. Like late evening before sleep being perhaps not a good moment for heavy metal. And running not a right time for slow classical music.

Which is weird in case of Apple Music because Apple knows exactly if I am sleeping, running or driving a car - just from reading my watch.

>Both Spotify and Apple Music algos have no idea about music mood vs time of the day vs activity.

You say that, but I attended a conference with Spotify- they are specifically working on that problem now.

If they're working on it then the point still stands that they don't currently have it as part of their product.

Time of day aware recommendations is something YouTube seems to have had for years. It always knows what to give me up top based on if I'm sitting down for dinner or lunch or if I'm looking for an audiobook for bed etc

Isn't that more just you thing though?

Why isn't late evening before sleep good for heavy metal? I listen to the same music I listen to all day before sleep if I had music on.

The music I listen too doesn't change no matter where I am or the time of the day.

What that really means is that they either:

1) aren't doing the sensor fusion we all think they are, out of inability to access the data.

2) the models (and therefore modelers) aren't good enough to use the data they have correctly.

Or 3) they run loads of A/B experiments to optimize engagement or some target metric and they've reached a local minima and are unable to escape it without a lot of political will or Product Managers willing to stick their necks out.

Google Play music had that feature. I really miss it.

THIS! People have been asking for different listening profiles for ages. But they keep ignoring those feature requests in their community websites.

Spotify has supported this for many years - it's called Spotify Premium Family[0]. You get 6 separate accounts for $16/mth. It is worth the price of admission just to keep my listening history / algo feed clean.

[0]: https://www.spotify.com/us/family/

You can turn off monitoring what you listen to for recommendations by launching a "Private session" in the app

That's a good tip though I'm not sure that there is an option for it in the car when I most often get a request for nursery rhymes, or more recently Disney songs.

  > Row row row your boat doesn't mix well with Anthrax

I'm not so sure about that. Have you heard Koяn's Shoots and Ladders?

https://www.youtube.com/watch?v=ZU2k-U2Ze0o

I'm trying to avoid this outcome by playing as many different genres other than "kids music" as possible to my baby. Probably won't know whether it worked for a few years but in the mean time I get to listen to real music.

That sounds like a great argument for easy profile switching, along with an option in your play history for "move this play history item to this other profile".

I keep hearing this, but it hasn't happened to me yet. I wonder if it's because my Spotify history without kids is ~10 years vs ~2 years with?

We listen to a lot of Disney music, etc in the living room on my account, but I've never had that type of thing show up in my Discover Weekly or any of the Daily Mix playlists.

I will be very displeased if/when it does happen though.... Discover Weekly is a major reason I've paid Spotify for so long.

Ha, my life exactly. Ever since my kids used my Spotify account to listen to music my discover weekly list has unbearable amounts of paw patrol, baby shark and others mixed in. Now to be fair it still puts in music that fits my taste as well (and often my kids also like that music too), and considering that this is one profile I don't know what it could do better.

A lot of the time it's not an algorithm in charge. Most of the time now recommendations are based on who paid to promote their song or podcast.

Algorithms cannot be left running totally when sponsored ads can be purchased on the fly by content creators and musicians unless a site is lying to ad buyers about ad effectiveness.

I abandoned my account, partly for this reason. Too much kids music damaged it.

The easiest thing for them to do is simply exclude Kids music from their recs. Kids music doesn't benefit much from it anyway. Apple Music does this as do most video services.

Don't these end up in separate mixes, though? My 13 year old makes a lot of grime requests on road trips etc. and now I have a separate grime mix under my top mixes.

It does it relatively OK with the daily mixes, but daily drive and top of the year don't. That's why I don't really get it, they already know fairly well what kids music is so let me just toggle it off?

My brother who has young kids did the thing where you can make a combined playlist with me. He has indeed kids music littered all over in the shared playlist but moreover he also listened to a 100 track audio book on Spotify (which was not a great experience since it was published as an album and did not have bookmark support). The algorithm thought: oh he must REALLY like that album because he keeps on playing tracks from it, and so chapters from this book were also in our combined playlist

driving to grime sounds like hell, that's parental dedication right there

Can't be much worse than "Jenny" on loop for hours on end, heh.

They do, but recently my weekly discover is infused with kids songs. And it’s taken a few weeks ok ignoring those tracks to get them out.

I feel like this could be easily solved by adding an "incognito mode" to the Spotify app.

Using the Spotify kids app is a solution to this problem.

dunno, "Indians" could probably be made to work with "row, row, row your boat" lyrics.

Can't really complain if they offer Spotify Family which gives you 6 separate premium accounts, I think it's only double the cost of a normal Spotify account.

So the solution is paying double so I can log out of my own account and into an account I created for my 2 year old so we can listen to a disney song? And when I want to listen to my own songs again then I have to log out of my kids account and back into my own?

Yes, if you want multiple accounts you pay for multiple accounts. It's a solved problem you just have to pay the $9 a month.

I have two children, myself. I just installed second account on an old phone which has no SIM card in it, it has all the kids playlists and I don't have to worry about them getting recommended any of my 90s thrash metal. :)

Always those complaining customers ...

(and I am back to caring for my own mp3 collection again and use spotify only for rare and new stuff)

That unfortunately doesn't solve it. There will always be moments where kids will want to ask for songs on your Spotify, like you're renting a car and have your phone hooked up.

Click on your account dropdown in the top header, choose Private Session. Now nothing you listen to will affect your recommendations.

Come on, guys, I'd expect better from HN than complaining about features which already exist.

https://community.spotify.com/t5/FAQs/What-is-Private-Sessio...

> Spotify's recommendation system is much better than youtube's (at least for me). I frequently "joke" that it is one of the few applications of ML that actively make my life as a consumer better.

I've had the opposite experience with Spotify - I'd say my discovery of new music has withered to almost nil since switching when Google Music shifted to YouTube.

The algorithm just churns stuff I've already listened to, or suggests artists with (consistently) two songs that feel like Spotify's own fake artists - session musicians generated to capture royalties in-house.

I have the same problem with Spotify, just the same stuff over and over and over. YouTube Music has a much more diverse recommendation approach than Spotify does in my experience.

> Spotify's own fake artists - session musicians generated to capture royalties in-house.

Is this something you have any references for? It sounds super interesting, like shadow kitchens for music.

https://www.vice.com/en/article/8xabb3/spotifys-fake-ambient... - from 2017, they have been up to it for a while

True, but it makes for another good use: building a live playlist. Youtube would cram all random things with the couple of novel interesting clips, while Spotify will keep the mood until I decide to switch.

Spotify's algorithm works really well until you let someone else use the account. I wish there was a "child mode", "party mode" or whatever to disable updating recommendations.

Perhaps Private listening is what you're looking for, assuming it's working as advertised.

https://support.spotify.com/us/article/private-listening/

Pandora gives people granular access to the songs that are being used as the basis for recommendations. You would just delete the songs used by guests.

There is. Called "private session" in the settings.

Oh ... today I learned. Awesome.

It is cryptically called "hide activity" in my translation of Spotify.

It does. It's called "Private Session".

I turned off youtube recommendations years ago so I can't comment on that, but spotify's have pretty much always been terrible in my experience.

Anything suggested on the front page is either songs i already have in my liked songs or completely out of place. And the songs that i've already liked are from a few specific genres and artists, which seem completely random.

A few examples:

* Ratatat and Röyksopp both are in my top 5 of all time according to Spotify's own stats, yet I never got any suggestion about E.VAX, Kunzite, or Röyksopp's releases (their Lost Tapes playlist and their latest album, released about 4 months ago).

* the 'recommended for today' section is regularly filled with random synthwave when it's a genre i barely ever listen to, and a lot of 'electronica/trance/organica/deep house/whatever you want to call it', which i don't really listen to either (at least this kind of electro music).

* still, the worst offender has to be podcasts. I have zero interest in podcasts, have never clicked on any and likely never will, yet it's always the first thing on the app homepage, just below 'recently played'.

I've tried to use the browsing categories but these are most of the time just as poor, they only contain a few playlists and the latest big releases of the genre.

Highly subjective, but i'm tired of personalized suggestions & feeds. Just because I watched or listened to something does not mean I want more of it. And imho 'just use a no history session' or 'click not interested' do not solve the problem, especially since the argument in favor of recommendations seems to be that it is 'more convenient for the user'.

Why can't I just browse instead? Spotify has a very extensive way of categorizing songs based on multiple characteristics (https://developer.spotify.com/documentation/web-api/referenc...), so why can't I just use that when searching for new music, directly in the normal app?

All these recommendation engine problems people aren't discussing here aren't bugs or problems, they're features. They're either pushing the content that makes people engage with the platform the longest, or that promote another feature that the platform is trying to push on you. With YouTube, it's MrBeast style videos, influencer bait, and shorts. With Spotify, it's for sure podcasts.

Podcasts are so heavily pushed by Spotify because they're trying to make themselves the centralized one stop shop for your audio consumption. They didn't pay Joe Rogan millions for no reason.

As a man under thirty, I see the JRE logo on the front page of my Spotify at least five or six times a month. I have no interest in listening to podcasts on Spotify. I very much dislike Joe Rogan. I hit the option to not request it to me again, and yet there it is.

They're not recommendations based on your taste. They're a mix of just enough of your taste so that you trust and buy-in, mixed with whatever flavor of the week the recommendation engine would love to sell you.

> Spotify's recommendation system is much better than youtube's (at least for me). I frequently "joke" that it is one of the few applications of ML that actively make my life as a consumer better.

I hear this all the time but Spotify just plays stuff I've listened to. It's not a discovery service for me, it just plays the hits.

I'm wondering if that's just what it is and everyone likes it because it's playing stuff they already like for twelve tracks and then one new song.

> It's not a discovery service for me

Last.fm used to be good discovery service until it went downhill years ago, there should be services willing to take that niche

> Happens to me a lot on Youtube too - you watch just one thing - and now your recommendations are full of things like that one thing.

I've heard reports of this happening but for well over a year, YouTube was still recommending relatively high quality videos to me that went along with my interests.

Until last week.

Somehow, somewhere along the line I must have clicked on some sort of "influencer" video. Now all I see are hundreds of hyper Minecraft videos, "I filled my house with 1 million packing peanuts" videos, etc. I have tried manually searching for some of the topics that interest me again, but these influencer-type videos still overrule whatever I try to manually teach it.

I don't know what happened, but it has ruined it for me.

You can review your history on YouTube and remove videos (or at least tell it to not use them for recommendations). You may find one obvious culprit a few weeks ago.

I have the same problem. I don’t know what I did to tell YouTube I’m interested in Minecraft (never tried it, never watched someone try it).

I can vaguely see why YouTube is pushing right-wing conspiracy videos on me, but Minecraft? Mr. Beast? I’ve blocked perhaps two or three dozen of these channels. Take the hint, YouTube.

Something I noticed after I watched an episode of Stoic Finance[1] about a possible "impending crash of the Chinese economy" is that YouTube started recommending about a dozen "independent" channels.

I put the word independent in air quotes because they all have near-identical looking title cards with the similar fonts, some variant of the phrasing "The collapse has begun!", and the same-ish content. Different presenters, different channels, same message. Over and over. And over.

Reminds me of the "This is Extremely Dangerous to Our Democracy" clip that edited together dozens of apparently independent local news channels saying the same script, verbatim.[2]

You've got to wonder if there's someone with deep pockets trying to influence public opinion on China? Or is it just the same thing as the "local" news channels, where a bunch of YouTubers have been "bought" by a corporation that send them material to read?

I would love to know if anyone on Hacker News knows something about this kind of thing...

[1] In case you're interested: https://www.youtube.com/results?search_query=stoic+finance

...and some copy-cats clips I recommend opening in a private tab unless you want to be inundated with even more clones:

https://www.youtube.com/watch?v=6slQLbT_fNY

https://www.youtube.com/watch?v=0-tLenP5NA4

https://www.youtube.com/watch?v=iDVNag9Pq7s

https://www.youtube.com/watch?v=4JEdz1eA2vQ

[2] Entertaining and terrifying: https://www.youtube.com/watch?v=ZggCipbiHwE

The average Chinese person needs to save 100% of after-tax income for >30 years to have the down-payment for the average Chinese house. And, then, the interest payment alone would still be higher than the person's after-tax income.

Either China has to indefinitely grow at 10%+ per year (it's becoming obvious that's not going to happen), or the Yuan or Chinese home prices (~50% of Chinese wealth) are massively overvalued.

>You've got to wonder if there's someone with deep pockets trying to influence public opinion on China? Or is it just the same thing as the "local" news channels, where a bunch of YouTubers have been "bought" by a corporation that send them material to read?

These are the least likely options.

The most likely case is that the channels are all copying each other. Even if they were all put out by a couple content farms, it's extremely unlikely that the content farms care about anything else but making money. The only thing they likely care about is that people watch videos like those.

People trying to make easy money is, by far, more common on YouTube.

It's probably just plagiarism.

The Youtube algorithm bothers me a lot because they have the subscribe system but as soon as you do subscribe to a channel that channel basically disappears off the recommendations/front page.

I literally ignore the front page 99% of the time and just go to https://www.youtube.com/feed/subscriptions

This is also the only way I can make sense of twitter, skip the 'feed' or whatever it's called[0] and go right to my curated lists.

[0] as an aside, I only just made the connection that 'feed' is likely an abbreviation of 'newsfeed' but also brings a strong and stark connotation that I'm livestock to be fed

Funny, maybe it's a language thing (French speaking) but to me the first thing that comes to mind when speaking about feeds is "RSS", so I genuinely never made the connection.

It probably is: English has "eat" vs. "feed", German has "essen" vs. "fressen", but duckduckgo.com translates both to French "manger".

Interesting. I'd have translated "feed" to "nourrir", not "manger", but they're loosely synonyms.

Just opened youtube.com and 6/8 of the videos on the top section of my front page are from channels I'm subscribed to. That happens for me every time. Sorry your subscribe button doesn't work.

Spotify's recommendation system is so bad, that when I find a song I like, I put it into Pandora's free tier to find more songs. I hate how Spotify likes to prioritize every other song from that artist, and every other song from that genre, over songs from other artists and genres that sound similar to the song I want recommendations for. Pandora seems to prioritize the sound of the song, and actually finds similar songs.

Sites like Pandora, and many other "underdog" sites, operate as expected because they are trying desperately to please users in a bid to capture market share. Once they gain market dominance they begin to operate in an unexpected/unfavorable manner, just like Spotify does now, because "tailored" recommendations (strategic content recommendations) generates the most annual revenue for them.

The saying "power corrupts" applies to this circumstance perfectly.

Spotify manipulates recommendations to what makes them the most money, not necessarily what is best for user satisfaction. They do it just enough to prevent users from cancelling their subscriptions as well (usually).

> Spotify manipulates recommendations to what makes them the most money, not necessarily what is best for user satisfaction.

I was once a heavy user of Pandora, I am now a heavy user of Spotify, and I'm considering switching back. I think you're right.

But, having experienced both services at the peak of their quality, I think "stronglikedan" is correct, that Pandora has a greater ability (in addition to a greater desire) to analyze the way that music sounds.

I think Pandora's Music Genome Project is just superior. It always has been for me, and more importantly, it's always been consistently superior.

It reminds me of stealing a cool car in GTA and then every car you see is that same model.

The YouTube recommendation engine seems to overweigh tangent content. For example. If your recommendations are all puppy videos (awww) and you go watch Jonathan Blow talk about why programmers aren't productive anymore, you'll get a whole new feed of "How to learn Rust in 100 seconds" crap.

>Happens to me a lot on Youtube too - you watch just one thing - and now your recommendations are full of things like that one thing. Or Facebook/Youtube, somehow: I watched a single video on Youtube about Viking sword fighting and suddenly my people recommendations on Facebook were at least 20% militaria fans, always visible from the profile picture already. It finally stopped but it took well over half a year.

>That's why I thought this story was about recommendation systems recommending either only narrowly what you already know, and finding something new is not really well supported or not better than random, or that one outlier can skew your recommendations for a long time.

This kind of problem is exactly why we are building alternative YouTube recommendations. Search a channel to find more channels making similar content:

https://channelgalaxy.com

Click on icons in the lists to go to that channel's list. You can "surf" along that way to discover new content.

That's the worst, on top of youtube replaying the same track(s) you just listened to in the last 5-10 minutes in loop.

One can only handle so much Cotton Eye Joe and Black Betty...

It's not a broken recommendation system, it IS the recommendation system.

With respect to news, these systems are driving polarization in politics.

Is it money laundering in that "dirty" money pays for the bots that rack up the streams, which generates "clean" money paid from Spotify to the fake artist?

Yep, exactly

ok so the problem for the money launderer is that as spotify gets better at detecting bots and spam accounts they will have to pay increasing amounts of dirty money to get out clean money, at what point does it become a losing proposition to them, that is to say when this form of money laundering is less efficient than older forms?

A bigger problem would be that Spotify pays very little. And to have the bot "listen" to a song (to be eligible for payment) takes minutes of streaming, which turns to tons of time and a big bandwidth bill, since they'll need to do thousands of them in parallel...

This is where malware run bots come into picture ... it solves a lot of problem like not having to pay for bandwidth, evading detection by distributing access to the songs among varied IP and geography, etc.

They wouldn't have to actually play the songs, could just use the offline listening reporting endpoints.

it's obviously worthwhile for them now to do this, but sure it is an additional cost that at some point will make it not worth doing.

Like an airgap for money

What is the money efficiency of that?

It's not about efficiency, it's about avoiding getting caught for criminal acts. Since criminals don't pay taxes on their earnings and don't have to pay for all sorts of things others do, they're already more "efficient"

Sure, but if you're going through a distribution service that has a 60/40 split with Spotify, are you going to be okay with 40% of your laundered proceeds getting eaten by Spotify? How does that compare to other laundering opportunities? And that's not even counting the costs imposed by your distributer. For example, if you have 1,000 tracks, on TuneCore that's $10,000 dollars just for track distribution, as well as a $50 per-album yearly fee. It does seem somewhat inefficient, and very easy for Spotify to notice and crack down on. (But they don't have a lot of incentive to look too closely, since they're getting paid handsomely for it...)

Diversification, I guess? Having multiple concurrent laundering streams, so if one dies, the rest can keep going?

Interestingly, searching for Danni Richardson on Tidal yields the following apparently real artists:

Deanie Richardson

Danny Richardson

Dani Richardson

So, they are typo-squatting as part of the scheme.

I asssume money laundering would work by setting up Spotify accounts with stolen money (that they used to buy gift cards?), then streaming the fake songs.

Does spotify pay artists differently for paid vs free accounts? If not, then presumably they use free accounts for the bots, so it isn't really money laundering.

I guess if you get some streams from real people, then using paid bot accounts might still be a reasonably cash efficient form of money laundering.

Don't assume they're just targeting Spotify. Spotify is (or used to be) the best at catching algorithmic spam, the other streaming services hardly care at all.

Likely they use a front service such as TuneCore which gives them access to all the streaming services. Just using Spotify is unnecessary risky if the goal is simply money laundering.

(However, if the goal is to do pick up some royalties from all those background plays, and I think that's more common, then Spotify is clearly better)

For an example of spam/fraud, look up the album "Angry man" on Deezer. Or rather, the albums. There are 200+ of them, all with the same stock image album art. That spammer is easily recognizable, he's been doing a similar bulk upload several times per month for maybe 10 years now. He's an example of a spammer who is present on very many streaming services. Spotify has kicked him off, though.

The spammer is still at it, I see:

https://www.deezer.com/search/%22From%20the%20box%22/album

FWIW this is all I see, not logged in (never had an account), from AU: https://imgur.com/a/HXpn9JG

Not sure if I'm seeing something different to you...?

Yes. I see a screen full of your second to last result:

https://twitter.com/HaraldKml/status/1562137336409985025/pho...

Huh... this is pretty interesting. My screenshot shows "15 results" at the top, which a reload of the page just now is still showing.

But when I go to load the site through my VPS in Germany... I see the "292 albums" you're talking about! And yep, umpteen copies of the same album art just like in your screenshot.

Wow. I think this might be due to Australian consumer protection laws dictating that faulty products MUST be returnable. (Buying stuff in Australia is actually kind of cool because if I don't get exactly what I paid for I don't generally have to make very much effort to get a refund.) IIUC this policy has taken impressive chunks out of pretty big players.

...And so it would seem I'm simply not seeing anything I might reasonably and correctly regard as faulty.

Well that's an interesting database column then.

(I unfortunately have less familiarity than I probably should about this status quo (given I live here :D), but I can at least offer the "Consumers" and "Business" headings (and all the individual sections in the associated dropdowns) at the top of https://www.accc.gov.au/ as a cursory indication of the level of pedantry and protection involved.)

It's entirely possible I have the wrong end of the stick here and something else is the axis point. I wonder what would happen if you loaded the site through an AU proxy/VPN. I'm also curious if any other countries cause this... different presentation.

This is just conjecture, but I think it may have to do with compilation rights. The spam albums contain a random selection of music from the actual artists. Maybe the spammer has some very broad license to make compilations - I can totally see record companies accepting this sort of spam as long as they get their cut. But such rights would perhaps be geographically limited?

There are also problems with this theory, though:

* "Their cut" should, from my understanding of copyright law, be 100%. I don't understand how the spammer would make money off it at all.

* From the metadata, the spammer's one-off label claims to directly own both the music and the recording rights.

* The spammed artists do not seem limited to one of the big three's back catalog.

* The spammed artists include some which are/were well known for not allowing third party compilation albums with their music.

The criminal gångs in Sweden are using Spotify to launder drug money buy buying streams on artists in their circle who they also to some extent control. One of those were just murdered last year.

do you have a link or source for this story? (not sceptical, just curious)

Here is an article in the Swedish equivalent of the financial Times. https://www.di.se/digital/expert-streamingtjanster-anvands-f...

How do the scammers get the money if they use someone else's (stolen) identity and bank account?

Also, how do they succeed in keeping this scheme for more than a few days after the identity theft is noticed and reported to the police?

And why would they not just take the very music they just bought and paid for, and publish that as any other artist do? There's no crime in buying the rights from another artist -- that's exactly what all music labels do.

My take was that it's about getting _some_ of the dirty money out. So the dirty cash they are spending to fund this campaign is returning some % worth of Spotify commission. Even better if it somehow hooks the algorithm to recommend it to non-botted accounts.

> How do the scammers get the money if they use someone else's (stolen) identity and bank account?

Because if you have a stolen identity and open a bank account, as long as that account isn't trying to defraud the bank (e.g. write bad checks, commit ACH fraud), it can go on living for a long time before it's discovered and closed.

In this case, they opened up a bank account with stolen identity, and then used that to get their Spotify payouts, and then eventually transferred that to another account (cash, crypto, whatever). From that bank's perspective, unless they were specifically informed of the bad account, there wouldn't be a ton of unusual activity that would flag the account.

> Also, how do they succeed in keeping this scheme for more than a few days after the identity theft is noticed and reported to the police?

Why would a stolen identity not being used to steal from the person whose identity was stolen, but just to process legal payments, be noticed in a few days? Or even a few years? Maybe at some point the IRS might find the account that used my identity and send me an angry letter, but that seems like it would be years down the line.

Is the idea that you spend dirty money on the bots in order to earn clean money on the fake artist accounts?

And it’s cheaper.

I think I came across a YouTube video a while back that said even the platform was sponsoring fake artist accounts to recoup/claim streaming royalty payments.

https://www.youtube.com/watch?v=pCAPll9A5F8

After years of grinding to break ground on spotify and seeing the names they promote taking the lion share of revenue it becomes very apparent that it's a cash grab that just can't be trusted. There are literally dozens of other sites that do the exact same schemes and only pay out a fraction of the revenue they rake in for doing little to nothing for artist and music discovery.

The social Internet just seems more scammy than ever, and that's why I keep most of my music work offline until it gets worked into a project. Eventually there will be a correction hopefully, but the vast amount of free/unrewarded work that is being hustled out of people from these sites is not sustainable, and eventually it will turn into a content creator drought/strike if it goes unchecked too long.

You shouldn't expect to profit if your product is completely substitutable with another 0 cost product (generic background music)

That almost makes sense, except that the money is paid out to the artist, so if you're using a stolen bank account, it remains 'dirty' after the round-trip.

What you'd want to do is use the dirty money for the bots that drive up the revenue, and the clean account for the artist to collect the royalties. This way the artist has plausible deniability, and if the bank account paying for the bots gets busted, it doesn't matter, you can get a new one.

This is right - the whole point with money laundering is to have a legitimate business where extra money can come in, apparently from real but untraceable customers, that is really dirty money you control.

Isn't that what the parent described?

No, they suggested using stolen bank accounts to receive the artist royalties.

I think they meant "stolen identities and legitimate bank account opened by the stolen identity"

This. Same thing with Google Play apps and similar where you can get a clean payout and you can buy gift cards for cash.

Back in the days, it was done with SMS payments and "party lines" with per-minute payments.

What you are saying is that the "artist" whose account the cleaned money goes to has to be in on it? I.e. no identity theft or bank account hijacking?

I had my Spotify password compromised a while back and random fake generated music would play sometimes from my stereo. Very strange experience until I figured out what happened.

How is this money laundering? If you use dirty money to sign up to Spotify and listen to specific songs so that you get some of the money “clean” back wouldn’t that in case of Spotify be the dumbest and most inefficient way of laundering in the history of money laundering? I mean doesn’t Spotify keep the majority so you’re essentially donating dirty money into the stock price of Spotify for the most part, no?

Not necessarily. Let's say you get a burner phone and hook it up at home. You need internet and electricity of course.

Buy a Spotify gift card for $10 USD cash, use that to get Spotify Premium (first use the three months free of course).

Play your (artists') songs, 30 seconds each (considered one stream by Spotify). That's 2880 streams per hour, 86400 streams per month.

Spotify pays around $2-$4 USD per 1000 streams, so that's ~170-340 USD clean money coming in. Some of that goes to Amuse/TuneCore/etc. Then music production, admin, etc. But there seems to be quite some money to be made, especially once you manage to fake your way into recommendations and get legit streams as well.

Sorry, but wow you got this wrong. I’m not sure if it was just a guess or something, but look at the comment below for what it actually is. The money laundering is in the form of purchasing botted streams.

As another comment [0] suggests, this is common at least in Sweden. I have it on first hand that a successful local rapper has the whole apartment full of phones streaming his songs. Floor, shelves, sofa, all full of phones.

I'm sure mixing it with bot traffic is also popular. I also assume that Spotify (et al) are more successful in filtering out bot traffic from central Asia than legit phones in the middle of the target group.

[0] https://news.ycombinator.com/item?id=32560737

Who launders pocket money? Anything less than hundreds of thousands of dollars is just not worth laundering. Nobody is going to ask you where you got a few 10k from.

You pay for robots and stolen accounts, you got money per listens.

Spotify doesn’t really get your money at any point. The bots don’t need to be Spotify Premium.

Spotify supposedly pays around 70% of subscription and advertising revenue to labels and rights holders. Since payments are not pro rata per listener but are aggregated over all listens, if your fake customer streams more than the average you could quite possibly make an actual profit from this technique.

I wonder how much of the traditional music and theatre businesses has been money laundering. One could put out a record (which presumably has a nice margin on it) and pay folks a small amount of money to buy many many many copies of the records with cash, turning the dirty funds into clean ones. Likewise with a theatre production and ticket sales. Seems like any business with significant cash inputs would be vulnerable to this.

And of course it would make it very difficult for legitimate businesses to survive: they would lack the implicit subsidy that the illegitimate businesses get.

I used to work in the music industry and gaming the charts was common.

Art and publishing too

By theater, what type of theater are you meaning? Stage performances like plays? If you put on a crap play, the critics will lambast it for the sham it is, and then no more ticket sales.

This happens in almost every company that acts as intermediary between buyers and sellers, I.e., a marketplace.

Uber had (still has?) similar challenge where fake drivers were created, who took fake rides to liquidate stolen credit cards.

That...Makes a lot of sense, now that I think about it. I guess it wasn't really apparent to me before it came to my attention like this, but making online content and artificially boosting it, so you can blend your funds within what you've obtained from ad revenue is so simple, yet so brilliant.

Makes you wonder what other media out there might be laundering schemes, too...

There is a shitton of crap movies in Hollywood. I wonder...

Of course Hollywood accounting is already "This movie took $100 million at the box office, we spent $98 million on 'marketing', using our own internal marketing company.".

I guess making crap Hollywood movies loses you money rather than laundering it...

> There is a shitton of crap movies in Hollywood. I wonder...

The movie doesn't need to be crap to be a product of money laundering! The Wolf of Wall Street[1] was infamously funded by money laundering (from the Malaysian Sovereign Wealth Fund),

[1]:https://en.wikipedia.org/wiki/The_Wolf_of_Wall_Street_(2013_...

Wow that's a rabbit hole!

https://www.theedgemarkets.com/article/prosecution-tells-jur...

That site screams "malware delivery" or spam at the least.

> That site screams "malware delivery" or spam at the least.

I was intrigued since I recall the 1MDB scam, but didn't know all the details about it so I archived it [0].

0: https://archive.ph/HuEbS

And now as far as we can tell there's a cottage industry built around algorithmically-appealing (and probably algorithmically-generated) mood music. What a weird outcome.

There is a similar problem with the YouTube recsys on cheaply generated videos (e.g. a physics + 3d rendering model to generate colorful videos of Spiderman and Batman driving colliding cars).

See this older piece by Natasha Lomas on Tech Crunch:

I watched 1,000 hours of YouTube Kids’ content and this is what happened… http://tcrn.ch/2iPXpIA

Somewhat similar, I'm not really a tik-tok-fan, but after watching a LinusTechTips video with "KallMeKris" I was hooked on her content (one woman plays like 20 characters, yea yea judge me later :P ).

Anywhoo I was surprised to find close to 20 Youtube accounts that only has like stitched-together 1 hour videos of all her content. With non a none negligible view count.

Fascinating. It's so simple and makes so much sense, and replicable on any service that trades eyeballs (or ears in this case) for cash.

This thread has quietened down now, so I'd like to ask about those forums you were talking about. I'm really curious about them and would love to have a closer look, but I completely understand if you don't want to divulge that kind of information.

I don’t doubt that people are using Spotify to launder money, but I really don’t think this is particular method is a viable approach. It would be a much better business model if you could buy off iTunes or something, where are you know that most of the money is actually going to the recipient.

You have to wonder about the economics of this scheme. The music probably isn't expensive to produce and clone, but Spotify pays so little per stream, how is it viable to purchase bot streams in large quantities?

I worked in music distribution and maintain a very active interest in this space. This is actually fairly unlikely to work, from three different standpoints: operational logistics of how money moves around; technical logistics; and economics.

I’m not saying that it never works - and I’m not saying that this was not happening in the past (in fact, it almost certainly was) but the chances of it happening today are really pretty low.

Here’s why - and I’ll also talk about what is more likely to be the reason for these tracks.

1. Operational Logistics

Although streaming activity is reported in real time Spotify accounts usage to labels and distributors in the middle of the month following the month in question. The label/distributor then invoices and receives payment. Depending on the deal the label or distributor has with Spotify they will be paid on a defined cycle. This may well be NET 30 EOM - meaning that when the distributor raises the invoice it will be paid 30 days after the end of the month in which it was raised.

2. Technical logistics

Both Spotify and distributors identify and flag unusual streaming activity. If you bot 1 million fraudulent streams this activity may well be removed in between the streams happening and payment being made. If a distributor sees sudden spikes in activity they may well hold payment until they can identify whether or not the streams are genuine.

Streaming fraud is a hot topic - and it’s obviously fairly easy to identify “non-human” activity when you have end to end control of the streaming platform.

Spotify has historically significantly reduced play count on tracks they think are botted: today I’d be surprised if many botted streams are even counted against tracks.

3. Economics

Essentially this is an arbitrage situation - can you get enough streams to drive more revenue than the amount you are paying for the fake streams?

The answer is almost certainly no.

Let’s say you’ve got 1000 accounts that are going to be your bot farm. There’s a lot of work is going to have to go into getting those accounts to stream in a human-like way.

Because Spotify does not have a user centric royalty model (where the subscription payments of a user go proportionately to the artists they stream) there is no fixed “amount per stream” - only broad indicators.

Streams from premium accounts pay a far higher rate than streams from ad supported accounts.

Roughly speaking a million premium streams is going to bring you somewhere in the region of $4000-5000 dollars.

If your bot farm is 1000 premium accounts then that’s costing you $10,000 a month. You’re going to have to have a certain amount of hardware - or hardware emulation - as if the streams on a certain track are all coming from web, and from the same network, that’s much more likely to flag than streams distributed across web, Spotify desktop app and Spotify mobile.

Now, sure, you can pay someone who already has the bot farm - but there are then three risks: first of all, that they take your money and run; second, that their bot farm gets flagged and the streams don’t pay through; third that Spotify doesn’t pay out on the streams.

So assuming that you’ve set up your own bot farm investing in 500 cheap android phones and doing the rest with emulation there’s a lot of cost - at least $10,000 a month.

Can you drive 5m streams a month? Well, each node in your bot farm would need to stream the track 5000 times a month. Is that going to look strange? For sure.

There’s also the question of finding distributors who are going to pay out easily and quickly if you can even convert the fraudulent streams to cash.

What is actually happening here? Well, mood music - ambient music, ambient electronic - is a really important part of the “make music happen” sell of Spotify. A lot of people don’t want to hear the hits. They want music to play while they work or run or cook or have a shower. And for these kinds of “activity driven” or “event driven” music it’s important that the flow of a curated playlist works.

You might have ten great tracks by known artists, but the transition between those tracks doesn’t work. Spotify has (according to media) begun commissioning songwriters to write these “filler” or “transitional” tracks. There was a huge tabloid storm whipped up about this by a music journalist some time back who said he’d unveiled a whole stable of “fake artists”. Spotify wants people to keep listening to music. I suspect - but have no direct insight into this - that people who listen to a lot of ambient playlists are particularly valuable subscribers - they will almost certainly be paying (as you don’t want your flow interrupted by ads) and they will almost certainly be over indexing against usage. They are probably also less likely to churn because of the lock in of saved playlists etc. It’s also quite likely that a lot of these custom transitional tracks are not available on other platforms so you may not even be able to replicate a saved playlist.

So the interesting bit is that there are hundreds of thousands or millions of Bot accounts streaming music on Spotify? Wouldn't this have massive implications for their Ad Revenue model?

The scam part here is that listen traffic is bot-generated? Irrespective of that, on the payout side, why do you need stolen identities? It is Spotify paying out legitimately, isn't it?

See also Twitch streamers, Amazon listings with weird pricing, etc. Give someone poor dirty cash, let them set up accounts and buy/subscribe. You lose a percentage but you get clean money.

I'm probably missing something but how is this money laundering?

It costs money to run the streams but Spotify pays some of that back out to the "artists"; that money sounds plausibly pretty clean. Until there was a thread at the top of HN about it and everyone suddenly knew how it worked.

Of course GP might just be a GPT3 experiment...

What does it cost an artist to run a stream? I thought Spotify hosted the content. Do you mean it costs money to run the bots that listen to the stream and get the artist recommended and heard by more people?

There are services that offer bots that increases your listening count on Spotify. People with dirty money create a bunch of artists and songs on Spotify, and use the services to increase the listening count. Spotify pay out royalties based on those numbers. So they pay the service with dirty money, and get clean money from Spotify

Dirty in, Clean out = Laundering

You need some piece of hardware for the bot to do the streaming on so it sneaks past the spotify bot detectors, and you probably also want to pay for a premium account for each bot so that you get a bigger payout on the other side, since premium users generate bigger royalties.

Making it expensive to operate an effective bot farm is part of the way that spotify tries to discourage botting. Spotify's bot detectors are good enough that it's not lucrative to just make a bunch of bots to give yourself streams for the free royalty money, but apparently not yet quite so good that it's not feasible to use bots to turn a large amount of dirty money into a lesser but still substantial amount of clean royalty money by way of funding a bot farm.

Getting out, say, half of what you put in is a losing prospect for somebody who wants a money printer, but might be acceptable to a money launderer.

You need some kind of publisher/label to get on Spotify; an individual can’t do it themselves (unless that’s changed recently). There are labels that do this for a reasonable amount of money, Like $30-50/year or so.

Distrokid et al. Easy

Who knows, maybe the bots even have legitimate Spotify listener accounts.

(In my opinion) Some hacker provides a paid service in which you have either hacked spotify accounts or well generated fakes that can run streams on fake artists accounts. Some random mafia pays these hackers to maintain their bots and to launder their money. You can even pay the hackers with the finally laundered money.

Now the hacker has to launder the money somehow. Fine, it's much less money, but still it doesn't feel like laundering rather than just straight up scamming.

Doesn't that imply to have millions of Spotify accounts to be used by the bots? Is that still profitable?

Money laundering is profitable because the money put in is devalued dirty money, more like a raw material input.

Very likely that the "musicians" that create those "music" are actually AIs.

hmm would this apply to steam as well? people buying up weird fake cards and on sell for less but cleaner money?

Doesn't it leave a money trail?