Record Labels Just Betrayed Every Musician, Dead or Alive

Comments · 31

• UN

  @UnderCurrentStudio · 13 days ago · pinned

  In 2006, the labels used Spotify to save their empires. In 2026, they're doing it again with AI. Do independent artists even stand a chance at survival over the next five years? What's your predictions?

  1

• EC

  @Econandcohorts · 2 weeks ago

  Here&apos;s the issue with this video it&apos;s using the lawsuit as proof it isn&apos;t proof of anything, since courts today do not typically turn away high dollar lawsuits.<br>Here the other issue, if the evidence in the lawsuits was factually you would be seeing YouTube and other platforms also suing and you would also see state and federal charges being brought against Suno. Since none of that is happing, it the label companies trying to either force a deal or bankrupt the company through lawfare. Same trick anti firearms groups try with firearm manufactures.<br><br>Also 3 federal courts have ruled that Suno and other AI companies have done nothing illegal, that training on peoples works is legal unless they hack firewalls, paywalls, etc. then at that point it becomes theft.<br><br>Then there&apos;s the fact that most artist do not have any rights to their music when they sign a recording contract with a label company and if you are a indie artist and you are putting your work on un-firewalled/un-paywalled platforms like YouTube and yes YouTube is a none firewalled/non paywalled platform, your works can be used to train computers, people extra without payments do you.

  7

• ST

  @stefanmanninen · 2 weeks ago

  You won&apos;t find an industry more resistant to progress than the music industry. Embrace the new technology. 😑

  3

• RA

  @radicallydyslexic · 2 weeks ago

  thanks for talking about this!!!

• MA

  @martinrevell1143 · 10 hours ago (edited)

  if you can ONLY type in lyrics and prompts without ever singing a note or playing a note, hell tapping your leg or a desk with a pen to get a beat- &nbsp;even if the data is trained off studio players and they get PAID to give the model the training data, it is bankrupt of all merit from then on. The only good thing is the session players will get paid. I say this as a 1 year user of suno, where i always input some or not all of a full or demo track, without lyrics as i make instrumentals. and soon you may be able to do just that, this produced music shouldn&apos;t be monetised period, only for the users enjoyment. and for the record i have no sympathy for the major industry, i do of smaller labels and independents.

• ES

  @eshamo2006 · 2 weeks ago (edited)

  I&apos;m looking for someone to run up in Suno&apos;s and pull a Luigi move. There&apos;s no way these Record Executives and Music producers were ripping off those artists the way they were. It&apos;s sickening and the music artists need more laws to protect them.

• WA

  @wayneclarke2890 · 2 weeks ago

  We&apos;re talking about an artist&apos;s skill not work. Labels have access to skills now which they never had before. It all seems like it was a well thought out plan. Robert Johnson would be turning in his grave. I&apos;m glad you picked up on the situation and how crafty these companies really are.

• SA

  @SAMSYANKS1 · 22 hours ago

  Embracing modern technology democratizes the entire creative process. Historically, transforming original lyrics or poetry into fully produced arrangements required substantial capital—often running into thousands of dollars in studio, session musician, and production fees. Platforms like Suno have fundamentally lowered these barriers to entry, enabling independent writers and creators to realize their artistic visions efficiently and cost-effectively without relying on cost-prohibitive legacy production models. Innovation expands access, and those who adapt to these advancements will be best positioned to thrive.

• UD

  @udiorockmeamadeus · 2 weeks ago (edited)

  Phil Lynot of Thin Lizzy. They didnt have many Irish Songs, they did Whisky in the Jaro just to get the support of dublin cause the lead singer is black.. Its covered in the documentary about Phil.. <a href="https://www.youtube.com/watch?v=HtJAzZKxRBc">https://youtu.be/HtJAzZKxRBc?is=2U3XWVG-OrX0pc0V</a><br>Huey &nbsp;LEwis and Midge Ure were in Thin Lizzy, Midge Ure of Ultravox and one of two producers who mixed the Band-aid &nbsp;&quot;Feed the World&quot; song.. &quot;LEt them know its christmas this year&quot;.. <br><br><br>This is my favorite, I love Robert Palmer&apos;s early stuff even though it wasn&apos;t best selling material, so I prompted, extended and cropped/extended this until it sounded like him then I &nbsp;extended it, and I could tell there was an environment to it, that I didn&apos;t ask for, so I extended it when it sounded environmental.. I was hoping the laughing happened &nbsp; and off mic stuff happened where the amplification was coming from but I figured , it happens in real life that the singers are in different spots than the speaker set up, but if that was true I&apos;d hear him singing in the background somewhere.. Thats why I believed it was in a small club, &nbsp;cause there was &nbsp;no echo due to propagation delay on anything but the drums.. Of course this is Model 1.0 which sounds like crap.. I think this video here says they trained it on youtube recordings, That &nbsp;makes sense.. They told us it wasn&apos;t based on mp3s.. Note, Youtube could argue that the music provided there was in fairuse cause the youtube video is a framed context, like recording a movie and selling it on the streets of manhattan, that old men were doing on the sidewalk in all places but wall street near the sculpture of the bull.. I once had a job in wall street for a short time.. It was a tech company who bought up a competitor and we trying to merge it into their current code base.. Whie I was working on it a manager was having an argument with a chinese developer discussing why he should comment his code, which is a big nono, I mean failing to comment code.. Obviously the Chinese guy was trying to make himself a legacy worker like one they already had who was sitting ina &nbsp;cubicle calling people and letting them know they had hackers in their network. While one of our artists was complaining that a guy from the business team was hitting on him.. I was upset cause the production tester wasn&apos;t permitted to tell us where in the code our source was &nbsp;failing.. The environment made me sick, and I was going crazy cause I had an irritation in my throat, that turned out to be acid reflux.. And I left the place early, the managers were discouraging me, cause I was leaving too early and they had not made back what they paid for the &nbsp;jobe search people &nbsp;and emplyment insurance.. So they lost money, but they had someone ready to replace me the same hour I left.. So glad im out of that crap.. &nbsp; &nbsp;<a href="https://www.youtube.com/watch?v=O49rpFhQWGA">https://youtu.be/O49rpFhQWGA?is=uzu4ASg6aPsP-B7n</a> &nbsp; <br>This process on udio is per segment, unlike suno, udio will not let you generate a song longer than 2 minutes, so anything longer than 2 minutes required extensions of 32 seconds per segment.. Even if you upload a song for extension, you can only extend it by 32 seconds. That means you write a prompt and some lyrics, and render out bunches of 32 second clips to extend the song, but you can only pick one.. And each segment of the song is like that. The longer the song, the more statistically imorobable it is. So for a song 15 minutes long, I can generate 2 minutes max, then I have to extend with 26 segments. For each segment I can generate any number if clips, but I can only use one.. So if I generate 10 clips per segment, and I listen to ten and pick one, I will need to listen to 260 clips and pick 26, and those are random generations, if the seed used is different every time, then only I had those choices, nobody else will have had the same choices..And if I liked a clip except for the last 3 seconds, I can&apos;t simply cut off the last 3 seconds after I extend it, I have to use crop/extend option and adjust it so the duration of the song is 3 seconds short and can add upto 32 seconds to the beginning of the song, cause you can only chop from the unextending end.. And after that I can extend again on the end that I chopped from. I&apos;ve extended songs completely from the beginning, if I like a clip but I want it to be an ending, I will from the back end &nbsp;and keep going that way until I hut the total duration. I can only ever influence the ai witha prompt or the neighboring chunk of audio, upto two minutes. This means songs over two minutes cannot borrow anything it us more than 2 minutes away from. Forcing the song to change. SUNO &nbsp;can&apos;t extend from the beginning, it even has a problem with songs tgat start monotone, it will starts the lyrics on the monotone lead, example is ASIA&apos;s ONLY TIME WILL TELL, the opening melody, lyrics will begin there.. And if you have a loud define melody it has a tendency to foreshadow it, so you cant surprise the audience &nbsp;when the singing begins.. A lot of people acting out against AI music expose the fact they don&apos;t know how it works, cause they onlyknow it uses a prompt, then they assume a lot of other stuff, but these are also programs whose interfaces are designed in a way to offer limited control, cause that will compel people to be creative. Its the limitations tgat create the sport. So the best artists if AI are those who have mastered all the idiosyncracies of the interface and methods of leading the AI and limiting its awareness of the material, cause the AI is stubborn in that it only chooses what is statistically probable, and to get originality, you have to nudge it, like english on a pinball machine. &nbsp;Tgats why the head of suno was claiming it could become a videogame, is that its an art of &nbsp;choice.. &nbsp;I think over time people will realize everything said about it, all the crap they were told was wrong, but even then they will want the lie to be true, cause so much of their identity and pride will be invested in it.<br><br>My Betty Boop..<br><a href="https://www.youtube.com/watch?v=C-f15T-KfmM">https://youtu.be/C-f15T-KfmM?is=GoCPAOoS9yjLvefA</a><br><br>SO PERIOD PIECES &nbsp;are what udio is for, not covers like suno.. Suno can do covers, udio &nbsp;can&apos;t its only good for deepfakes and original works but you can&apos;t ask for celebrities, there are so many tricks in getting it &nbsp;to work, one is you have to talk to it like its alive and it can reason, and you have to praise it when something is a success even asking it to try something seems to work.. I actually get the feeling that there is a ghost in the machine, you prompt it segment by segment, you will even get renders tgat sound spooky.. I mean really eerie.. likes its a line to a underworld, possibly hell, who knows.. Dead musicians, and they sound thankful.. Notice the subject matter of the lyrics, its like the twilight zone.. <br><a href="https://www.youtube.com/watch?v=C-f15T-KfmM">https://youtu.be/C-f15T-KfmM?is=GoCPAOoS9yjLvefA</a>

  3

• UD

  @udiorockmeamadeus · 2 weeks ago

  its actually impossible to tell the sources of the music unless the information was tagged when it went in. I have 57- songs , 290 are from udio.. &nbsp;I have a backup before the new terms of service, of UDIO, those a registered as mine legally. It is like 243 gigs in size, most of it is replications of the same track but one 32 second extension for every generation.. If a song made was 15 minutes (macimum length of an udio track) long and I used 10 generations per segment.. &nbsp;Well it gets to be a lot, but I can show the complexity that was required to make any of those tunes.. In some cases I changed the prompt on every segment.. There are a lot of tricks on how to get good generations, but you can&apos;t simply make any of my songs just typing up one prompt. Its a bit more akin to what a producer does.. I tell it what it can and cant do, I give it lyrics, specify the genre , I control whether it can add more variety from the prompt or if it can only create content like it already produced , I can tell it how many seconds of the previous two minutes of the generation it can use to make this next segment.. I can tell it how close to the lyrics it can use, If I set that to zero and singing is permitted, it can&apos;t sing anything meaningful, it sounds like scat.. The more complex it gets, the more control, the harder it is to know precisely what went to the production. I have songs that sound like sting and donald fagen, I have a song that sings with both the vocal styles and tones of peter gabriel and phil collins. And Ive found tgat if I know who Im hearing in a generation, I can steer the AI to who I want precisely by being careful with each segment I extend by.. So if I&apos;m intent on getting Boz Scaggs, I just select the segments that sound the most like boz scaggs and looking at the context window, the AI will assess the best singer for the job is boz scaggs.. I can then cut off that segment and keep it for later in case I want to make a song with boz scaggs voice. I have a segment in my archive where I caught olivia newton johns voice, but UDIO&apos;s moderation wouldn&apos;t let me extend it after a minute, cause it was too close to her style.. The songs I made with udio, if they sound like someone, it wasn&apos;t enough for the moderation AI to recognize it as an artist voice.. The one of mine tgat has a lot of likes, sounds distinctly like Micgael Jackson, it even has a point in the song when he says &quot;I believe in forever, even if it is a dream, &nbsp;let love be what it means&quot;.. <br>I don&apos;t write lyrics, I eitherhand it over to an LLM or let the engine generate.. The spooky thing about this us they sound so real, its like the AI captured their soul, and if they are dead, I imagine they can only be brought back to life through the AI.. But I think its that we know so much about these people that in oneform or another, we are recording something more than is in the recordings, and an AI could be trained completely on the segment choices made.. It needn&apos;t know how music sounds, it only needs to know if we like it, to replicate it.

  2

• JA

  @JAC-hf8yn · 2 weeks ago

  WMG took around 4-6% share in Suno at settlement...forgo the NDA...

• DO

  @doommega · 2 weeks ago

  No One wants to talk about this... this vid has it! no one wants to belive the record companies are the biggest villians in the room not Suno.. suno is the final result the activity been happening post-pandemic, and we don&apos;t wanna admit it, because it would mean we slept!

  2