🔴 We're going to expose Udio tomorrow 👀
Вставка
- Опубліковано 5 чер 2024
- Tomorrow I will release a video showcasing how Udio is not only possibly training their model on Copyrighted music, but also is able to generate exact match sound-alikes of the music it might have been trained upon.
Hey Jesse, another excellent video. This has to be fully discussed from every angle if we have any hope of avoiding disaster.
My problem with using it like a "really smart friend" you chat with for information, is the massive potential to brainwash people and centralize "truth" or codify "information" down into one (indisputable) source. Then it gets to who is training the institution, and what are their values and motives?
Here's an example.... My anecdotal truth about health and healthcare is that:
- I will be healthy and my body will fight infection via my own immune system if I eat organic foods to reduce toxins
- eliminate gluten because it causes inflammation
- eliminate dairy because it provides more harm than good in our physiology
- exercise keeps my body strong and flexible
- a lean bodyweight is optimal for healthy systemic functions.
However, Institutions do this
- The medical establishment allows toxin laden foods that are FDA approved on our shelves
- The Medical Establishment promotes the "Food Pyramid"
- Big Pharma promotes a drug for every symptom and side effect, but ignores root cause of illness (and in many cases causes them).
- Political institutions are working to normalize being overweight and normalizing fat being unlinked from poor health.
- etc....
My question then is, who owns the truth (Truth = trainers of the AI data)? Will we be permitted to debate and dispute anymore once it's all codified and delivered to you on your phone. Recent world events seem to point to a hard no on that. Who gets to decide the morals and values surrounding these issues, and then how people are permitted to live their lives?
I don't mean to be dark, but I see so much rot and corruption in institutions. How will this institution (AI Information center) avoid the rot and corruption? I guess I'll leave it there, but I could go on about this topic.
I truly want to believe that people in control will realize the threat to humanity and design it in such a way that it will only provide incredible good for people. I'm just not seeing many institutions, and certainly not the biggest ones, that demonstrate that ability.
All good questions and thankfully there are many open source models that will compete in the marketplace to allow us to chose which AI models we want to engage with.
Very true. People don't know that ChatGPT supports the rights of Drag queen 'expressing themselves' to children under the age of 7 and the rights of Adult babies...
Wow this thread went off the rails quickly. I'm glad Jesse kept it diplomatic. 😸
I actually thought about starting a youtube/twitter/etc channel dedicated to making explicitly copyright-protect copies of famous artists and songs, and tagging those artists when I do, with the explicit design to get major labels upset. The problem is eventually people will have their own personal AI they can make so anything they want and millions of people will have their own Udio.
Do it!
We can all agree and feel 1000% confident about one thing. AI Music is not going anywhere.
My gracious, Sir Jesse, that snippet of the track you're writing now is FANTASTIC! As the chorus was playing, I found myself harmonizing with it, it's SO enticing to BOTH the ear & the heart, couldn't help it, singing is actually more my strong suit than trying to compose stuff of my own, at which I'm currently toiling away. Keep going, you & your chosen vocalist fit like a hand in a glove! Beautiful work, respect. 🙏🏽
Thanks so much man!
You should also try with Queen! It sounds ridiculous similar when I prompt Bohemian Rhapsody
In the end I deduce they would be shaking hands or bought by big labels .
Big boys who owned a long list of copyrighted catalogue materials. And repurpose in wicked way those materials
I'm my opinion ai music was being trained by others copywrite material but now with millions using it soon. All future music will be trained on non copywritten music created by the ai generated music.
inbred beats 😂
@@Pharrellsburner Yep 🫣
I have a few examples of those soundalike prompts that came up, especially ones in the epic orchestral genre, that sounded exactly like 2 Steps From Hell or some other recognizable cues from trailers and such.
Thank you, you are awesome
definitely quacks like a duck
12:05 🙌🙌🙌
LOL holy cow - the hilarious irony of you picking ChatGPT's brain about how AI will impact production music. it's great, and scary and hilarious, and mind-blowing --- wow, this is really going down in MY lifetime.... damn
A handy little check-box on the metadata (page/sheet) when you upload a work to a library would be, "Do you approve of this work serving as training material for an AI model? YES. NO". Might be a good preliminary way of saying, "You do not automatically have access to everything," to the AI companies.
Udio is fun to play with as an end user. But it is obvious to anyone that has spent any time prompting the model that they have trained on mainstream music. As a country music fan, I have been trying to see how many country artists I can get it to copy, so far I have a Duet with Miranda Lambert and George Strait, a Diamond Rio and a Keith Whitley song published on my page. Also got snippets of songs with the voices of Tom Petty, Sammy Kershaw and Lee Ann Womack. I have also Liked several other users songs, One is straight Dierks Bentley and another user generated a Weird Al Yankovic song. It's going to be interesting to see how it plays out.
Out of my own "The Evolution Of Life" ones I think I got these likenesses although not sure how many are the blatant real voice and not only likeness.
Indian Vocal Ones: Arijit Singh
Evolution Chart (My Evolution Time): Buddy Holly & The Crickets, Bobby Darrin ect era artist.?
Chart Clash Rhapsody: Iron Maiden?
Global Rhythms Crown: Janet Jackson
Chart Showdown - Michael Jackson
Frequency Ascent: Depeche Mode (Early 80s)
Rhythm's Crown: Depeche Mode (Dark Era Late 80s)
Chart Show Coronation: Slade
Chart The Hits NOW!!!! Pt's 1, 2, & 3: The most voices you can get in a tune 30+ artists in all C3PO is even in the 3rd part.
Rock 'n' Roll Rebels - Sweet?
Mixtapes Devotion (Synthpop's Genesis) - I know exactly who these are but forgot the name of them well known 80s group. May be both a melody and voice generation by them as well.
Disco Train Rhythm's - Another I know but can't put my head on the name of them.
Cosmic Dancer - Same as above.
Sheher Ka Jashan - CAP's?
Queen of the Disco - Michael Jackson, Five Star, Janet Jackson
Glam Rock Soldiers - Slade
Glam Rock Soldiers - David Bowie & Billy Idol
Bansuri Visions: Lekha Rathnakumar - Sonoton, Weststar, Royalty Free & Custom StarPlus Music From A Central so called Private Server in one track but if Udio can rip it for a Jhanak BGM melody it's not private for AI only humans (My Track Fully Though As It Was A Fully Custom Routine Experiment with terms of music written to change the sound up fully like the old AMIGA, Fairlight and ST original DAWs days where it was called a tracker I wanted to see if I can write a full customized routine making this fully my music as opposed to Udio music with the same melody the entire track as a rebellion thing to show off it may train off peoples work but I can make my own too outside of the prompt I wanted it to do first)
Distorted Horizon's - Production Music Rock with Mainstream Prog Melodies (Same as above)
Solo Synthesis (Experimental 2024 New EDM Sound) - (Same as the last 2 making it without much prompting more experimenting on custom routines like a real production artist would do if making a track in AI to make it completely there's but in a genre I haven't been so fond of since the 90s as well)
My last 3 so far I think might be my favorite outcomes other than Chart The Hits NOW!!!! which is so bad at taking voices I loved that auto routine one too. As swerving the generation with fully custom routines it's like going back to the commodore amiga days way of picking an electronic tracker instrument and having to write a length code to it and a pitch code hex as well to make your own not AI's generated track anymore but it's fully already in there no more hexes just write the instrument or the style of music or melody ect. of the part and away you go. Im probably the first ever AI pusher to really do the custom thing proper like an AMIGA DAW in it.
I couldn't ever figure out the hex thing having to put length and tone in hex after choosing your instrument from the program despite always having an AMIGA around even now but in this thing I can do what I always wanted in that mess around with proper routine real music.
It's like I'm swerving the AI to be the anti AI my personal musician who isn't AI at all with me making the music fully myself after the prompt for the style. I like my new way of experimenting with this tech putting all those real musicians music therapy autism helping my anxiety lessons into fully practice allowing me to make music for the first time. It was the coding which let me down on the AMIGA I never could get along with the interface of the hex and that but AMIGA was the last days of the EGM genre I really loved as it was the end of half midi half chip real electronic music until now I guess as I can probably write EDM 90s or Dance 90s and get AMIGA at a touch of a prompt.
But that also puts this debate into my head how may royalties are being payed off for the training do they pay a public broadcast PRS or Cocatalog licence and s their eone thats ay extra for AI company owners yet as the whole world is using it now so it's broadcast media so they should be paying a broadcast licence at the minimum.
Great stuff as usual, man. Let me ask you this: did you try having a conversation with this new Chat GPT thing about stuff that you do know about? I haven´t played with this yet, but in my experience playing with regular text Chat GPT ALWAYS when I talk about stuff that I know about it gives me a whole lot of wrong information, specially when it gets to more specific subjects but in general too, it doesn´t have to be that specific, so I would assume the same thing would happen with this new version that came out, so I think we really really need to be careful with that stuff.
Yes I asked it about Sync Licensing and it got most of it pretty accurate.
Logic has been using “AI” with “drummer” for years. The new bassist and keyboardist is the same tech.
It's different because they didn't train on copyrighted music as far as I know, definitely not for the drummer and it's not really AI anyway.
@@RubenMillsOfficialthey didn’t train on existing music for the new bassist and keyboardist, none of it including drummer is AI it’s just randomization of existing samples.
Goodbye universities!
Weaver Beats exposed UDIO in his “UDIO Steals Music” video.
We’re taking one step further tomorrow…
Let's see Ai music as a cruch for some emptyness moments. Let AI compose, inspire me, use it as template, do cover of that AI generated song making all instruments (vocal if it's necessary) and then deleted the original AI track. What left? All the project tracks by myself with all efforts, challenges and prouds of achievements.
What i think is happening is udio is hiring musicians and singers to create "type" music. They are making The Beatles type music. Mariah Carry type music. When you put a txt promt of "i need a vintage 60s band sound, the system goes into motown "type" Beatles "type" sounds that the hired musicians did in the spirit of. So it sounds like something like it made Beatles music out of thin air....
lol just wait until you see the video I release today...
That track you played sounds like boy band music from the '90s such as NSYNC, Backstreet Boys or 98 Degrees!
Is this supposed to be an original piece by you?
It sounds like something I've already heard before.
It's funny, as I was writing it I def reminded me of 90's boyband stuff too! But yes that is a song I recently wrote.
Other humans also makes better tracks than most of us. Udio is faster but until there is a 100% possibility the rearrange/remix/change everything inside an udio track it is still a "one shot" tool. The biggest loss might probably be what "the hit" singers will take from this technology, degrading/saturating the singing efforts. The beat behind a singer can be accepted with lower quality (Lo-fi etc) but a singer need to be top notch in performance, and if AI can replace that it raises the bar singers need to be able to jump over. Who wants to watch eg Idol (for the quality aspect of singing) if AI generates better results (pitch, timing, feel, range etc)? We need new genres AI Pop, AI Rock, AI Soul etc to be able to separate AI prompted music from human created.
I listened to one of the featured songs on UDIO yesterday and it was a complete ripoff of Fleetwood Mac.
I quickly lost interest in AI music after generating a couple joke songs for the family, maybe because I know how record and produce my own music for real... Not to mention the audio quality is still garbage on a decent pair of studio monitors/headphones.
So it's okay to take jobs away from psychiatrists and counselors, but not musicians?
Let’s be honest if people make money w sync libraries from submitting background sampled 808-beats for a Kardashian episode is that “song writing” to be protected? no. And Netflix could easily hire monkeys in-house to prompt Udio for their briefs. This business model was waiting to be disrupted. And we’re all upset now that Taylor Swift or Drake are getting ripped off by AI? Are there not countless human artists signed by labels precisely because they are clones of Taylor or Drake? The entire business is derivative thanks to humans lack of creativity - so perhaps the music business and Netflix need LESS human clones and more AI clones. The actual, talented human songwriters will make a fortune and will stand out among the AI tracks. If the “beat makers” cannot distinguish themselves among the AI beats then go make $ doing something else. There’s zero talent to program a beat compared to Udio - the ultimate 808-machine 😂
Like you said, heavily biased, but I've yet to hear a good argument as to why AI replacing human creativity is a good idea, outside of a bunch of people get to make a profit off of it.
Well obviously we all know that people will still make music and won't get totally replaced but outside of that AI is literally just a tool just like a DAW who "replaced" a bunch of musicians. Just like AutoTune was intended to just be a tool to fix certain parts of a singing performance well AI music is the same i mean think about it nobody ever said "Hey go generate a full song with AI to compete with artists" literally people started using it as such but just like a tool you use it to maybe help you with a song idea but nobody is stopping you from making full songs with it just like nobody is stopping you from putting AutoTune on a full vocal take instead of a little bit of AutoTune on a part you didn't like... I rest my case it's just a tool... People will get very creative and inventive with it instead of the generic music people are making now...
@@Songinatorr XD any other tool is not trained on data from other people. Comparing ai as a "tool" is inappropriate
That's always down to what people will do with it . TBH Ideas never enough, because someone can say training machines meant to preserve the old music spirits if someday our culture forgets about those things.
But what people do with it might be in the opposite direction from those ideas.
That's why Fair Competition Law between Machine VS Humans should be prioritized .
It's not worth our time and attention. "Whatever one focuses on, GROWS".
Blockbuster shouldn't have paid more attention to Netflix?
Due to the emotional toll that this tech is having by challenging the subjective perceived continuity of people's livehood ,we can see reactions like yours on repeat loops ,first or most notorious on the arts were the visual arts ,now music and next is video ,the first two are a done deal ,video is halfway done ...this is reality evolution is ruthless those that do not adapt simply die ,no music industry is large or powerful enough to stop AI,you can't sue evolution out of existence
😐 You again. Did you even watch the video? Jesse is trying to get us (and himself) through a painful transition.
Unlike you, frolicking from channel to channel, pasting prompted propaganda, or demonstrating your robo-fellatio skills.