Rick Beato on how easy it is to create AI music

One of my first thoughts when seeing this is, for those folks that make a living from writing jingles for advertising etc are all now out of a job :(
 
I'm surprised how easy it is with very little information.

Suno produces a huge range of output from stuff which is basically rehashing other artists work through to stuff which is more or less original content with the 4.5+ model and a big range in quality - some genres it really struggles especially ones like trance where it has trained a lot on tunes which use a lot of low quality samples other genres it is producing close to studio level now.

In terms of what Rick is talking about I wonder where you draw the line - I made a bunch of music years ago by dragging and dropping samples in various pieces of software like Reason and EJay, etc. which wasn't that much more work than creating a prompt for AI and curating the resulting generation.

One of my first thoughts when seeing this is, for those folks that make a living from writing jingles for advertising etc are all now out of a job
:(

AI generated music is still a long way from being able to be truly sculptured for a specific purpose, for the best output you need to be able to describe what you want using technical terms which in most cases people would only know by having an equivalent level of ability to create music themselves. But the rate of progress is scary.
 
Last edited:
Suno V5 model recently dropped and while it is has some caveats, the better generations are hard to tell are AI now and it has significantly more range and variation than the older models and less reliant on rehashing existing work.
 
yeah seen this a while back. AI is good for medical advances etc but we seriously need to keep it out of the arts/music
it's not something i'll ever want to listen to anyway
 
Perhaps the focus for the future of music artists will be more about live (and therefore obviously authentic) performances.
 
Perhaps the focus for the future of music artists will be more about live (and therefore obviously authentic) performances.
For a lot of artists, I think this is true.

Similarly, if AI actors and AI scripts and whatever for films comes in, I could see theatre becoming more popular.
 
Naa, last few decades mostly. 90's and 80's slop at least had character. :)

Rose tinted specs (or headphones) doing a lot of heavy lifting there!

"Popular" music has always been generic and derivative, with the very occasional standout track, we just conveniently forget all the dross and remember the 1-2 songs every year that were actually worth listening to!

I've found significantly more good (new) music in the last 5-10 years than in the previous 30 combined!
 
If you can only find music you liked in a specific decade or two you either

- have insanely limited taste in music
- have never really crate dug or looked outside the mainstream for music
- got old and grumpy
- were never really that into music to begin with
- associate your favourite music with memories of when your were young and life was simple/fun, thus nothing new can ever compare.

At least one of these is always accurate.

It’s amazing how many middle aged people are convinced everyone stopped making good music at the exact moment they got old… a truly remarkable phenomenon, doesn’t matter what generation. You hit 40, you look at your balding head and rapidly expanding waste line and then just like that… the realisation hits that absolutely no one makes good music anymore.
 
Last edited:
If you can only find music you liked in a specific decade or two you either

- have insanely limited taste in music
- have never really crate dug or looked outside the mainstream for music
- got old and grumpy
- were never really that into music to begin with
- associate your favourite music with memories of when your were young and life was simple/fun, thus nothing new can ever compare.

At least one of these is always accurate.

It’s amazing how many middle aged people are convinced everyone stopped making good music at the exact moment they got old… a truly remarkable phenomenon, doesn’t matter what generation. You hit 40, you look at your balding head and rapidly expanding waste line and then just like that… the realisation hits that absolutely no one makes good music anymore.

The sad thing is that this seems to be true for the vast majority of people :(

It's easier than ever to both make and find good music these days, the barrier to entry has been dropped dramatically with the spread of the internet and (relatively) cheap home studio hardware/software.

Of course that means there's a lot more dross to filter through, but it also means that you can listen to some amazing indie artists from across the world, who 10-20 years ago would have been confined to sharing CDs among their mates and "touring" a few bars in their local area (if they ever even got out of their parents' garage).

I think this is a separate topic to AI music creation though, and while @Rroff has compared it to creating a track using samples in a DAW, I think that's doing artists who use samples a disservice; there's still a skill in finding the right samples and knowing how to make them fit into your track, manipulating them correctly, mastering etc. and - no disrespect intended - there's a world of difference between slapping a few samples together into something that sounds ok, vs writing a professional sounding and complete track.

I suspect it's going to be a while still before AI can do anything more than the former without someone already relatively skilled at music production doing most of the work; at that point, where do you draw the line between "AI created" and it just another being another tool/instrument?
 
Last edited:
If you can only find music you liked in a specific decade or two you either

- have insanely limited taste in music
- have never really crate dug or looked outside the mainstream for music
- got old and grumpy
- were never really that into music to begin with
- associate your favourite music with memories of when your were young and life was simple/fun, thus nothing new can ever compare.

At least one of these is always accurate.

It’s amazing how many middle aged people are convinced everyone stopped making good music at the exact moment they got old… a truly remarkable phenomenon, doesn’t matter what generation. You hit 40, you look at your balding head and rapidly expanding waste line and then just like that… the realisation hits that absolutely no one makes good music anymore.

Probably the last 2 for me.
I don't find much new music good anymore. And stick to the sound style I liked when I was younger.

Its not good or bad music. It's just tastes and how brain develops I guess. Nostalgia is very real. And music is very powerful at stirring memories. Even with people with mental issues.

Its annoying when people say "xyz genre is bad" as music is music. But liking what you like? That's fine too.
 
Last edited:
I think this is a separate topic to AI music creation though, and while @Rroff has compared it to creating a track using samples in a DAW, I think that's doing artists who use samples a disservice; there's still a skill in finding the right samples and knowing how to make them fit into your track, manipulating them correctly, mastering etc. and - no disrespect intended - there's a world of difference between slapping a few samples together into something that sounds ok, vs writing a professional sounding and complete track.

I suspect it's going to be a while still before AI can do anything more than the former without someone already relatively skilled at music production doing most of the work; at that point, where do you draw the line between "AI created" and it just another being another tool/instrument?

True there is a big range of skills required when it comes to creating a tune with samples in a DAW and that making something good, though the latest update to Suno has a Studio mode where you can build tunes by breaking down AI generations into stems and extracting bits to build loops, etc. to create tunes.

For example though this is something I slapped together with samples in eJay years ago when I was like 14 with rudimentary music knowledge:


(No prizes for guessing the inspiration)

It didn't really take that much more effort than building a decent style prompt for AI music and curating the output to find something you like.
 
True there is a big range of skills required when it comes to creating a tune with samples in a DAW and that making something good, though the latest update to Suno has a Studio mode where you can build tunes by breaking down AI generations into stems and extracting bits to build loops, etc. to create tunes.

I guess there's a very fuzzy line between a person telling an AI to create music, vs using an AI to generate the sounds for them to use to create music. I'm sure a similar argument was made when people started using synths to make music e.g. you're not playing an actual musical instrument, it's just a computer making an electronic sound, (or from the rest of the tribe when Ugg started hitting a deer skin with a stick in his cave instead of singing :p).

A couple of my synths and midi controllers have generative functions, where I can pick certain notes and hit a button and it will randomly order them, or even pick certain parameters and it will tweak and build a sound around those parameters, using AI is just a logical extension of that (IMO); you still need a musician's ear and knowledge (both of which I apparently lack :() to turn that output into something musical (or in fact to set up the inputs in the first place)

For example though this is something I slapped together with samples in eJay years ago when I was like 14 with rudimentary music knowledge:


(No prizes for guessing the inspiration)

It didn't really take that much more effort than building a decent style prompt for AI music and curating the output to find something you like.

As I said before though, that sounds "OK", but ultimately is just a relatively generic sounding loop, the mastering is all off, the "instruments" don't really sound cohesive, and there's no structure or progression to it. It's inoffensive, but nobody is going to be adding it to their Spotify playlist! (no offense intended, my stuff is no better!)
 
It’s an ideological matter at the end of the day. Are people making music as artistic expression, or do they just want to churn stuff out for mass consumption? To the extent that it’s the latter, why would Spotify (or whoever) pay third parties for that, if they can just generate the slop themselves? At some point there’s a risk that the bottom falls out of the music industry - if the “music” can be generated at essentially zero marginal cost, why would anyone pay for it?
 
It’s an ideological matter at the end of the day. Are people making music as artistic expression, or do they just want to churn stuff out for mass consumption? To the extent that it’s the latter, why would Spotify (or whoever) pay third parties for that, if they can just generate the slop themselves? At some point there’s a risk that the bottom falls out of the music industry - if the “music” can be generated at essentially zero marginal cost, why would anyone pay for it?

I guess that comes down to the listener's choice.

If there's no skill or cost involved then they can cut out the middle man and just make their own music surely?
 
Last edited:
Back
Top Bottom