Dunno if you’ve been seeing these promoted posts for Suno music on facebook… they are ads featuring the likes of massively successful pop producer Timbaland extolling the praises of Suno. And also unknown people dancing around saying how great it is that Suno made their lyrics into a song.
In case you have not been following this and need a catch up:
In the past year, some companies have been developing products that create music from text prompts. Two of these companies, Suno and Udio, are each valued at $500million, and charge people for their service of creating music from text prompts.
I tried both of them for free at a point to see what they were like. Using prompts like “create a techno jazz track with a Detroit-influenced sound and bebop trumpet melodies” led to… nothing. Something so far away from what that sounds like in my head that it was utter failure. 0/10. Goodbye.
Why was it so bad at creating a synthesis of these 2 niche genres? My guess is because it’s expensive for these companies to farm data to feed their LLM (Large Language Model) so they don’t bother with niche music, they concentrate on feeding them all the popular music. Try asking it to write a country song, or a pop song, or any combination of popular music genres, and it will do a decent job. It’s comparatively easy to synthesize these genres because A) there is so much of it and B) most of it is so similar.
So, my own personal interest in using it stopped immediately after using it once. And I thought, well it’s not gonna affect me, because what I do is millions of miles away from what it can do, and at least SOME people will be more interested in hearing what I can do.
But, I have been seeing these ads and reading the comments under them, and I’ve noticed something very curious.
MOST comments align with my thinking that is: if you write some text into a prompt and ai spits out a song, you are NOT a musician. You are NOT making music. If you think you are… you are deluded.
It is the comments from people who think that this is the most wonderful new tool to enable them to express themselves that interest/scare/worry/frighten the beejazus out of me.
Here’s an example I saw this morning under a Suno advert:
Yes Paul Sergio, absolutely, I agree with you 100%. I don’t care to hear anything that is “perfect”. I want to hear the PAIN, I want to hear the EFFORT of YEARS AND YEARS of unrelenting slogging it out to GET SOMEWHERE you’ve never ever been before!
Listen to Coltrane playing Transistion! Why does it move me? Why is it powerful? It’s not because of the notes or the tone of the sax or the band thrashing out uptempo swing at 4am in a recording studio after playing their 5000th gig! It’s because Coltrane WENT somewhere deep inside himself that Ai can NEVER even come close to capturing.
All it can do is recreate and resynthesize what’s already been created by humans. The very nature of it is that it can’t do anything new. ALL I WANT is to make something NEW!
So back to the scary thing, it’s what Dave Wiseman says. That Paul is wrong, and that he is creating beautiful songs that move him to tears (using ai). What’s scary about that? It is very close to, or perhaps is, a case of ai pyschosis.
Quoting from a bbc.co.uk article: “a new condition called “AI psychosis”: a non-clinical term describing incidents where people increasingly rely on AI chatbots such as ChatGPT, Claude and Grok and then become convinced that something imaginary has become real.
Examples include believing to have unlocked a secret aspect of the tool, or forming a romantic relationship with it, or coming to the conclusion that they have god-like superpowers.”
https://www.bbc.co.uk/news/articles/c24zdel5j18o
THIS IS SCARY. Many, many people are actually believing that they are not just musicians, but GREAT musicians, because the ai tool they use somehow unlocked some super, special level of creativity within them that they could not otherwise access, and unleashed THEIR amazing music on the world!
Real musicians and musicmakers might be stunned and laughing at this (especially when those same ai musicians start complaining about how hard it is to get anyone to listen to their music lol)
This guy is sad because no one is listening to his ai generated music. Boo hoo! Why isn’t anyone reading my ai generated book? Why isn’t anyone buying my ai generated art for millions of $$$? Well, because it is worthless. Sorry.
Now here is a really sad truth… some people will feel attacked and angered by what I’m saying. Because they are in a kind of ai pyschosis, their experience is that ai helped THEM to do something great without all the years of practice and study that it actually takes to do something great. So they’ll develop a big attachment to this wonderful ai tool and anyone who says anything negative about it is just a technophobe who doesn’t understand this is just the advancement of the human race.
Could I be wrong? Could some of this ai music actually be great? Will ai actually produce the next great symphony, the next great underground house music record, the next great big band jazz arrangement?
I want to say no, it can’t.
But I’m biased, I spent the past 34 years playing music, studying, learning, practicing, practicing, practicing… so OF COURSE I don’t want to believe that was all a big waste of time and I should have just sat around waiting for wonderful Suno to come along.
BUT, despite all of this, it is perhaps possible that we are moving into a new, dark time of humanity when music becomes something else… something that each individual generates for themselves to listen to, using LLMs, maybe there is no more need for the advancement of art and music, maybe we’ve come to the end of all that.
Pfffff I doubt it.
Oh btw I have a new album coming out soon LOL