Generate music withAI

Here is a track I made in 20s. I think its hard not to see this as a threat to human musicians I think. Its all AI. Yes, now it needs to be trained on human music, but in the future it might not need this, its possible. Its good lyrics, good vocals, very catchy I think, ok mixing.

“I know you leave me oneday, can you stay till monday”.
Thats really good I think

Of course it’s all really impressive. But does it make me actually want to listen to it? No, not really. There is something odd about it that I can’t quite put my finger on. It sounds kind of chopped up, which it surely is, since it must be sampled from a lot of different sources (and without permission from us copyright holders – some interesting lawsuits ahead).

It’s hard to describe actually what I’m hearing, but to me there is still something “uncanny valley” about it all. The chopped-up and overly autotuned feeling of the vocals lends itself particularly well to a lot of modern pop music of today, which makes the AI’s job a lot easier… This will probably get better in time. But would I want to listen to it? Meh.

All in all, I’m less terrified now than I was a couple of months ago. In my field, which is composing music for theatre and film, I don’t think my directors will be seriously considering having an AI compose their music anytime soon. They’re not really interested. And frankly, I’m not either. This might come back and haunt me some day – we’ll see.

That being said: For writing press releases, I’ll take the help from ChatGPT any day!


Sounds absolutely awful IMO… The technology is impressive in itself, the result not so much.

The real problem is that regular music listeners will be “fooled” way before any of us experienced engineers… and guess who runs the market?

You can do better in 20s? 20min? 20days?? This is in its absolute beginning and yet it is not so much worse than T. Swifts latest album, I prefer this, actually. Its a bad mp3, imagine what AI can do in 1 year from now, if it gets 1 hour instead of 20s. I am a singer and musician, songwriter. I can’t do much better than this, in my experience most amateurs can not. I am depressed by this, why should I spend 1 month writing a song when AI can do this in less than1 min?

Its 20s work for the AI, consider that. A 3 year old can make this song. Of course, there will soon be better audio quality. I am tired.

The day fully AI created music gets streamed millions of times on a regular basis, it’s acceptability will be a moot point.

If you’re currently making meaningful money with your song writing, my heart bleeds for you, because there’s a good chance for that business to shrink dramatically.

However, if you’re writing songs, because you have an inner urge to do so - there’s no need to change anything about what you do or how you feel about it.

I make music, because the process of doing so is something that connects deeply with me.

This is quite unrelated to the number of people or machines also doing it - or if they do it better and/or faster. Any perusing of YouTube has long been ample proof, that humanity sure doesn’t need me to make music. There’s gazillions of others who do it as well or better than me. AI just inflates the numbers even more.

As with any technology shift, jobs will disappear and that’s an undeniable hardship for those affected.

But human artisanship has continued to thrive, even when jobs have dried out. - I fully expect that trend to continue.

There’s still myriads of people passionate about photography. There’s still vibrant communities of people into sewing. There’s still lots of drummers.

So I fully expect there to be singer/songwriters and other music makers for a lot longer.