[deng]

> I've got 25 years of loops that basically to finish them need better arrangements

Welcome to the club. You need to learn how to actually finish a track, which is the most difficult but also the most rewarding part. Why would you use AI for that? I mean, just listen to that demo track Codex made in the above repo, you surely don't want that.

There's a good book about this, published by Ableton, you can read it for free here:

https://cdn-resources.ableton.com/resources/uploads/makingmu...

[bschoepke]

To be fair, the demo track was one of the first I had it make, and I didn't put much effort into it because I thought it was especially funny with the macos "say" command vocals.

It's a garbage-in, garbage-out situation. If you give it more musical direction you will get more out of it.

[deng]

It's not just that the track is garbage (the "say" vocals are actually the least of its problems). Even if AI would make a good track: why use AI for creating your arrangements in the first place? Why this resistance to actually getting good at something? I can understand if your livelihood depends on it and you just need to be fast, but why for stuff you do for fun?

The book I mentioned has a good suggestion when struggling with arrangements: just copy. Take a track you really like, put it into your DAW, sync the speed and replicate its structure. You'll see that in many genres, structure is often exactly the same anyway. This can be an eye opener, and once you've realized this, you'll be able to experiment with structure in ways you couldn't do before. That's the fun part.

[bschoepke]

Well, I do that too. I actually spend most of my music hobby time with my Ableton Move. I just happen to find it very fun to play with Codex as well.

[deng]

I agree that having an MCP Ableton can make total sense. After many years of use, I would say I know Ableton quite well, but nowadays, I regularly ask ChatGPT if certain things could be done differently/more efficiently, and it often surprises me with new ways of doing things. For instance, sometimes Ableton has gotten new features over the years I'm not aware of. It surely would be nice to have this integrated via MCP.

I think you would get much better feedback if you'd focus on these use cases: flattening the learning curve for newcomers, and new ideas for experienced users, rather than creating tracks completely by AI. Because in that case, why even go through a DAW and not use Suno directly?

[bschoepke]

Thank you for the feedback, I'll record some more co-creating examples! And yea, it's also fun to take stuff out of Ableton and run it through Suno, to get real vocals and such.

A cool thing about this MCP is you can ask Codex Ableton questions and it will go and read the state of your current Live Set and answer based on that. You don't have to have it change anything for you if you don't want.

[2bewithu]

on a side note, his early 2000s output is genius

[tim-projects]

They had the most emotion because I didn't know what I was doing. Then for any ten years I got lost trying to make it sound technically good. Now I must make stuff and there technical part happens automatically

[tim-projects]

> Welcome to the club. You need to learn how to actually finish a track

I don't think you understand. I've got thousands of songs. Why would I use Ai to generate arrangements... Maybe for ideas?

Maybe because certain things I'm lazy about?

Maybe because I've got thousands of songs?

It's not actually difficult to finish a song if your output is high enough. Sometimes the songs just come out without any struggle. But most of the time they don't.

I wrote and finished my first song around 1996. Using Cakewalk plugged into a midi keyboard.

HN is full of people who think using AI means you are lazy or can't do something. The fault is yours not mine. Adapt