Wonderful question. I suspect it's partially the culture issue you point to, but also a practical issue of composition. If we decompose sound into the basic waveforms, similar to the subject pdf on page 18, we then have parts that we can reassemble. We can take the defense-funded DSP math of the likes of a John Cooley or a John Tukey and build an engine for assembling the parts of sound.

All this being said, I think that's a process of convenience and a historical path not a absolute constraint. We have some more flexible means of communicating with the machines today. And I strongly encourage someone to work on a new UI for computer music. "Jazz trio piano, upright bass, and drums. start drummer laid-back, piano blowing over the changes, then piano on top."

In short: Not really.

As another commenter below has said, "mathematics might be a useful way to understand music", but it's not how compelling music is made.

Mathematics are fundamental to scales and the harmonic series, and knowing about them will help you refine certain choices, but it's not going to help you write a dramatic melody or an emotionally resonant chord progression, or play an energizing rhythm, even if there are mathematical explanations sometimes.

Good music comes from being a good listener, having a strong sense of what's possible, where it could go, and then delivering something surprising. Telling a story with your melody and supporting the arc of that gesture with harmony that accentuates or contrasts it.

Again, there's a mathematical explanation for harmony and dissonance, but players aren't thinking that granular. They're operating at a higher level of abstraction one, two, or three levels above that: They're thinking about telling a story, evoking an emotion, and exciting an audience in the moment.

Likely historically true, but not anymore.

As a software developer I see that LLMs are better at the "craft" of making software.

Software developers training are overwhelmingly analytical.

Musicians will experience the same. That the quality of Ai generated music is superior. But it will come more as a chock for the reasons you explain.

> Is thinking about music as applied mathematics a good way to create good music?

As an instruction, I think clearly not, the fact that lots of musicians aren't mathematical at all but create great music seems to prove it to me.

But it is interesting to think about musicians who do seem to think about music this way. Bach is definitely a good example where the system of counterpoint is very complex. I'm not sure if she'd describe herself in these terns, but I've always got the impression Laurie Speigel thinks about music a little like that too. Then there's stuff like Coltrane's Giant Steps, where the whole piece is based around a sort of music theory "trick".

So maybe not generally, but there's definitely some awesome music out of that kind of relationship.