FinalGod
Member
You mentioned Urs Heckmann as an example of genuine creative software development, and I certainly respect his work and the achievements of developers like him.I don't think, people understood my point. Let me try to simplify it (and don't jump on the simplification now)
But let's take a closer look at that example.
Did Urs Heckmann invent the mathematical foundations of oscillators, Fourier analysis, or subtractive synthesis?
Of course not. And that doesn't diminish his achievements in the slightest.
Many of the mathematical and technical foundations used in modern synthesizers were established long before software synthesizers even existed.
Just adding the dates here to give you an idea of how old some of the scientific and mathematical foundations behind your oscillator example actually are.
Joseph Fourier (1768–1830),
Heinrich Hertz (1857–1894),
Edwin Howard Armstrong (1890–1954),
Hermann von Helmholtz (1821–1894),
Harry Nyquist, Hendrik Wade Bode, and Robert Moog all contributed, in different ways, to the scientific and technological foundations that modern audio developers can build upon.
Developers use existing mathematical knowledge, combine established concepts, modify algorithms, experiment with different approaches, and create something new from them.
That's how a great deal of innovation works.
So why should an AI-assisted developer not be allowed to make use of the same established mathematical knowledge?
And more importantly, why do you keep assuming that the AI itself has to be the creative part of the development process?
It doesn't.
There is still a human being behind the software.
That person can have an original idea, define the requirements, experiment with different concepts, evaluate the results, and decide what works and what doesn't.
The AI can act as the programming department, implementing those ideas, suggesting technical approaches, generating variations, and even assisting with testing.
For example, I could ask an AI to implement ten different mathematical variations of an oscillator, listen to the results, select the most interesting one, and then develop it further.
The creative direction still comes from me. The AI simply allows me to explore those ideas much faster than I could by implementing every variation manually.
And if the result is a genuinely interesting instrument, why should it matter whether I wrote every line of code myself?
As for the growing number of generic plugins and synthesizer clones, we already had plenty of those long before the current AI boom.
AI may make it easier to produce even more of them, but it certainly didn't invent the concept of copying existing products or flooding a market with similar alternatives.
And here's another important point: Nobody is forced to buy any of these products.
That's called freedom of choice.
If you prefer software developed entirely by traditional programming teams, that's perfectly fine. You can continue buying and supporting those developers.
Other customers may prefer an AI-assisted product because it offers different features, a better workflow, a lower price, or simply does something they find useful.
Ultimately, the market will decide which products people actually want to use.
And I think we should stop treating AI-assisted software as if it somehow appeared out of nowhere without any human involvement.
Behind every AI-assisted software product, there is still at least one human being making decisions about what to build, what to change, what to keep, and what to release.
The quality, originality, reliability, and usefulness of the final product should be what matters.
Using AI doesn't automatically make a developer less creative, just as writing every line of code manually doesn't automatically make a product innovative.
If AI-assisted development with agents, source code access, and proper testing is perfectly acceptable to you, then aren't we essentially agreeing that the problem is not AI itself, but poor development practices and inadequate quality control?