Dreams, nightmares… tomato, tomayto.
The future sucks.
E. reformat, mix, filter, re-synthesize, edit or alter the sound samples and other musical content connected to the software for use in any kind of commercial sampling product/package or software, without the express written consent of the licensor.
So, if I’ve used a couple of instruments from the library in my tracks, does that mean I’m already breaking the law?)))
That is not what I read the license saying.
I am quoting only what is written.
It’s written that you’re not allowed to repackage the provided samples/demo songs/… in another sample pack or in a software you’d sell. similar license as for any sample packs out there.
You can do tracks with them…
Riffing on this, not intended as a direct reply to you, but more brainstorming here…
What helps to get to the “next level” is Version Control: Git. But could also be SVN, or whatever replaces Git in the future. It acts a code changelog that the AI, if it wanted, could scan through. It also lets you start fresh sessions, from where you left off in Git, without staying in the same chat. (Add a CLAUDE.md file or AGENTS.md file describing the role and the project conventions for good measure. The AI can generate it for you if you prompt it to)
About automated testing: This makes it so that when you add a new feature, if a test breaks, then AI knows the new thing broke the old thing, so it needs to account for it. If you’ve set it up as CI in GitHub (or equivalent) then these tests can be run at the end, outside of what is more important. Less tokens.
Also, with this setup you can open smaller issues, well described chunks of things you want done, and multiple agents could work in parallel using the Version Control system, opening individual pull requests. Divide and conquer. Less tokens.
Good time.