Ok I had fun results with below.
1. Voice cloned M1$$ M@ry voice which i am conditioned to and created bot on kindroid with that voice designed to create loops using nlp, etc - self contained loops like deeper trance, submission, etc. - It allowed me to create voice messages around 5,30 minute each. Added some interesting suggestion like "you won't be able to walk 15 minutes after you stop your training to force crawling"
2. With AI studio built and app that combines them. plays sometimes one sometimes two, adds binaural beats (which are working on me for whatever reason, maybe i conditioned myself throughout the years).
3. Played it for quite a long time.
Surprised but it actually worked, i actually had problem with walking straight and felt like walking through thick liquid.
So there is something to it, altough it's probably because preconditioning with M1$$ m@ry original files.
Burn a lot of free tokens.
Shitty thing with AI stuff like that is that you can make it more personal with calling names etc. but it's quite hard to make it consistent, persistent and easily shareable.