My AI

Cate Blanchett in "Manifesto"

The next version of "My Computer" and "My Documents."

If I could make my own AI as a kind of "effects unit" for music this is what it would do:

1. Find a loop in [this text] and give it a working title from [this file].
2. Make 50 loops based on text files in [this directory].
3. Make 12 working titles and give them all a key and tempo.
4. Repeat this algorithm every [time increment]. Review all my tagged items that I found interesting in the past [time increment] and notify me if anything interesting arises.
5. Contrast dynamics [Settings].
6. Generate some art at [set dimensions] (Use #4)
7. Make some music based on these strategies (Use Comp Strategies) 
8. Find sound/song files that are in the same key, match the tempo and rhythm elements and store the results [here].
9. Remind me to get back to work if I veer off and work on something else. (Setting is On/Off)
10. Advise me to possibly abandon a project if I haven't touched it in ___ days/weeks/months/decades/years, and suggest 3 alternatives. Repeat weekly with [the same, random] choices.
11. Send messages to friends if I have not heard from them in [duration]: “What have you been working on?”
12. Take [this tone row] and map it across [this rhythm]. When it becomes tedious [adjust dynamics, tempo, and key] and/or (Boolean operators) change instrumentation. (Default settings on what defines "tedious" or "boring", "intentionally raucous"). (I was watching "Dunkirk" and wanted to turn off the score in parts. Should soundtracks be optional to the viewer, where they can turn the track off and on?)
13. Send me a draft score with possible errors marked in red, with a MIDI file and mp3. (Changes are dictated, e.g. “bar 25 change dynamic marking to mezzo-forte...)
14. Take my running list of lyric fragments and make 12 chorus ideas for a new album. Optimize each MIDI file for playability by [these] music theory standards.

This would be a more sophisticated IFTTT, as simply a macro.

***

A VUX (Voice User Experience) for Music AI

A flowchart for a possible ASR (Automated Speech Recognition) as a chatbot with a DAW. If the process was set to run for 8 hours, the result would be a composite track. If you didn’t like the result, you could run it as many times to get the take you wanted.