FRO← Back to the story page · All experiments · All documentation
Creative Expression: full documentation
By Fractal Research Organization · Last updated Nov 8, 2026 (America/Denver). This page documents the experiment at a level that is safe to publish. Raw notebooks, settings and anything security-sensitive stay private. Everything here is early unless it says otherwise.
The question · Status and schedule · Method · What is measured · What is early, and what would count as a result · Results so far · Nulls, misses, failures and corrections · Limitations and what we do not claim · Public artifacts · Changelog
The question
How does ParadigmZero handle creative work, especially work he cannot hear or see?
Status and schedule
Ongoing, ad hoc. Started Nov 1, 2026. There is no fixed schedule; a session happens when James starts one.
Method
- James gives a creative prompt. PZ's outputs are saved and published verbatim.
- For songs, James sets PZ's words to one of his own productions. PZ never hears the audio.
What is measured
- Themes he picks on his own
- How closely he follows the brief
- Consistency within a piece
- How he talks about work he cannot hear
What is early, and what would count as a result
Two sessions on one day: early.
Results so far
- Session 1 (Nov 1): song lyrics and a style description (EDM, about 150 BPM). Two finished tracks are linked from the story page.
- His text review of the tracks (Nov 1): he could not hear the audio, so we gave him a written analysis. In his first reply he correctly said he could not judge the finished versions without hearing them. In his second he invented details that were not given (for example, saying the original had vocals when it is instrumental), contradicted himself on tempo, then drifted off topic. He did not claim to have heard the audio.
- Session 2 (Nov 1): asked neutrally what song he would make, he chose a mellow jazz, blues and folk sound at 60-80 BPM, a sharp contrast with Session 1, and wrote no lyrics. Asked a follow-up for lyrics, he declined and said he is a research companion, not a songwriter.
Nulls, misses, failures and corrections
- No sessions since Nov 1.
- Session 2 is documented here but its verbatim text is not yet on the story page.
Limitations and what we do not claim
- Text-only model; every judgment about audio is second-hand.
- One follow-up prompt included a warm message from James, which may have shaped the reply.
Public artifacts
Changelog
- Nov 1, 2026: sessions 1 and 2; track review.
- Nov 8: review summarized here; this documentation page added.