innernet labslive
the resonance model
a model of how a person thinks and works, read from how they answer, not from what they claim. you can take the quiz now.
what it is
the resonance model is a v0 instrument. a person answers a short run of open questions across six chambers, and a reader scores the depth of what they wrote: surface, structural, generative, recursive.
it is checked against scripted personas whose depth is known, with controls for the ways an instrument like this fools itself: rewarding length, rewarding brevity, rewarding words that only sound deep.
on the last bench three of the ten checks failed. a recursive persona scored under a generative one, a long hollow answer scored like a structural one, and the same persona moved too far between runs. until those pass, it is a draft of an instrument, and it is described here as one.
the evidence
- checks passing, 2026-09-06 bench7 of 10apps/web/lib/resonance/bench-results/latest.md (bench 2026-09-06T20-56-07)
- sessions14 (7 scripted personas, two each)apps/web/lib/resonance/bench-results/latest.md (bench 2026-09-06T20-56-07)
- known groups, failssurface 30.5 < structural 41 < generative 64.5, but recursive scored 60.5, under generativeapps/web/lib/resonance/bench-results/latest.md (bench 2026-09-06T20-56-07)
- length control, failsa verbose, hollow persona scored 41.5 against structural 41, on about three times the wordsapps/web/lib/resonance/bench-results/latest.md (bench 2026-09-06T20-56-07)
- retest, failsthe same persona ranged 18 points, against a spread of 30 between groupsapps/web/lib/resonance/bench-results/latest.md (bench 2026-09-06T20-56-07)
- 14 sessions of 7 scripted personas, two each, not people
- three of ten checks fail, so its scores cannot yet be read as a measure of anyone
papers and notes
nothing written up for this area yet. what it has is the evidence above.