moranetz

What I actually do

Every product here started from the same operating question: what invisible human pattern is deciding whether this software gets trusted or deleted?

I build that layer directly into native iOS apps. The recurring subjects are cognitive workload, trust calibration, whether practice transfers.

Read the Room looks like an iMessage inbox. Underneath, every character runs on a psychological profile and every response is scored against it. The scenarios test whether you can read the person in front of you under pressure.

When the final character turned out to be immune to every influence technique except authenticity, I kept it. The system produced that result on its own.

Reckon and WeighIt take Heuer’s analysis of competing hypotheses, the method CIA analysts train on, and turn it into a touchable decision log: hypotheses ranked by what has survived refutation. Stillness regulates arousal with fully procedural scenes and synthesized audio; there is no asset pipeline to hide behind. SoloFinance reduces freelance money anxiety to three numbers on one screen.

The research side holds the apps to a standard. A pre-registered Bayesian meta-analysis of 39 named influence techniques found 15 of them have zero peer-reviewed support. A 245-record corpus tracks how attention is captured and held, compiled for retention work and kept current.

So when a team says, “the feature works, but users still don’t adopt it,” I usually find the same underlying failure: the model of the user is too shallow. The interface assumes rational behavior where the real driver is anxiety or ambiguity or social risk.

Off the keyboard, I host a recurring working session for people building in neurotech and aerospace: EEG, neural decoding, behavioral reward, human performance. Small room, recurring members, referral-only.