Multiplicity

Projects · Vol. I

Software, tools, and research.

Five works across software and research, on making behavior legible.

Each is a way to make behavior legible — human or AI. What acted, in what context, and what followed. Some are in daily use; others are research programs still in formation.

Gallery · Drag to browse

1 / 5
A watercolor scene of a parent and small child at a kitchen counter: the parent holds out a bag of pretzels and an applesauce pouch; the child points upward, manding. Surrounding the scene like a behavior analyst's hand-drawn working notebook: an 'active program: mand training' panel with sliders for MO sensitivity, schedule (FR-1), and prompt level; an 'active reinforcers' panel showing snack 72%, attention 18%, escape 10%; session timer 02:14 / 30:00; engagement bar at 62%; a session events log reading 'prompted mand /cookie/,' 'reinforced FR-1,' 'spontaneous mand.' Title 'BehaviorStream Simulator' hand-lettered across the top.

Working prototype

BehaviorStream Simulator

2026- · In active build · Behavior-analytic simulator

A behavior-analytic simulator. Motivating operations, schedules, response variability, and consequence dynamics are exposed as parameters; behavior emerges from their interaction. Procedures like functional analysis sit on top.

Historical cover-style illustration for Formative Grapher showing a large behavior chart and smaller cumulative, log-scale, and stacked chart variants.

In development

Formative Grapher

2015- · In refactor · Time-series graphs

An active browser-first rebuild of the Excel-based graphing tool for time-series data. The historical release remains available while the new version moves toward structured graph settings, quality checks, reproducible exports, optional analyses, and richer behavior data.

A pinstripe-suited lobster on rollerskates standing on a high-rise rooftop ledge above a foggy city skyline, with the title 'ClawSuite Relay' set against the cloudbank.

Current system

ClawSuite Relay

2026- · In use · Coordination layer

A coordination layer that lets named agent specialists delegate to each other, maintain continuity across exchanges, and stay visible throughout. An MCP version, independent of OpenClaw, is in development.

A stylized chart titled 'AI Behavior Analysis Benchmark.' A rat climbs ascending bars labelled with behavior-analytic dimensions — responding, reinforcement sensitivity, stimulus control, generalization, durability, fidelity — toward a flag reading 'emergent behavior.'

Research proposal

AI Behavior Analysis Benchmark

2026- · In planning · Empirical benchmark

An empirical benchmark for behavior-analytic reasoning in AI systems. Expert-written items, human BCBA baselines, informed by publicly available practice standards. Text vignettes first; data, graphs, and session artifacts later.

An engineering blueprint of a small robot labelled 'Project: Kind Gesture,' shown handing a bouquet to a parking meter. Annotations describe the scene from the outside — 'emotional output: hopeful,' 'meter response: no audio output detected' — without claiming access to inner states.

Research proposal

Functional Analysis of AI Behavior

2026- · In planning · Research program

A research proposal on describing AI behavior under black-box conditions without defaulting to mentalistic vocabulary. It asks what can be specified from the outside: conditions, response classes, and interventions that change outputs.