This project is a discontinuous journal that helps people in prison develop Bayesian thinking and evolutionary psychology thinking. It allows individuals to reflect on their experiences while providing anonymous data for research on personal growth.
While such an approach may not offer immediate or large-scale transformations, even a small improvement in the reasoning of one person could have disproportionate long-term impact. In some cases, the presence of a single altruistic agent might matter more than saving a hundred others who cannot act in turn.
1. ReStructID: A Lightweight, Gamified Self‑Assessment Tool
We propose a system to help incarcerated users capture and categorize “reference moments”—episodes of high/low motivation and satisfaction—across four process axes (In/Out, Physical/Personal, Informational/Social). It then maps these data points onto eight identities (Guardian, Warrior, Versatile, etc.) using pie‑chart visualizations, guiding users to define goals, habits, and tasks tied to XP and weekly checkpoints.
Key decisions:
Flutter framework for cross‑platform (Android/iOS) ease of maintenance, version in paper optional.
Reference moments approach as a minimal viable input method, requiring only brief, structured entries.
Identity mapping & XP create intrinsic motivation loops with IA.
Continuous tracking empowers users to update their priors about what works for them.
Compared with traditional in‑person curricula—where coverage is limited by instructor availability and static content—ReStructID offers on‑demand, personalized feedback, reducing marginal cost per user toward zeroRAND.
2. Why It’s More Efficient Than Other Proposals
Digital + Low‑tech hybrid: A paper planner replica ensures accessibility when devices are restricted, without added staffing costs.
Self‑guided learning: Users learn to infer their own patterns (Bayesian updating) rather than depend on scheduled sessions.
Gamified engagement: XP and identity levels sustain motivation, addressing the “drop‑off” typical of one‑off workshops.
Minimal training: No need for specialized facilitators—any staff or volunteer can onboard users in minutes.
By borrowing the self‑reinforcing loops of large‑scale AI training but applying them to human cognition, we can achieve comparable adherence at a fraction of the per‑capita cost of standard correctional educationWikipedia.
If the combination of Bayesian reasoning and AI-assisted reflection can support the reconstruction of even one individual’s identity in a prosocial direction, the cumulative benefits could extend well beyond that case. We acknowledge the limits of this hypothesis, but suggest that helping one person become a more thoughtful and altruistic actor might, indirectly, lead to outcomes more meaningful than many direct interventions. Perhaps even saving half a gorilla is not a trivial result—if it’s part of a broader shift in human decision-making.
We welcome feedback, potential pilot partners, and co‑founders eager to bring more‑leverage tool to life.
Summary
This project is a discontinuous journal that helps people in prison develop Bayesian thinking and evolutionary psychology thinking. It allows individuals to reflect on their experiences while providing anonymous data for research on personal growth.
While such an approach may not offer immediate or large-scale transformations, even a small improvement in the reasoning of one person could have disproportionate long-term impact. In some cases, the presence of a single altruistic agent might matter more than saving a hundred others who cannot act in turn.
1. ReStructID: A Lightweight, Gamified Self‑Assessment Tool
We propose a system to help incarcerated users capture and categorize “reference moments”—episodes of high/low motivation and satisfaction—across four process axes (In/Out, Physical/Personal, Informational/Social). It then maps these data points onto eight identities (Guardian, Warrior, Versatile, etc.) using pie‑chart visualizations, guiding users to define goals, habits, and tasks tied to XP and weekly checkpoints.
Key decisions:
Compared with traditional in‑person curricula—where coverage is limited by instructor availability and static content—ReStructID offers on‑demand, personalized feedback, reducing marginal cost per user toward zero RAND.
2. Why It’s More Efficient Than Other Proposals
By borrowing the self‑reinforcing loops of large‑scale AI training but applying them to human cognition, we can achieve comparable adherence at a fraction of the per‑capita cost of standard correctional education Wikipedia.
If the combination of Bayesian reasoning and AI-assisted reflection can support the reconstruction of even one individual’s identity in a prosocial direction, the cumulative benefits could extend well beyond that case. We acknowledge the limits of this hypothesis, but suggest that helping one person become a more thoughtful and altruistic actor might, indirectly, lead to outcomes more meaningful than many direct interventions. Perhaps even saving half a gorilla is not a trivial result—if it’s part of a broader shift in human decision-making.
We welcome feedback, potential pilot partners, and co‑founders eager to bring more‑leverage tool to life.