Minds · experiment notebook
Small systems.
What can they learn together?
Start with local estimates, connect them, and test their predictions. Each experiment keeps the selected rules, observed results and unanswered questions together.
Completed experiments
Experiments and what to build next →Experiment 01 · Explored
Two spaces exchange estimates
Watch two local estimates agree, keep swapping, or retain a gap. Their average can stay still while both parts move.
5 selected conditions
Explore the recorded results →Experiment 02 · Explored
What changes the answer?
Change tick size, message timing and references. Quiet values can still disagree, and missing messages differ from choosing not to respond.
24 conditions · 72 runs
Explore the recorded results →Experiment 03 · Explored
Predict before the observation arrives
Continue the last two samples, then check the next value. The line succeeds on ramps, loses to hold-last on steps and alternation, and leaves missing outcomes unscored.
12 selected sequences
Explore the recorded results →Next: when does a broader view help? Test whether observing the whole can warn of a disturbance or request useful detail, using information available before the outcome.
What makes a useful result?
Agreement, accurate prediction and useful action are separate questions. We retain misses, unresolved work and comparisons with ordinary exact arithmetic. The aim is to find where the structure helps and where it does not.
This work is motivated in part by Ingressing Minds and its invitation to examine unfamiliar forms of problem solving. These small experiments test concrete behaviors; stronger claims about cognition remain open.