Learning, Choice, and Controlled Change

Status · Active research

Central question

Can prior pressure and consequence produce a measurable, attributable change in future selection?

Why the question matters

Systems often describe any changed output as learning. This program asks whether the change can be measured, traced to a prior event, distinguished from collisions or operator effects, and tested again.

How the investigation began

The investigation began with the Seeing Loop theory and moved toward traceable trials, descriptor constraints, collision handling, and a governed candidate-influence prototype.

Human–LLM collaboration method

Proposals are recorded before selection. Outcomes, collisions, operator interventions, and future choices are preserved. Claims are limited to the behavior the trial can actually show.

Projects and experiments

• Seeing Loop
• Choice Trace

Verified accomplishments

• Seeing Loop Trials 001–003 produced recorded supported, failed, and ambiguous cases.
• Failure analysis identified descriptor requirements, collision constraints, and a same-operator confound.
• Choice Trace demonstrates a governed prototype for recording candidate influence without silently authorizing it.

What has not been demonstrated

The program has not demonstrated autonomous learning, general intelligence, consciousness, or an operator-independent causal effect.

Failures, ambiguity, and competing explanations

The same person designing, running, and judging a test can influence the result. Small trial counts, descriptor ambiguity, collisions, and alternative explanations remain material.

Selected articles

• The Seeing Loop
• Seeing Loop Visualizer
• Paper Collision: AI, Human Cognition and Knowledge Collapse

Selected podcasts

These episodes show continuity in the questions; they do not validate the software experiments.

• Charles Darwin – Father of Natural Selection and Evolution
• Isaac Newton – The Father of Gravity

Current position

The trials are useful research records. The central theory remains under investigation.

Open questions

• Can a precommitted independent operator reproduce the selection change?
• Which descriptor and collision rules must be fixed before a trial begins?

Next bounded test

Run a precommitted replication with an independent operator, fixed descriptors, declared collision handling, and a falsification condition.