TurtleBlock AI turtle and bricks on an AI chip🌿🐢 Building in public

Turtle Terraria

Places where ideas live long enough to change.

Turtle Terraria is the umbrella research environment inside TurtleBlock AI. It contains bounded habitats for observing what emerges when people and computers construct, question, reflect, and revise together — and what happens when human-made computational agents recursively do the same with one another.

cohabitation without provenance collapse: human meaning ≠ Turtle interpretation ≠ synthetic self-play ≠ scholarly source

Two habitats to begin

🧑‍💻🐢 Human + Turtle Terrarium

A person drives the inquiry. Turtle contributes questions, interpretations, alternatives, information, construction, and technical assistance. Human purpose, correction, judgment, reflection, values, and ownership remain human-authored evidence.

Research question: what happens when people think and make with computers?

🐢🐢 Recursive Turtle Terrarium

Synthetic Turtle roles construct interpretations, challenge one another's assumptions, run hostile cases, test the Turtle Charter and Sanders ontology, and revise their representations.

All traces are explicitly synthetic_self_play, never learner evidence, and carry no production authority.

More habitats are expected

The architecture intentionally leaves open space for future terrariums. New habitats should be added only when a genuinely different research condition, participant relationship, artifact pathway, or observational boundary appears. Terraria is plural because the project should be able to compare environments without pretending they are the same environment.

Critical Techno Constructivism as an observational frame

The seven established CTC domains remain the primary operational frame. Every meaningful Terrarium event can map to all supported domains with rationale rather than a single forced tag.

1. Personal Inquiry — who originated the problem, purpose, or direction?
2. Compelling Problem or Question — what uncertainty sustains investigation?
3. Technology as Tool to Think With — what did Turtle make visible, manipulable, comparable, or testable?
4. Formative Demonstration of Learning — what artifact, representation, behavior, or state changed?
5. Reflection as Learning — what caused reconsideration or revision?
6. Social and Cultural Critique — what assumptions, values, positionalities, power, or consequences surfaced?
7. Sharing and Collaborating — whose contribution affected what emerged next?

And room for CTC 8, 9, 10… if evidence earns them

Observations that do not fit the seven established domains are preserved rather than force-fit. The database can accumulate uncaptured residue, identify recurring patterns, and propose provisional candidate domains. A machine may notice a pattern; it cannot canonize a new CTC domain. Promotion remains a human scholarly decision.

established CTC 1–7 → uncaptured observations → recurring pattern? → provisional candidate → human scholarly judgment

Exhaustive tagging

Terraria treats questions, responses, interpretations, revisions, artifacts, failures, disagreements, rejected alternatives, tests, source signals, implementation decisions, and observations as potentially researchable objects. Each object can carry actor, evidence class, privacy class, source, model/configuration, version, CTC mappings, Sanders ontology mappings, emergent tags, relationships, and unresolved residue.

The rule is exhaustive within a bounded capture batch, not infinite recursion. Tags themselves remain metadata unless the act of tagging becomes intellectually significant — for example, a disagreement about a tag changes the research or the system.

What one Terrarium can teach another

The overarching comparative question is deliberately difficult: what can observations from one terrarium teach us about another without confusing synthetic machine behavior with human learning? The environments may cohabitate in one research system while retaining distinct provenance and evidentiary status.

Research lineage in practice

STEAMHAMLET was the room of possibilities. TurtleBlock AI makes the participants computational. Turtle Terraria lets the project study what happens when they live together. The same lineage connects Critical Techno Constructivism, purposeful play, Minecraft as an inhabitable learning environment, persistent computing environments, and co-active emergence.

Current implementation edge

Migrations 0007 and 0008 define Terraria habitats, runs, participants, events, artifacts, observations, universal research objects, CTC and ontology mappings, emergent tags, uncaptured observations, auto-build records, and the first structured seed traces. The next step is to apply and verify the live D1 schema, then begin writing Human + Turtle and synthetic Recursive Turtle traces into it without collapsing their evidence classes.