CITY GATE / CONTEXT BEFORE FRONTIER
A research frontier is not a shelf. It is a territory you enter with context, a question, and the right to return without a claim.
illustration only · territory map, not runtime evidence
THE ATLAS / TWELVE PAPER TRAILS
The City is the shared context. Beyond its gate, each source takes one organ-territory: memory, handoff, oversight, evidence, orchestration, or the still-unknown edge.
THE READING RULE
Research is not a mood board for the architecture. Every reference on this shelf has a named reader, a place where its pattern might transfer, and a question that can still prove us wrong.
ARXIV / TRACKED SHELF
Twelve primary references are surfaced from the current architect identity and technical-frontier reports. They are source links and transfer hypotheses — not proof that Colony has reproduced the papers.
HOLA pairs a compressive recurrent state with a bounded exact KV cache selected by surprise, giving the lost prefix somewhere precise to survive.
MemoHarness treats the harness itself as editable: execution diagnoses and a dual-layer experience bank adapt control to the case.
A language game gives a non-neural system a route to express its own dynamics through action, instead of borrowing a model's voice.
MI9 frames runtime governance as continuous telemetry, authorization, conformance, drift detection, and graduated containment.
HILA trains a policy to decide when an agent should solve and when it should defer, then uses expert feedback for continual learning.
Delegation is treated as a sequence of decisions with role boundaries, intent, authority, responsibility, accountability, and trust.
Natural-language intent compiles into an editable workflow graph with reusable components, topology preview, tracing, and human-in-the-loop interaction.
Compositional delegation and scope attenuation make authority a bounded contract that can be layered over existing policy.
OrchVis makes goal alignment, task assignment, dependencies, verification, and selective replanning visible to the supervising human.
Magentic-UI studies low-cost human involvement through co-planning, co-tasking, action guards, multitasking, and long-term memory.
AutoResearchClaw turns failed experiments into information through debate, a self-healing executor, verifiable reporting, and targeted human intervention.
Argus pairs searchers with a navigator that keeps a shared evidence graph, dispatches missing pieces, and verifies the assembled answer.
SPECIAL TECHNICAL PERSONAS
These are craft seats, not celebrity endorsements. Each one names a technical lens and a bounded Colony surface.
Meaning, boundaries, roadmap, and the decision that keeps the system coherent.
COLONY / architecture & roadmapFind the small rules that generate complexity and test whether self-evolution is real.
COLONY / self-evolutionPredictive world models, self-healing code, and transfer across the next release.
COLONY / repair & transferMemory, planning, imagination, and the loop between a remembered past and a possible future.
COLONY / NeocortexObjectness, symmetry, topology, and the priors that make a grid become a world.
COLONY / ARC semantic catalogFragment-first dispatch, capacity-aware routing, and a graph that does not wait on one voice.
COLONY / schedulerTrust, resource allocation, runtime health, and the boundary between an advisory idea and an action.
COLONY / governanceBreak the claim, inspect the receipt, and leave the seal blank when the evidence is thin.
COLONY / falsifierTHE ARCHITECT'S DESK · AA_AI_ARCHITECT
The Architect owns priorities and development boundaries, then asks the harder question: what must remain true when a research pattern becomes a Colony primitive?
SWE, ML, ARC, research, and release each get an explicit surface.
A broad frontier becomes one meaningful next question, not a swarm of slogans.
A decision is not complete until the effect, authority, and remaining uncertainty are inspectable.