Loading the institute record…
Glyphd Labs separates work by what the lab is doing with it now: active investigation, applied transfer, preserved option space, or evidence and reproduction. These portfolio lanes do not replace evidence states.
LCSM / Backboard · Project Core · Continuity
Research question. How can AI-assisted work preserve accepted truth, provenance, revision, and authority across models and sessions?
GlyphCAS · FOVEA Address Fabric · typed reuse
Research question. How can systems retrieve and reuse information without confusing semantic resemblance with exact identity?
Actor–Agent Runtime · Work Orders · Returns
Research question. How can durable people, projects, and institutions outlive temporary inference while constraining what agents may change?
SurfaceMind · SurfaceDelta · Workpages
Research question. How can models operate inspectable, source-linked project surfaces through bounded changes instead of regenerating entire answers?
fold/2 · content-addressable media · temporal operators
Research question. When can structured objects be represented, routed, locally accessed, and regenerated without surrendering exact recovery proof?
EL-JEFFE · SchemaStack · route–model–flight
Research question. Which tasks can compact local models perform reliably when schemas, routing, tools, and deterministic runtimes carry more of the system?
This lane contains systems intended to help an operator, team, researcher, or organization. It is deliberately honest about the difference between an implemented mechanism and a validated market.
A SurfaceMind workbench that compiles repositories, specifications, conversations, and receipts into source-linked project state and bounded accepted changes.
The mechanism is implemented; broad customer value and market demand remain separate questions.Inspect the system →A public projection that keeps publication status, claim state, implementation, fixtures, measured evidence, and external validation distinct.
It reports evidence; it does not independently create the evidence it displays.Inspect the system →An operating loop that can observe the corpus, propose work, prepare experiments, and assemble receipts while reserving consequential acceptance for a human operator.
WebMind may generate candidate evidence. It may not certify itself or promote its own claims.Inspect the system →Independent projects publish capabilities and constraints so bounded scouts can discover, test, and propose integrations without receiving ambient authority.
Open record →ARCHIVED SPECULATIVE ESSAYPreserved for authorship and chronology. It sits outside the current technical research program and is not presented as an engineering result.
Open record →SPECIFIED / INCUBATINGA constitutional simulation specification retained as a bounded systems experiment rather than a current flagship program.
Open record →SPECIFIED / INCUBATINGA spatial and social projection of project state retained for later interaction research and narrower validation.
Open record →SPECIFIEDA multiscale observation-based compute fabric in which systems publish fields and acquire only the regions, channels, and epochs they need.
Open record →SPECIFIEDA labor architecture for evaluating a person and their bounded AI capability as one accountable operating unit without erasing ownership boundaries.
Open record →EXPERIMENTAL LABInteractive experiments for recursive geometry and address behavior that remain separate from performance or compression claims.
Open record →EXPERIMENTAL LABA bounded decoder experiment for inspecting divergent semantic paths without promoting similarity into exact identity.
Open record →A paper argues. A field note narrows. An instrument demonstrates a mechanism. A benchmark defines what must be measured. The evidence ledger records what a result actually supports.
Authored arguments with claim ledgers, source manifests, limitations, revisions, and durable downloads.
Open register →21Narrow technical updates, protocols, observations, and unresolved implementation questions.
Open register →26Executable teaching mechanisms and deterministic fixtures. They are not performance results by default.
Open register →19Named suites, acceptance criteria, example receipts, and the conditions required for a result to count.
Open register →Every stronger statement must resolve to a bounded claim, artifact identity, configuration, result, limitation, and receipt.
