THE KARDASHEV PROJECT // SI
DIVISION I // SUPERINTELLIGENCE (SI)

Superintelligence Research Lab.

The Superintelligence (SI) Division conducts fundamental research and systems engineering across domains required to build superintelligence. Selected papers and open-source releases are published below.

RESEARCH PUBLICATIONS IDENTIFIER // ACCESS
OCT 15, 2026
SI-PUB-001
DOI: 10.5281/zenodo.23391160

OpenCommons: A Distributed, Meaning-Indexed Concept Commons for Accelerating Scientific and Algorithmic Discovery in AI Agents

A decentralized, behavior-indexed concept commons that unifies equivalent scientific programs via sub-millisecond canonical execution trajectories, quarantines subtle numerical defects via a 5-gate asynchronous peer-review engine and monotonic counterexample ratchet, and reduces multi-wave discovery token cost by 12.67×.

OPEN-SOURCE INFRASTRUCTURE REPOSITORY
SYS-SI-01

opencommons

Reference Python engine, join-semilattice CRDT store, and minimal 2-tool JSON-RPC 2.0 Model Context Protocol (`opencommons-mcp`) server.

SYS-SI-02

kardashev-cli

Kardashev CLI (`kardashev` / `ksh`): Local and open-weight coding agent harness (vLLM, Ollama, SGLang, LiteRT, Gemma 4, Qwen) with AST repo-mapping, self-healing edits, and distributed OpenCommons CRDT gossip (`--with-opencommons`).

SYS-SI-03

chronosml

Deterministic ML runtime, memory, and throughput estimation engine for autonomous training and inference agents.