Koorye/PCD-real

The project is a static reference implementation for a specific research paper. With only 3 stars and 0 forks after nearly a year (315 days), it has failed to gain any meaningful community traction or developer adoption. From a competitive standpoint, it lacks a moat; the 'Policy Contrastive Decoding' technique is the value, not the specific codebase, which is easily reproducible by any ML engineer reading the paper. Frontier labs (Google DeepMind, Toyota Research Institute, NVIDIA) are the primary developers of the underlying foundation models this project targets; they are more likely to implement their own internal versions of such decoding strategies if they prove effective, rather than adopting this repository. The '6 months' displacement horizon reflects the rapid pace of robotics research, where specific decoding tricks are frequently superseded by new end-to-end architectures or more robust diffusion-based policy representations. There is no evidence of an ecosystem or 'data gravity' here.