Research / provenance / retrieval
Knowledge that keeps its context.
Exploring how AI can work with information that remains connected, inspectable, and useful over time.
Shared Object Networking
Bill’s Shared Object Networking research explores reusable knowledge objects and the relationships around them. The work considers persistent context, incremental updates, provenance, and distributed retrieval.
The published papers describe an evolving architecture and research direction.
Read the SON and distributed GraphRAG papers (opens in a new tab)
Questions guiding the work
How can an answer remain connected to its evidence? How can useful context survive between tasks? Where should information remain under local control? These questions inform the experiments and architectures under development.