Current status? Site Blyg-fied; configuring research agent and its operations.
Clarifying some of my intent for the Blygger Hackathon. The essential idea:
- Convert Kinocomputing to a fully automated Blyg (FAB)
That may involve:
- Site alignment with Blyyger spec
- A research agent (e.g. PI’s Humboldt)
- A protocol for agentic research
- Ties to World Machine theory
- Guidelines for tiny experiments
- Engine to productise experiments
- Paywalls and payment rails for output
- Software factorisation generally
I’ll accomplish (some of) the above by managing the night shift.
Other things that I’m thinking about:
- Utilising software engineering oracles
- Casey Handmer’s metascience audit and other epistemological supply chain hardening
- Blygger’s mathematical notation
- Capacity for non-textual media formats (e.g. diagrams, interactive explainers)
- Subset’s HTTP peer capability providing research materials
- Human-computer interaction patterns for guiding research agents (e.g. intelligence briefs, Principal Investigator x researchers)
- Autonomous productisation of tiny experiments via product risks, archetypical cases and Prime Radiant style prophesies