Josh Lehman explains why calling a decision model from application code can be useful for structured judgments that do not need a conversational responseA decision model selects or scores possible answers for a defined decision.. He describes potential speed and cost advantages while discussing Jev.
Lehman and Graham McBain discuss giving OpenClaw users a configurable decision model and making it available to plugin authorsAn agent plugin extends an AI agent with packaged tools, integrations, skills, or interfaces that are not part of the agent's base capability., so the community can experiment with integrations.
A proposed application is checking whether a message in a multi-person chat actually addresses the agent before it responds. Lehman says the integration was still being developed when the conversation was recorded.
Watch on YouTube
