Blog
Persona6 writes around the work: Discord community agents, grounded persona graphs, continuity systems, evals, and research that sharpens the product instead of decorating it.
Pipeline proof
A deliberately temporary post that verifies the weekly Persona6 blog workflow can select an unpublished queue item, generate a structured article, mark the item published, commit the output, and deploy the blog.
Evaluation
A practical evaluation framework for AI personas in live products, with checks for evidence coverage, multi-turn consistency, refusal behavior, source traceability, and operator usefulness.
Packaging strategy
Delve AI's pricing shape shows that buyers will consider AI personas as a packaged product. Persona6 should not copy that packaging. Community agents, twin runtime, and API access need separate value meters.
Research lens
Anthropic's PSM is a useful lens for assistants. Persona6 applies a parallel idea to users: grounded persona graphs that stay inspectable and product-useful.
Testing
Why persona graphs are a stronger foundation than prompt-only fake users for testing AI products, onboarding flows, and adaptive systems.
Foundational idea
Why AI chat memory does not equal identity, and why persistent products need persona graphs instead of loose summaries.
Discord use case
Persona6 starts with Discord: read server history, build grounded member personas, and let agents keep conversations moving without generic bot drift.
Core demo
The simplest credible demo for persistent identity: match a fresh anonymous conversation back to the correct persona and explain why.