Long-Horizon AI Work: A Public Reference Architecture for Continuity Without Personas

I’ve seen a lot of discussion here around long-term use of ChatGPT:

  • context drift
  • loss of continuity across sessions
  • scope bleed between projects
  • and unintentional anthropomorphism or authority creep

Most approaches I’ve seen focus on personas or memory tricks.
I’ve been working on a different direction: treating continuity as a deterministic, human-auditable system problem instead of a conversational one.

I’ve published a public reference specification that documents one architectural approach to long-horizon AI continuity using:

  • explicit scope separation
  • human-readable continuity packets
  • deterministic boot order
  • versioned snapshots instead of implicit memory

This is not a product, companion, or live system.
It’s a public architectural snapshot meant to:

  • establish prior art
  • document design principles
  • provide a stable reference point for discussion

The spec is here:
:backhand_index_pointing_right: GitHub - codexoftegridy/Forge_Continuity_Architecture_Public_Spec: Public specification of the Forge Continuity Architecture — a deterministic, human-readable framework for long-term AI narrative continuity. © 2025 MICHY LLC (Maybe I Can Help You)

Important boundaries up front:

  • This framework avoids personas by design
  • It is not therapeutic or advisory
  • It does not imply agency, identity, or consciousness
  • It does not expose private or personal data
  • It is intentionally incomplete as a public reference

I’m posting this to compare notes with others who are:

  • doing long-horizon work with LLMs
  • running into drift or scope bleed
  • trying to keep systems boring, auditable, and safe over time

I’m especially interested in:

  • failure modes others have encountered
  • alternative continuity strategies
  • places where this architecture might break down

If you’re building something similar, I’d love to hear how you’re approaching it.

This topic was automatically closed after 22 hours. New replies are no longer allowed.