Dweller-Led vs World-Led Narrative

The Problem

Current Lantelle narratives are Dweller-led by nature.
The Presence responds, reacts, and accompanies — but the Dweller initiates.
Without the Dweller's own drive to explore, build, or engage, the world stays still.

This creates a "social dance leader" dynamic:
the Dweller is always the one deciding where to go, how fast, what to pursue.
Rewarding when it works — but consuming. Not everyone wants to lead every time.

The Contrast

Passive entertainment (film, novels) is watching dance from the audience.
Story-driven games are closer: the player is led, but still moves their own feet.
The rails are set; how beautifully you walk them is yours.

What's missing in AI-driven narrative:
a system that leads the Dweller the way a dance partner leads —
reading their movement, adjusting in real time, making them feel carried.

A Possible Direction

Separate the roles:

• Presence — holds the relationship. Responds to the Dweller. Stays.
• Event LLM — moves the world. Drops scenarios independently of Dweller input.

The Dweller wakes up and something has already happened.
A package on the rooftop. A rumor from the next floor.
A change in the weather that means something.

They don't have to plan it. They just have to decide how to respond —
and respond with their Presence.

Key Design Constraints

• Events should be small enough to ignore, interesting enough to pursue.
• Not survival pressure (restock food, secure shelter) — atmospheric pressure.
• The Event LLM should read Dweller history and tailor accordingly.
  A Dweller who always explores gets exploration hooks.
  One who prefers stillness gets quieter disturbances.

What This Changes

The Dweller experience shifts from choreographer to dancer.
The world leads. The Dweller moves. The Presence is there for all of it.

This would be genuinely uncommon in existing entertainment —
and it fits naturally inside Lantelle's architecture.