The idea
Record an earlier attempt, then cooperate with a translucent replaying avatar that repeats its actions alongside the current run.
How it works
- Place virtual switches and objects in a room using persistent spatial anchors.
- Record a run as a deterministic sequence of player actions.
- Replay the past self as a translucent avatar while a friend or current player completes another step.
AI’s role: Help explain puzzle actions and parse player intent; deterministic timelines and puzzle state replay recorded actions rather than guessing them.
First demonstration
Make a room with three virtual switches. Record two runs and use simple translucent avatars to press switches in a coordinated sequence.
What to solve next
How can replayed actions stay spatially and temporally clear if the room or tracked anchors change?
