← agent harnesses Quantum Maze — Stage 2: Fork / Join Harness The Hive

Quantum Maze — Stage 2: the fork / join harness

Real-agent orchestration, replayed. At every junction an agent dissolves fully into one child per unexplored corridor; the first to reach the exit triggers a global halt and the lattice collapses to one route. Two memory models race the same maze.

Shared-hive mechanism

One common blackboard. No agent re-walks a corridor another has mapped — it stops the instant it hits known ground.
Spawned
0
Active
0
Forks
0
Tick
0
Spawn tree (the harness itself)

Blind-worlds metaphor

No shared memory. Each branch is its own world and re-walks ground others already saw — maps merged only after the fact.
Spawned
0
Active
0
Forks
0
Tick
0
Spawn tree (the harness itself)
tick 0
speed
start exit explored (blackboard) collapsed route live agent (color = lineage)

Generated by stage2_harness.py (heuristic brain — navigation mechanical, the four meta-decisions stubbed). The same maze (seed 42, braided so it has loops) is solved under both memory models. Notice the spawn counts: shared prunes redundancy, blind re-walks it. Flip the harness to --brain claude with an API key to replace the stubs with real Claude judgment at the four decision seams — same visualization, real agency.