Measure the speedup before calling it 100×.
Assign every task once, count handoff cost, protect learning time, and see the real bottleneck across parallel lanes.
Batch and constraints
binary assignment
minimize makespan
minimize makespan
Synthetic task durations illustrate the computation. Replace them with observed focus time; the optimizer does not launch agents or claim general productivity gains.
Synthetic ten-task batch loaded. Ready to optimize.
Solved dispatch
one task placement
per binary variable
per binary variable
No assignment yet.Run the solver to compare sequential focus with the optimized parallel bottleneck.
MEASURED BATCH RESULT
—
SEQUENTIAL—
TIME SAVED—
EFFICIENCY—
LEARNING—
Boundary: The result is a deterministic schedule for supplied task durations. It assumes independent, divisible-by-task work and fixed handoff overhead; it does not measure the source author, Cursor, AI expertise, quality, rework, dependencies, or universal shipping speed.