What it was for
Scrum solved a real problem: coordinating a team of humans whose individual output was slow, opaque, and hard to forecast. Sprints batched work into inspectable increments. Planning rationed scarce hours. Reviews and retros forced feedback into a cadence that human-speed delivery could not produce naturally. For its constraint, it was honest engineering.
The verdict
DISSOLVES — mostly. The framework's machinery exists to manage human throughput, and that is precisely the constraint that is ending. When the inner loop produces working software in hours, a two-week batch is not a cadence; it is a queue in costume. What survives of Scrum is what it borrowed from deeper traditions: the empirical inspect-and-adapt instinct, which predates it and outlives it.
What changes
Sprint boundaries stop marking delivery and the outer loop's revolutions become the real cadence — set by how fast humans can frame intent and verify evidence, not by the calendar. The Scrum Master role, whose entire purpose is coordinating human throughput, is the single most exposed position on the old org chart. The retro endures, because organisational learning is human work; almost everything else around it does not.
The strongest objection
Scrum gives anxious organisations a shared language and a safe on-ramp, and ripping it out overnight produces chaos, not speed. Conceded — sequencing matters. But a framework kept for comfort should be named as comfort, not as method.
Falsification: this softens to TRANSFORMS the day a team shows its sprint boundaries still ration a genuinely scarce resource rather than a habitual one. The verdict is on the sprint as a delivery batch, never on the inspect-and-adapt instinct underneath it.