Confidence: medium. Evidence: strong gains in selected regimes; weak standardized accounting. Last substantive change: 2026-07.
This subsystem owns the money and the capacity: what the factory actually costs, where the bottleneck sits, and how to spend a fixed budget of compute and human attention.
The conclusion
Optimize the whole queue for accepted, durable outcomes and for scarce human attention. Parallelism is valuable only until validation, integration, infrastructure, or operational capacity becomes the bottleneck, at which point adding agents makes things worse. The right denominator is cost per accepted, durable outcome, not tokens, lines, or raw task success.
How the thinking got here
Early framing counted tokens and lines of code. That gave way to per-task cost, then to cost per accepted change, then to attention economics, review backpressure, and downstream maintenance burden. The recurring surprise is that apparent technical deflation can coexist with rising verification and coordination cost, so a cheaper token does not always mean a cheaper outcome.
Credible alternatives, and when each is right
| Approach | Right when |
|---|---|
| Frontier-model abundance | quality dominates and budget is ample |
| Cheap-model cascades | most work is easy, escalate the rest |
| Hard quotas | preventing runaway spend |
| Value-based budgets | tying spend to expected outcome value |
| Capacity-aware schedulers | the bottleneck shifts across the queue |
Where it fails and what we still don't know
Failures include a Jevons effect where cheaper generation increases total spend, and review queues that collapse under throughput. Evidence shows dramatic production gains in selected regimes but weak standardized accounting across retries, review, maintenance, and incidents. Open questions include full-cost accounting, marginal value curves, cost attribution, energy and carbon, and build-versus-buy.
What would change our mind
A standard full-cost accounting that includes validation, review, maintenance, and incidents would let factories compare themselves honestly rather than on token price.