This seems like it is tapping into the same risk management strategy as in Agile methods? Essentially allowing for more frequent course correction. I assume "small" here blurs together cost and latency .

The tradeoff of this kind of incremental planning and execution is that it becomes more reactive and myopic. You can end up stuck at a local maxima or worse, just executing a random walk.

I think a large part of becoming "quick" in an effective way is to improve your triage skills. This is a meta-decision process where you quickly estimate the time-dependent risks and priorities.