Spending a few weeks on exploratory or speculative work is not really a huge time commitment. If it was going to drag on for months I would question it.
I also challenge the “deliver in small PRs” idea. Yes in practice that would be ideal. But often when you are doing something completely off the map there is no point to submitting the PRs for the 5 things that didn’t work. The author is falling into the same trap that scrum advocates and agilists fall into. Anything can be broken down into teeny tiny bit size deliverables and managed that way. Again, it would be ideal if that were the case, but sometimes you just need to give your best engineers some time and space to swing for the fences.
I also challenge the “deliver in small PRs” idea. Yes in practice that would be ideal. But often when you are doing something completely off the map there is no point to submitting the PRs for the 5 things that didn’t work. The author is falling into the same trap that scrum advocates and agilists fall into. Anything can be broken down into teeny tiny bit size deliverables and managed that way. Again, it would be ideal if that were the case, but sometimes you just need to give your best engineers some time and space to swing for the fences.