So the author of this has a vested interest in their "solution" being the correct one, but I fundamentally disagree with it.

I agree with their hypothesis that we are bad at planning/estimating, but the solution is not spending more time on it, but rather less.

Firstly I really like this humorous look at the problem: https://www.quora.com/Why-are-software-development-task-esti...

You cannot estimate what you don't know, and you don't know what you don't know and no amount of upfront planning will surface those. In their own example, you can't know upfront that there is a road closed or an accident or that only one cashier is working until you get to that point.

In my opinion the most reliable solution is to break the work down into small pieces that deliver value (ideally less than a week's work). Prioritise and then deliver the first piece. Have regular reviews/checkpoints with stakeholders to decide whether any value was delivered, are there new learnings that we need to apply to the rest of the project, or indeed new 'pieces' that we've discovered now need doing, what is the next piece we need to do and is it worth continuing.