This is a bit of a nuanced topic with lots of shades of truth or context, and overall I agree with the article, but that said:
Sometimes you know your team is going to do the stupid shortcut if you let them and you save them from themselves by solving the robust solution before they have a chance to "be pragmatic" and pointlessly bikeshed about things that don't matter to waste your time. Sometimes you know it will take you a week to do it right or a week with input to do it half right and you just bite the bullet and do it right the first time and, to the articles point, suffer the consequences if or when you get it wrong or partially wrong.
Realpolitik is part of being a senior engineer. Sometimes you make the call to take the shortcut without mentioning it, sometimes you force the full solution with the big PR (or, even worse, a series of staged small PRs to make it look like you're taking feedback); but most of the time, if you're working somewhere good, you get to be completely transparent about what you're working on and the feedback you get is corrective and flexible.
I have found that the attitude of "just silently do the robust and proper thing immediately and push it, instead of letting the entire team discuss this important infrastructure topic forever in order to ultimately come up with an abomination of design-by-committee-architecture, just to then find out that the remaining time now is way too short to implement it, which is why they then take all the stupid shortcuts to get it done in time, ending up with something barely holding together with duct tape that somehow works for the moment, but immediately tumbles down as soon as the next guy tries to build the next layer on top of it" is a very important senior dev skill, especially at the beginning of new projects when a lot of basic building blocks on which the more complex features have to rest later are still being created.
The reality is that it's not your call, it's the customer's call, and you need to stop trying to make a decision for the customer. Let them tell you what they want.
Sometimes you know your team is going to do the stupid shortcut if you let them and you save them from themselves by solving the robust solution before they have a chance to "be pragmatic" and pointlessly bikeshed about things that don't matter to waste your time. Sometimes you know it will take you a week to do it right or a week with input to do it half right and you just bite the bullet and do it right the first time and, to the articles point, suffer the consequences if or when you get it wrong or partially wrong.
Realpolitik is part of being a senior engineer. Sometimes you make the call to take the shortcut without mentioning it, sometimes you force the full solution with the big PR (or, even worse, a series of staged small PRs to make it look like you're taking feedback); but most of the time, if you're working somewhere good, you get to be completely transparent about what you're working on and the feedback you get is corrective and flexible.