I don't think that's entirely accurate. We tolerate steep learning curves if the result is an increase to our power. That's why we are power users.
Vim, for example. Some of us put effort into learning it because it's a powerful and efficient editing language that will enable us to easily accomplish hard things.
Caring about the system itself and being willing to put effort into it are what separates power users from normal users. We don't do this because of masochism, we do it because we want to increase our power.
Normal users want the system to just do what they want without any thinking or effort at all, as though it was a highly specialized tool for their exact task rather than a powerful programmable general purpose computer.
Personally I don't care at all about how "less experienced users will tolerate or adapt". This unceasing focus on the wants and needs of normal users frequently hinders us. Developers typically reduce the complexity and feature set of a piece of software in order to turn it into something a normal user can deal with. We want more power, not less.
Normal users don't really matter unless they are directly paying our salary. We should all favor ourselves unless we're getting paid to focus on someone else.
1) Exactly what you wrote: power users don’t even realize they clicked through five things in ten seconds any one of which would have derailed a weaker user, because they’re so used to bad UI that it’s almost invisible to them.
2) Zero care given to what’s most natural for the platform, for cost of development & maintenance, or what’s easiest for the user, when (bad) designers and (bad) “product folks” get involved and are way too into the wrong kinds of “consistency”, especially brand and cross-platform consistency (yo, I’m not sure it’s worth spending extra money fucking up contrast and platform norms on your inputs so your fucking radio buttons are “on brand” or whatever, like, I’m deeply skeptical of the actual business value of that kind of thing, though I can see the PowerPoint presentation value… and that’s why it happens)
Every human has an opinion on practically everything. But has that human put in the effort to justify pushing that specific opinion?
In this case, is the opinionated engineer humble enough to realize that using software in their day to day life does not equal using software in our customer's context?
Neither answer is necessarily wrong, you just need to make a choice.
My mom can use gmail, but she doesn’t even know about its hotkeys and accelerators, or Labs and whatnot
This fallacy is at the heart of the failure of modern software.
Making things work for the median user is almost entirely about defaults and intuitiveness. If everybody is sending messages all the time, there should be a conspicuous button for sending messages.
Making things work for power users is about allowing those defaults to be changed. It's fine if this is five deep in a menu somewhere. It's fine if there is an option for "advanced mode" that opens up a bunch of menus that are otherwise hidden. It's fine if this requires you to write your own filter rules etc., as long as that's available. What's not fine is to make the limited interface the only interface.
Simple things should be easy and complex things should be possible.
I love this.
Building on this thought: When you are starting to build something, you have very limited resources. Focus only on making simple things easy and forget everything else. Once you have product market fit expand into making complex things possible. This applies to 90% of all products.
It's totally fine to have an opinion on UI/UX if it is informed by cognitive psychology and other relevant HCI subfields, in the same way that it's OK for UI/UX to push for designs so long as they are also informed by business and engineering.