For this scenario to unfold the following has to happen:
1. OS is set to auto mode 2. It's late afternoon 3. User finds they have (for some reason) in the past turned on the Website dark mode (presumably in the daytime) 4. User wants it light now.
This is a pretty tortured hypothetical though. I admit as an engineer, who understands that there's a light/dark signal being passed down, and there's an ability to override that signal, I appreciate separating the "should override" control the way a "3-way" switch allows.
But I have to admit, I cannot imagine a real person getting bit in any way by the implementation the author describes. 90% of people set their OS to the "mode" they want, whether it's full-time or sun-based, and that group never even needs to touch or look at this control (We do default to auto, right?). The other 10% are most likely here because they don't understand they can set this in the OS and they apparently want to toggle to the opposite mode. They might be thinking "I want this dark mode until morning" but "we" in the website actually can't really help them with that unless we start prying into what time it is there and build our own time-based mode. So the best we can do for that user is to shift them into "latched dark mode" and let them come back (as they've demonstrated a knowledge to do) when and if they want the 'lights' back on.
I'm the first to criticize UI designers who rob users of choice in the name of minimalism -- the Alan Dye approach, but this actually screams "Designer thought hard about the actual intent of the users" to me and I like it.
If I'm setting this up during the day, when I normally work, those external sites would easily end up with the "auto" behavior by this very poor design in the blog, when that is not my actual intent. This would result in me having to set it twice (the second time in the evening) to have my desired behavior which is a nuisance. The explicit option is better here, not whatever this silliness is in the article. [0]
At least I've read this article so sites and apps using this stupid option wouldn't be too big of a surprise to me, I'd hate to have to explain this to other people who haven't read it though. Principle of least surprise is a good thing, not something to deliberately work against like this author wants.
EDIT:
[0] Actually, this demonstrates a crucial flaw in the system's design. If I want to set the current default as the permanent setting, I have to wait until a particular time of day to do so. That's even worse than I'd previously realized. You can only set permanent light in the evening (or by changing your OS/browser settings, changing the site, and then changing OS/browser back). You can only set permanent dark during the day (without that same stupid exercise).
This is a terrible design concept.
There is no way any one with any common sense can think that this two-choice option (one is magic, one does what you would expect; which is which depends on time of day for many users) is better than a three-choice option where each choice does exactly what you'd expect.
In my experience, software that is designed to work only for the most average user ends up feeling terrible for most everybody, because most people deviate from the average in one way or another, at one time or another.
It feels especially bad when the software makes assumptions about what the user wants, to the point of including a button that doesn't always do the same thing when pressed, and which doesn't even tell the user it doesn't always do the same thing when pressed.
Just look at all the complaints Microsoft gets for the way Windows constantly reverts or ignores user settings.
Or the way it feels to google a very specific set of search terms, only to receive back 20 completely irrelevant listicles that answer a statistically more likely question you didn't ask.