What is the 'sunk cost fallacy'? Is it ever a good thing?

This is such an odd title to an article. Is the fallacy ever a good thing? The fallacy itself is a concept - so not really good or bad. Using it as a logical premise in an argument is recognized to be problematic.

Are we actually asking: is ever good to keep doing something you yourself hate only because you yourself already spent money on it? The answer seems clearly to be no.

But be warned, this doesn’t work for everything: it seems that spending wildly on a wedding ceremony or engagement ring doesn’t have a “sunk cost” effect – it fails to increase the likelihood of staying married.

LOL

IMO, it is a fallacy because the mind couldn't distinguish if the past it is trying to compare correlates to the current situation.

It is similar to "what-about-ism" but with prices instead of other values being measured.

is it... a good thing?

there's some very situational cases where you'd actually want to fool yourself to fight a bigger issue (like laziness/ procrastination). that's why some people advise goes in a pattern like "start small" [if trying to start a new project] or "just write something" [for writer's block].

I remember an episode of EconTalk that discussed the sunk cost fallacy. It was Mary Hirschfeld on Economics, Culture, and Aquinas and the Market. I'll just quote the transcript at length:

Mary Hirschfeld: Okay. But first, I just--I love the observation that in Chapter 1 of an economics textbook, they'll say, 'We're just describing people as they are. People are rational, and this is how it goes.' And then, in Chapter 2, they'll say, 'And most people fail to'--you know, 'respect the sunk costs.' And they never even notice that there's a small tension there.

But, anyway: Now, there's some places where you actually do want to ignore your sunk costs. If I'm playing you at poker and I spent a lot of money bidding up the pot, and then on the last card it turns out that I really am not going to win, I should not follow up my losses with more betting, if I think I can con you out of it.

But, there's a good--there's plenty of places where you really should ignore your sunk costs. But, if you think about it as a practice in general, if I'm always saying to myself, 'Oh, well, whatever it was that happened yesterday is a sunk cost. I have to decide today,' that means I'm always going to be emphasizing what are my feelings in the moment. They are just going to have an elevated sense of importance in my decision-making. And, the problem is, my feelings in the moment might often be impulsive or not wise. And, part of self-discipline--part of growing in self-discipline and virtue, is learning that 'Well, sure, I might like the donut today; but in the long run, I might like to cultivate the habit of not wanting donuts, because it would be better for my health.'

And, so--I mean, there's classic places where this happens. If I want to start exercising more, I might go down to the gym and pay for a membership fee. And then if I have the idea that I should honor my sunk costs, I'm more likely to act on that higher impulse and get to the gym. But, if I do what the economists says, on the day when I'm thinking I don't want to go to the gym because I'm tired, and then my little quote-unquote "irrational" voice says 'You paid that money for the membership; you should go,' well, the economist would say, 'No. Ignore that voice. That's a sunk cost. If you don't feel like going, don't go.' And then I never get this chance to discipline myself.

Now, economists certainly have models that allow you to try to do this self-committing kind of behavior. They have complicated models that try to handle these kinds of situations. I just think it's a lot more common than they think. And I think the main reason I think that is actually personal. I really was formed as an economist. I started grad school at the age of 20, 21, whatever I was--it was young. And I always found myself thinking on the margin, ignoring my sunk costs, and not developing anything remotely like self-discipline. So, I'm really--and I'm still struggling with this. The legacy of this is still in my being.

tl;dr: According to Hirschfeld, employing the sunk cost fallacy can help us more heavily weigh our past decisions to inform how we should act in the present. We may have made decisions yesterday in accordance with some higher ideal of ourselves, and perhaps we shouldn't discount those decisions entirely just because they're in the past.