While the article touches good point overall, I have a small comment on the Conclusion part, where it is said, that 'abstraction is about finding "the common" to reuse.

Correct me, this sounds like first we have a lot of "common" spread out all over and then start searching which common things to reuse.

I usually like to "get things right" from the very beginning, which means building enough understanding of the domain in question to at least see which dimensions the code will most likely to grow.

This way the code is much more coherent, because it is built with longer-range understanding.

This means, that the exercise described in the article does not have good answer at all without wider context.

Thanks for this article! It gives valuable insight that "abstraction thinking" is harder to detect as a candidate can't really know the wider context, so the interviewer should not press for one way or another, but probably a discussion why one way and not another would be more fruitful.