I did one of these interviews on short enough notice no review was possible. The process was a little jarring but on these type questions I could just talk through the problem space and come up with something reasonable and explain it; although I doubt I answered anything precisely as expected in the half dozen interviews I had. Apparently it went well.
To be fair, I wasn't a fresh college grad, and I had been recently working as a full time researcher. But they sent me through the SE interview process and none of the specific questions were close to anything I had thought about in years.
I think most people cannot ever understand these topics well, and from their perspective, it'll always be about cramming. -- It doesn't fit in their world view that others can solve these problems by thinking, not cramming, when they cannot?
Whilst to you, it's more about thinking, coming up with new things, but trying to learn it all by heart would seem a bit weird? -- And then maybe in a sense you're talking past each other?
No, you absolutely do need to cram to get through the interviews. In fact the questions on leetcode are designed for this very purpose. Every question has one common theme. And one 'catch' that you absolutely need to know to solve them.
If you haven't looked at a question in X months chances are you will struggle to solve it quickly even if you have solved it before.
If you are out of a job, just buy leetcode premium. Keep doing it for 15 hours a day. Then start applying for interviews.
Its the dumbest way to hire and get hired.
Not even least surprised people can't hire a person to get a darn thing done.
I just described a situation where that was not true, so I don't really know what you are responding to. If you are generally on top of problem solving, you don't need to "cram".
This generalizes - "cramming" has origins for formal courses, as a technique to try and pass an exam when you haven't actually done the work and don't understand the subject. It's a last-minute approach; might get you a pass but is hardly ideal.
I approach it more as learning different techniques to solve different problems, rather than learning how to identify the trick and recall the solution to specific cues.
Trying to memorize solutions isn't really effective for interview problems, in my experience, unless you really can dedicate time to cramming as many solutions as possible right before an interview and hoping that they ask you one you've seen before.
The problem with this approach is that you won't be able to remember the details of problems older than X months ago, at least not to the level that will reliably help when in a stressful whiteboarding interview.
There might be entertainment/knowledge benefits, but it doesn't replace cramming for a known upcoming interview-event.