Now, after layoffs, and the fact that Amazon doesn't hire entry level engineers anymore, the bar is much higher'
We're not telling people to grind Leetcode because we think Leetcode is great, we're telling people to grind Leetcode because this is usually one of the steps you need to do to get in.
I dunno how other companies work, but when it came to Amazon, if you were struggling with a coding problem, and I gave you a hint on the solution, and you ran with it, usually that wasn't a big concern if you could talk about your experience (even during basic college projects), where you showed evidence of being able to deferentially think about things, not just follow patterns that you have been taught.
I do agree that you should know how to do Leetcode and spend time practicing, but grinding them out isn't what I would recommend. The vast majority of Leetcode medium problems are some form of pointer manipulation, linked list (single, trees, or n way like Trie structure), and some O Log N search, which usually involves a sorted data set. The key to learn is how to map a problem into those things, and how to operate on them, not how to solve a specific set of problems.
Its the same reason why ECE/EE people usually have no issue getting coding jobs at FAANG, because they are so used to low level memory operations and optimizations on how to do stuff that they can easily recognize the patterns.
There is still a demand for competent engineers in big companies, the issue is that most people who claim they can code aren't problem solvers in the sense of reasoning through a solution, but rather trying to find something that they have done in the past and shoehorn the current problem into that solution.
This is why even on this thread, people are trying to tell you to grind out Leetcode - the idea is that you memorize enough patterns to be able to solve the problems.