[southphillyman]

Aren't sites like leetcode partially to blame for this? Years ago asking a candidate to push all zeros to the end of an array , in place, probably eliminated 75% of the pool. Now anyone who adequately prepares can probably solve that optimally in 10 minutes. I remember when asking for optimal solution to isPalindrome was a "bar raiser". I was recently told that I am expected to solve 2 medium-hard questions with clean code and good variable names and test cases in 30 minutes for phone screen. The recruiter emphasized that speed was a factor. The bar will keep increasing as people optimize for the test.

[christiansakai]

I think sites like Leetcode is not to blame, because they merely filling the gap. Btw Leetcode is really good for this sort of things because they actually put questions tagged for the companies based on people who have interviewed. The community there anonymously (obviously) always post and tag questions.

Back to your question. I think personally one (at least me) can blame: 1. The insane salary that these companies give 2. The not so amazing salary that the mid size and startups give, and these days, startups are imploding left and right, low hanging fruits are gone, IPOs are longer, VCs are more greedy, etc.

The money is to blame here, and this is a direct result of tech giants growth being unchecked. If somehow we can level the playing field more, then this DS&A problem will go away.

But if we don't solve the money problem, then no matter how many experiences and senior architect titles you have in your career, if the money doesn't beat FAANG money, then a lot of people will constantly chase FAANG money as the end goal. As a result, a lot of companies that knows this know that it is hard to contest FAANG in terms of compensation, therefore these mid-tier companies and startups are okay living with the reality that one day their employees will jump ship anyway, so we don't need to raise their salary as much and just get a fresh body to replace him/her when that happens.