Competence is most absolutely not one spectrum. Competence is measured in dozens (hundreds, really) different axis. Everyone will have different scores (if we could realiably narrow it to a score, which we can't) on different axis of skills. Which of the skill axis are most relevant for any given role will vary, obviously. A generalist will have decent scores in many, a specialist may have mediocre scores in many but 99+percentile in their chosen areas.
Leetcode interviewing measures along one single uninteresting axis, the one corresponding to memorization of algorithm puzzles. That axis happens to be entirely irrelevant to any job I've ever hired for. So I don't test for that because I care as much for your skill doing leetcode as I care for your skill juggling frogs. Neither is relevant to the job.
Have you considered that with a strong understanding of elementary algorithms and data structures it is possible "solve" rather than "memorize" these algorithm puzzles?
If that were even remotely possible, why isn't that person cranking out multiple novel algorithm research papers per day? They should be outpublishing Knuth, for sure.
So, no. The only road is to memorize extensively and practice acting skills on the delivery.
EDIT: Obviously even some of the most elementary Algo 101 problems were research questions at some point, and led to published papers, but that's the reason people write textbooks and use them in college, so newcomers don't have to redo the foundational work.
You're right in that HR of their previous (likely current) employes won't say or allow saying anything.
But of course they're likely not giving as references their current boss, for obvious reasons.
Their previous bosses though, who have also left that company, will speak to you freely.
What could be niche about this? It's the most overwhelmingly common case.
Most people change jobs every couple years and your interviewees ex-bosses are also in the industry doing the same job hopping. It is a near certainty that nearly all (if not all) their ex-bosses have moved to different companies already.
This is the most common case, by far.
Are you suggesting going further back, to the candidate's previous previous employer, and hoping that their direct manager from years ago is able to provide reliable and accurate testimony on the current state of the candidate ?
I'm surprised this is not a universal experience based on these responses. Every candidate I've interviewed has given at least one ex-boss as a reference. Every reference I've ever given has been ex-bosses.
I've previously hired contractors with minimal interviewing because they came from a trusted partner to my organization, so I asked the partner about their productivity and skills and they were given good reviews relative to what we were trying to accomplish.
Each of them were a net negative with respect to the project. The time it took me to review their pull requests and help them write code which was passable was longer than it would have taken me to write the code myself. Additionally, the code that was passable has been a constant source of bugs.
A more thorough technical interview may have avoided hiring them. In retrospect, the correct course of action would have been to not hire anyone until someone qualified was available. This anecdote does support the grand-parent posters theory that operating with fewer incompetent people is better.
Additionally, their proposal is generally workable (which again, yours is not).
Competence is a spectrum and so at some level of competence there will be people who are competent but unable to pass overly complex interview questions. The fact that this overlap exists seems to be what you are complaining about. Even still, not hiring people in this overlap is probably safer than hiring people which make the system worse in my experience.