Interviewing is hard, requires skills precisely zero people are trained to do, and has huge long-term upsides and downsides to a business.
I'm a programmer. I trained for years to do that. Then I'm asked to hire someone. A task I have zero experience or training at. Spoiler alert - any actual competence is a fluke.
At my peak I was hiring maybe 2 or 3 people in a year. Mostly I hired people I'd heard about who came with a recommendation. Results varied. Most were OK, some were exceptional, some didn't last long.
Some companies are hiring tens, hundreds or thousands of people a week. Presumably they have full-time interviewers. But "good judgement" isn't a scaleable property, so you end up with filters and processes to try and narrow the field to "good candidates". Maybe thats "must have degree", maybe that's leetcode problems.
Hiring into a big company is painful. They have a process which probably makes no sense, adds little value, and appears to be arbitrary and doesn't lead to significantly different outcomes. But how else could it look? It's a sausage factory churning out sausages.
Hiring into a small company can be turgid - they're a small company but it seems like everyone there has to sign off on the interview first. They're scared of making a mistake, so they never seem to get to a decision.
Or it can be simple. There's one guy. He either likes you or not. He either thinks you can do the job or not. Takes an hour, and there's an offer on the table. But this guy hires 1 person every now and then, not 100 per week.
Hiring us hard [1]. It should be done by professionals. There should be degree programs training people how to do it. But even then, companies growing at severe pace may as well just pick every 3rd CV and save themselves, and applicants, a lot of time and effort. Yes it would be random, but you shouldn't hire unlucky people anyway.
[1] hiring for skilled people, where actual skills are hard to measure, to work in a team environment, where their output is disproportional to their rank, status of pay.