(1) The time cost isn't as high as it seems. If I get as far as a phone screen I'm getting an offer, nearly guaranteed. The flow chart then looks like (a) if I don't get a phone screen, that awful process is no worse than the status quo, (b) if I do, I just find out in the phone screen if the battery of interviews has high latency (an 8hr day is fine, 8hrs over 2 months greatly complicates a job search) and drop out early the 20% of the time that happens, and (c) from there I have an 8hr day to a guaranteed job. Each offer then costs roughly 0.95/0.8~1.2 full interview processes, or 1.2x 4-8 hours.
(2) That time cost is a bit annoying when it comes to competing offers (a single solid day isn't crazily expensive when you're about to get a 40% pay boost, but 5 solid days for a job search is ... 5x as expensive ... fine, but not ideal, and hard when you have finite vacation days). Somewhere around 1/2 of employers don't actually seem to care about producing proof of competing offers though. If they make an offer, you counter that you're worth some fixed XYZ instead (ideally doing enough research to choose the levers they're most likely to accommodate) and will sign immediately if they can make that happen, they'll go above and beyond to agree to your counter-offer. It's a waste of time for the whole industry to require counter-offers in the cases where everyone knows what you're worth (admittedly, when that's not the case, counter-offers are an unfortunate necessity to prove your worth).
(3) The average tech interview is 10-20min solving the techno-babble and 30-50 probing what they know about the company and the team. New companies are very risky, and they work fairly hard to keep damning evidence out of the face of the public. You can mitigate a lot of risk by background-checking the executive team, but a present-day boots-on-the-ground view of things, ideally with the several overlapping/competing views you get from multiple team members, allows you to bounce out of problem situations early.