Yes, I have some technical chops - started coding on a TRS-80 Color Computer II in the 80s and never stopped. It's not something I do much of anymore (sadly) as a growing crop of grey hair has convinced my day job c-levels to shovel me into managing groups of technical people.
I also do lots of work outside of pure development in my day job, and it has wildly different sets of requirements, but the process is more of less the same. I've hired and managed a few more of these types than straight developers. But I've managed to have about the same success with both, or if we did hire duds for some reason (CFOs cousin or whatever), I noted the potential problems in their file and came out reasonably accurate.
The problem with trying to hire the next legend in the field is that more often than not, you are not hiring the next legend in the field no matter how they feel about themselves.
That's why legends are notable, they're unique and rare. If they are really that awesome, they'll become legends in their own time and you'll know about them before their resume hits your inbox. Building a hiring process for legends is fraught with peril. Legends are almost by definition not hire-able around repeatable hiring processes. So companies fake a process. Some companies select for specific schools -- like the too often true stereotype of top finance firms only hiring Harvard/Yale grads or top tech firms only hiring Stanford/MIT types. I've even heard of some hiring with a heavy weight on SAT scores or average KLOC per month!
As much as we all like to think of the lone swashbuckler programmer, coding his way to greatness in the fewest undocumented KLOCs possible, I'm hiring people to build actual things that other programmers of various talent levels from other organizations might have to work on months or years later long after my rockstar has moved on to greener pastures because we didn't offer him the office chair he wanted or whatever. In other words, I'm hiring for the best of typical. I need engineers, not crusading nights. I'm building the Golden Gate Bridge or the Empire State Building, not the Pioneer Space probe or filming and episode of MacGuyver. I've hired the rockstars before, there's a reason organizations tend to eventually move on to other processes.
I've worked with some absolutely unknown solid developers who I'd put up against the best in the field any day of the week, because they produced miracles on time and within budget. Absolutely stunning, elegant, rock solid software. But I'd rather focus on their steady solid performance over time. Rather than have a rockstar come in and dump down a few thousand lines of impenetrable code then flame out. I've found that solid, steady developers will ultimately outperform those guys every time. They'll do it right, and the bridge won't fall down or the building won't fall over.
If you are looking at tremendous number of resumes per week, you have to build a hiring process that selects for the best of typical. If you spot a super star in the pile, by all means talk to them, but you'll eventually find that most of the time they have overinflated egos. Hiring people is almost exactly like American Idol. 90% of all the candidates you see are bad, often spectacularly, often embarrassingly. The worst suffer from some of the most intense Dunning-Kruger effects I've ever seen. But you might find an honest to gosh amazing developer every once in a while and it's worth it to hire them. But don't build a process trying to find them, they'll find you.
If you don't build this kind of selective process, you'll spend literally all of your waking hours interviewing people. If you like this kind of thing, and your company is willing to let you do this, awesome. I have other things to do.
An all day technical interview sounds great, but now do this with 30+ candidates. Bye bye month of November. You need to weed people out and get it down to a manageable pool that you can then focus your time on.
I'm know what I'm saying may sound harsh or overly selective based on some arbitrary metric but it's all there for very good, repeatable reasons. But in practice these are only guidelines. If you went to school, awesome for you, I don't care where, be prepared to talk intelligently about the experience, be proud that you demonstrated the ability to complete a grueling, often mindless multi-year project on your own. If you didn't, but managed to autodidact your way to a great skill set, fantastic, be prepared to show and talk about why that worked better for you than going to school.