Labor is a market exactly as any other: paying the highest price does not always get you the best quality but paying the lowest price always ensures you don't.
In any case, I think always is close enough to fit. Yes, sometimes someone gets something very cheaply but that's usually because of ignorance or urgency from the selling party, deception, etc., etc. In a natural market interaction, if you're paying bottom dollar you are not and cannot be getting the best quality (unless you are in a scenario where the offering is flat: the best offering is also the worst).
Yes: that is what he is saying.
> always is close enough to fit
No.
It is very unlikely to hire top talent at scale, but it is still possible to hire top talent occasionally even for below market rates.
Those who get kick out of these technical challenges are willing to do it for nearly free - we only gotta make sure, their other needs are taken care of.
If you offer $1m per year (which is a ridiculous premise from the get go), you'll maybe get more good candidates (though a lot would assume there was something weird going on and would balk), but you'll certainly also get way more bad candidates, including many who are cunning and able to take up a lot of your recruiting bandwidth, and maybe cunning enough to actually talk their way in - making your problem worse.
It's just too simplistic to assume that throwing money at the problem will solve your talent problems.