I disagree. I've interviewed people that failed on our rather easy interview questions but had other redeeming qualities. We ended up hiring these people and nobody has ever turned out well; it would have been better to have no position than to have them. My new rule is, "would I go into business with this person". If not, no hire. Sorry.
When I wrote "Sure, it may be worth it - I certainly think it's worth it", I intended to communicate that I think it's worth putting up with a high number of false negatives to avoid really bad hires... so I think we actually are in complete agreement on this point.
It seems to me that the unspoken issue here is the dichotomy between founder and employee. Not every employee is going to be someone of whom you would say "I would go into business with this person." While I agree that companies should keep standards high, at some point (and I would maintain that point is not very far out on the curve) if you can only hire people you'd form a business with you are going to hit a wall when it comes to growing your company.
I'm not a founder though, I'm employee eight billion at Bank of America. It's just that hiring people that are bad at programming is so demoralizing that it's better to not hire someone unless you're 100% sure they will turn out awesome. Otherwise, you'll spend all your time "helping" them, and neither of you will get any work done. (Nobody ever gets fired because it makes you look bad, or something, so if you make a bad hire, they're here for life. Especially since they are getting paid above market.)
