When you're in school, deadlines are well-tested because everyone's doing approximately the same work. Unless you have a health crisis or death in the family, you don't have an excuse for a late assignment and the typical 10%/day policy is more than generous. When you're at work, though, 90% of "deadlines" are just someone's opinion and a good 50% are impossible. VC-istan is about exploiting young kids who haven't learned that yet and who would rather stay at work till 3:45 am than miss a "deadline" set by some VP with Shit's Easy Syndrome. See: http://steve-yegge.blogspot.com/2009/04/have-you-ever-legali...
The older programmers are really good at what they do, and they're a lot more flexible than the stereotype gives them credit for. The problem, however, is that they're a threat to the cult. Bring someone on board who does his best work between 8 and 11 (when the kids are just starting to roll in) and goes home at 5:30, ignores the deadlines that don't actually matter (i.e. the ones set by egotistical bosses, not hard deadlines that must be met if at all possible) and suddenly there's a breakdown of the Cluelessness. Couple this also with the Dunning-Kruger effect in reverse; the Clueless young people don't see that the older guy is 4x more efficient than they are (because they're inefficient and don't know it, cf. D-K). All they see is a guy putting in significantly fewer hours. Either the young Clueless will see him as a piker for his shorter work hours, or as a badass for his high effectiveness (making him a legitimacy threat, as the younger engineers see him as the real alpha).
I would also add that there's a culture of extremely harsh age-grading. VC-istan is most concentrated in high-COL areas where you can't raise a family on a (non-financial) programmer's salary; at two you can do it, but two career jobs, plus kids, necessitates maids and daycare and that's expensive and complicated, too. Because the only way to own a house in these cities is to become an executive or financier, you have a lot of 22-year-olds who see themselves being executives-who-code in 5 (!) years-- having managerial control of the division of labor and using it to give themselves the fun stuff, but being full-time programmers-- never mind the fact that that's extremely impractical in almost every organization. Very rarely do executives spend less than 20 hours per week in meetings. A 40-year-old who's been around the block and still programs might tell them as much. That's a problem. To counteract this, the 35-year-old VCs and 30-year-old founders discredit the 40-year-old programmers as "bitter" and he reacts by leaving, just after the one-year cliff so as to take a couple lottery tickets out the door with him because, well, strange things happen.
The short way to say this is that older programmers aren't the problem. The issue is that most of these formulaic startups are founded on exploiting young talent that overestimates the career value of low-end engineering work (because these people don't have the experience to tell what quality work is) and most older people can see through the bullshit. That's dangerous, insofar as it represents the threat of the younger Clueless engineers also learning to see.