There's also the assumption that all these companies have a burning need for exceptional talent, which is nonsense. Most companies are not doing anything exceptional, and there'd be no reason to pay exceptional programmer wages when you could pay competent programmer wages.
99% of startups do not need someone to "invent things an ordinary programmer would never even think of." They need someone to write bog-standard uninteresting[1] Javascript/RoR/whatever for their webapp which, to be sure, might be making our lives better but isn't doing anything technically special.
To me the "handful" of consulting firms he mentions is massively downplayed. Most H1Bs, however talented, are being sucked up by those firms or by firms engaging in similar strategies, not startups. PG's goal would be just as well accomplished by focusing on the crackdown than opening the H1B gates wider - if we do that, then the usual suspects will simply and gladly suck up proportionally more.
[1]EDIT: Uninteresting is probably not quite the right word. Even well-understood, solved problems can and are really interesting to some of us. But you don't need to find them interesting to re-solve them efficiently and competently.
99% of startups do not need someone to "invent things an ordinary programmer would never even think of." They need someone to write bog-standard uninteresting[1] Javascript/RoR/whatever for their webapp which, to be sure, might be making our lives better but isn't doing anything technically special.
To me the "handful" of consulting firms he mentions is massively downplayed. Most H1Bs, however talented, are being sucked up by those firms or by firms engaging in similar strategies, not startups. PG's goal would be just as well accomplished by focusing on the crackdown than opening the H1B gates wider - if we do that, then the usual suspects will simply and gladly suck up proportionally more.
[1]EDIT: Uninteresting is probably not quite the right word. Even well-understood, solved problems can and are really interesting to some of us. But you don't need to find them interesting to re-solve them efficiently and competently.