Interesting theory that maps onto lots of my recent experiences collaborating with people who rely heavily on AI when I don’t use it at all. In one case, a collaborator seems persistently unaware of how much he
doesn’t know in favor of instead regurgitating the same things over and over, courtesy of AI. I get and can appreciated how early bullshitting can help someone cross the threshold into new domains. I myself engage in such strategic bullshitting when venturing into new areas, all while scrambling in the background to learn the new domain and bring myself up to speed. The problem is that AI has imbued my collaborator with such (unearned) confidence that there’s no scrambling in the background to figure things out. He simply doesn’t think he has anything to figure out. And worse, he isn’t curious about what he doesn’t know. It’s frustrating to me from a work perspective (as collaborators on an independent project, this is
not tenable in the long term) and distressing from a large-scale societal perspective. It feels like a combined lack of self-awareness and death of curiosity.
One point of disagreement I have with this article:
> So, how much does intrinsic knowledge matter? What's the difference with a calculator, isn't it the same ? For everyday output, I'd say less and less, and pretending otherwise is just looking backward
The calculator example is always so telling to me because I think people say this without realizing that kids in early math are in fact not allowed to use calculators. The point is to help them learn principles and concepts first by doing the math themselves. Intrinsic knowledge matters because it allows humans to think, develop ideas, and propose solutions. It allows people to actually do the thing that makes us uniquely human — to intellectualize and theorize, to consider and ponder.
That said, I understand that the essay is really just about the business environment and is thus presuming that productive efficiency is what matters above all else. So yes in most business environments, intrinsic knowledge might not matter much. I think this is a grave error though that in the long term will have dire social consequences.