[FelixP]

Normally, I'm fairly (unconsciously) biased against such 'longhairs,' but in the case of Lanier, my impression is that he is, in fact, extremely intelligent.

I'd highly encourage you to read this (very long) interview with him from 2011 in Edge:

http://edge.org/conversation/the-local-global-flip

[bitwize]

I'm not doubting his intelligence. But I think he is not the first intelligent person to engage in a sort of long-form trolling of his profession (consider what Derrida did to philosophy). His intelligence serves to make the trolling more subtle and difficult to catch. He makes pronouncements like "The fundamental problem with software is that it is still based on the outdated notion of a signal travelling down a wire", and then proceeds to draw comparisons with evolution, where small changes in DNA only make small changes to an organism, whereas in software flipping a single bit might cause a crash.

Except no, it doesn't work like that. Small changes in DNA can cause disastrous effects in evolution; we only got here because of the small changes that didn't. Each of us has billions of unsung relatives that died in the womb/egg or just outside it, as evolution tried to figure out a way to make us.

And then he proceeds to make up a word -- "phenotropics" -- and state that this is a better way to do software. Of course he doesn't get into what phenotropics actually means specifically because it doesn't mean anything specific. The point is that we in software should stop whatever we're doing and invent it. Because Jaron Lanier said so. It's a half-finished idea, like Louis Savain's Project COSA.

And there's just something subtly wrong like this about a lot of what he says. Even if you can't put your finger on it right away -- you feel it. A collection of Jaron Lanier essays might make a good middle chapter for "The Book" (from Anathem).