However, I think Nicholas Carr's points are not completely without value. The problem is if you attach qualitative judgment to it (the changing mind is good / bad, etc.) -- because it's very difficult to judge those things without knowing the future. For instance, take this note from the post:
UPDATE. A reader reminded me that Socrates was worried about the impact of written arguments many centuries ago. People have been wailing about technology spoiling everything for a long time.
Arguably, this 'linear' way of thinking that he argues pre-Internet writing cultivated -- arguably, that has real limitations. If you've ever studied ancient Greek, you might've noticed how anti-linear the language is (because of variable word order, cases, tenses, particles, etc.). In many ways Greek was a much more powerful way of thinking. And in many ways it works better in an oral environment. There is perhaps a reason that the Iliad and the Odyssey were created as oral poetry. As with a lot of archaic lyric and choral poetry that later formed the basis for Greek drama. And similarly, Socratic argument is something like the antithesis of a 'linear' mind (or certainly it complicates it).
So one might argue that linearity was a major problem that prevented, in certain areas, more dynamic, Renaissance-like thought until other languages developed a sufficiently rich and malleable base. Elizabethan English is anything but linear. Shakespeare, the spark and/or a central figure of that, was not linear in the written sense. 'Would that he had blotted a thousand lines...' was the general consensus until people paid better attention to his antilinearity, arguably.
But I mention all of this because it's important to study how different languages and mediums affect how we think. It's unlikely that anyone will figure out the present way of thinking, definitively, during the present. But I admire Nicholas Carr for trying (if I don't admire his judgments).
Perhaps he gets it completely wrong. Perhaps the internet allows us to have a non-linear renaissance of discourse, etc. But being precisely wrong is often half the battle, at least among analytical understandings of how mediums work, etc.
Or take Hacker News. The fact that you can't edit something that you post after a certain amount of time in conjunction with comments being attached a numerical 'value' from up/downmodding -- could dramatically affect what types of things are discussed here. I can understand the rationale (what's said is said -- editing after the fact is potentially disingenuous). However, people in turn may be less willing to take argumentative risks because they fear being locked into a 'trial' of an idea. Maybe this is good (for signal/noise). Maybe this is bad (low discursive risks, low discursive rewards). And I'm sure there are many other factors. But arguably it's worth thinking about all the effects of any type of medium.