To be honest, I see a bit of a contradiction here... what we now call ‘computers’ are far more than initially intended (and I’m not referring to the “‘computer’ used to be a job description”): the whole original theory and assumptions of computer usage was that they would be used to calculate and generally enhance already existing rational methods of consideration. Lambda calculus and Turing Machines and so forth were single devices that executed tasks — not nodes in a network whose only purpose is to enable communication between the users that posses them. Those who pondered them and their implications did not see them as many-to-many communication tools. Largely, this is the source of the disconnect referred to (and the disagreements going on between responders): the utopian premises of computing were based on expectations of rationality because that’s what stand-alone computers are best used for; when they became communication devices we (culturally) manufactured the genre of cyberpunk to accomodate for the implications of this (and other things).

It's pretty hard to get a computer to do what you want. You're at the mercy of whatever software is installed, and most of this software is "stupidity-amplifying" rather than "intelligence-amplifying" -- especially the software that an untrained person is able to pick up and use for their own purposes.

Yes. The same can be said about chisels. When I was small, I tried to use chisel to carve something. I was untrained, so I only hurt myself. The same is with computers. But should we limit access to computers for untrained people? It would make my life much easier, but I still think we shouldn't.

I don't think there is a problem with the former statement at all:

Amplifying stupidity must mean it can solve problems without external intelligence. If it required more intelligence, you'd have a less useful machine.