I argue against AGI in this piece: http://everything-will-happen.com/ai/mind/2017/04/03/an-argument…
Nitzan Hermon

Yes, You Argue Such, But There’s No Evidence

Cute, but no sale (and I’ve read all of Brockman’s book and its essays). First, you’ve juxtaposed AGI for ASI in your infinity argument which is explicitly intended, as you say, to be “pro-human.”

The “arguments” about analogies to biology and the brain are self-serving and specious as you know (e.g., planes are not direct analogs to birds). And, of course, you omit common phenomenon in all evolving systems, complexity and nonlinearity.

While true “most” machines are domain specific, that doesn’t preclude machines dedicated to integrating or aggregating multiple domains and or unsupervised learning (e.g., self-driving cars are a multitude of domain specific machines integrated into a single system) in a network.

I could go on (see, Artificial Intelligence (AI) Community Is Playing A Risky Game With Us), but you get the point. Btw, while I’m also pro-human, I’m not sure we warrant praise (see, Our Human Stain, Civilization’s Anti-Human, Not Machines, and Our Twilight Zone & What Comes Next)

Doc Huston