Hacker News new | past | comments | ask | show | jobs | submit login

It's one thing to say "VR will be commonplace in 10 years" or "Self driving cars will be commonplace in 10 years". That, I can believe because we already have prototypes and I've messed with them and could imagine advancing them.

It is quite another thing to claim that human adult level strong AI is coming "very very soon" and could happen overnight.

The latter is just science fiction wishful thinking. I have no reason to believe we will ever have truly thinking, sentient computers, let alone "very very soon." Sure, our Siris and Alexas and whatnot will get better and better at responding to our queries how we want, but that's way different from an adult level human intelligence AI. Machine Learning has limits and will not yield conscious machines anytime soon, if ever.

> I have no reason to believe we will ever have truly thinking, sentient computers

Unless you think human brains work on magical pixie dust, I'd say you have quite a few reasons.

I'm not even convinced that the human brain can be modeled in terms of a Turing machine. How does does the brain achieve free will - the ability to "decide" between two arbitrary, equally-weighted things? RNG? And yet, it doesn't "feel" random.

Firstly, free will isn't what you think it is. A random choice might seem "free" but it's not "willed" in any meaningful sense. So let's leave aside the free will question because that term is pretty much undefined at this point.

What your post sort of hints is what's known as the hard problem of consciousness. I recommend reading the Wikipedia pages on it and qualia if you're interested.

Suffice it to say, given our current understanding of physics, we are no better than finite state automata (see the Bekenstein Bound). The only escape from this inevitability is if we collectively decide that the hard problem of consciousness is irreducible, and then something like panpsychism becomes preferable.

This is unlikely though, and we've been through this once before in the debate over how living matter differs from non-living matter. Must there be some "secret sauce" added to non-living matter to bring it to life? This was the proposal of vitalism, but eventually biology came to prominence and all of those who insisted living matter had to be different just died off and we were left once again with a reducible, mechanistic understanding of living matter. So it will be with consciousness (see [1] for an example of how this might work).

[1] http://journal.frontiersin.org/article/10.3389/fpsyg.2015.00...

This presumes free will, which is itself contentious. I think it's clear we have the appearance of free will, but it's not clear that consciousness actually makes decisions rather than an emergent appearance of agency, claiming decisions as made for actions that are already in process. If you're going to make a claim about the feasibility of AI that relies on free will, you'll have to prove free will first.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact