[fuligo]

  Computers are very, very, very far from being like humans, especially 
  when it comes to consciousness. The problem is different, that the system, 
  the military and economic and political system doesn't really need 
  consciousness.

Consciousness is a poorly defined concept with many unfortunate connotations, but let's assume in this context it's the same as self-awareness. I assert it's very likely that nature didn't evolve conscious brains by accident, it's probably a byproduct of making an intelligence that can reason about itself and its environment.

I know it's just a thesis, but when you think about what our mindless AIs lack, it makes sense. They're characterized by a complete incapability for global reasoning and an inability for personal consideration. You might argue, as the article does, this is exactly how we want our tools to behave, but then we might have to accept there could be hard limits on the complexity of mental tasks these systems are able to perform without access to higher reasoning.

[woodchuck64]

> I assert it's very likely that nature didn't evolve conscious brains by accident, it's probably a byproduct of making an intelligence that can reason about itself and its environment.

I think you're exactly right; Michael Graziano's theory of consciousness it that it starts as a necessary function for modeling the attention of an agent and turns into awareness when the brain itself is modeled as an agent. First, a specialized brain function for agent modeling of predator/prey/rival/mate evolves as a good trait for survival. Part of that function entails modeling what an agent is attentive to. However, once this agent model discovers the brain it is running on, a new agent is recognized, and attention now becomes redundant and is instead reported as awareness of attention, which has to feel subjectively like a secondary aspect of primary sensory information. It's a nice logical progression. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3223025/

[shasta]

"Self awareness is a byproduct of an intelligence that can reason about itself" looks like a tautology to me. I've heard similar theories proposed that extend to broader notions of consciousness, but I don't find the argument very compelling. It's very easy to see why self-awareness is required in nature: intelligence won't do any good for an organism unless it is self aware and has a survival instinct. That pressure clearly doesn't apply to man made machines.

[AndrewKemendo]

>it's probably a byproduct of making an intelligence that can reason about itself and its environment.

I think this is exactly right. In fact my suspicion is that "consciousness" is an emergent property of feedback from high resolution sensors. Stated another way, it's a constant internal inventory of everything that can be controlled through the same volitional system.

[rivd]

The word "Consciousness" is almost like the word "God".

Many people seem to know what it is, no one can actually define it, and it has never been measured, located or proven to exist.

Maybe we should stop using it altogether.

[TeMPOraL]

I wouldn't go as far as stopping to use it altogether, but when discussing things it might be worth a while to taboo it [0]. It's a very helpful trick; [1] explains it nicely.

[0] - http://wiki.lesswrong.com/wiki/Rationalist_taboo

[1] - http://lesswrong.com/lw/nu/taboo_your_words/

[ForHackernews]

We should either stop using it, or at the very least own up to what it is we're [trying to] talk about: http://consc.net/papers/facing.html

[toomuchtodo]

I wouldn't go that far. Consciousness is the result of algorithms properly functioning in my brain chemistry which allows me self-awareness.

Our brains process information in a certain way due to evolution, therefore, we are.

[agumonkey]

I root all these things into survival. So far computers are 2 years old, without humans to lay infrastructure, renew, fix computers would stop functioning pretty fast.

[nova]

> it's probably a byproduct of making an intelligence that can reason about itself and its environment.

See Jaynes' bicameral mind hypothesis. Even if it's most likely a mistaken theory it's still interesting as a distinction between intelligence and consciousness.

[ForHackernews]

> it's very likely that nature didn't evolve conscious brains by accident, it's probably a byproduct of making an intelligence that can reason about itself and its environment.

Strictly speaking, everything in living nature is an "accident". Natural selection acts on "accidental" random mutations. It's not a directed process.

[fuligo]

I'm not sure the original comment deserves this clarification, you'll be hard pressed to find someone on this forum who didn't already know about the nature of mutations.

The point is though that while the mutations themselves are "accidents" and the result is always characterized by a certain randomness, evolution as a problem-solving algorithm isn't itself accidental. While some (or even many) if the characteristics of an organism might be incidental, some major features tend to have a good reason for being selected. Self-awareness, I assert, is such a feature, because it carries implications too large not to have an effect on selection.

[Nogwater]

The mutations are random, but the selection isn't... it's governed by the environment.

[aikah]

You realize that randomness and "accident" are man made concepts, they don't exist in nature. Something we can't predict isn't random at all, with enough data one could forsee anything, even the fact that you were about to write that message and the exact words I would use to answer you.

It's beyond science, but I don't believe in randomness or chaos, which doesn't mean I believe in religion either( which are just a collection of myth, which says absolutely nothing about the fundamental nature of 'god').