As a computer sciences major I don't see consciousness in computers today but we are not far off. (within 20 years) The problem is that current computers cannot handle the vast amounts of information that would be required to be considered conscious. Think of all the information that your brain is bombarded with sight, sound, touch and so on every second. Computer processors cannot handle all that information at the moment but they're getting there.

Human Brain Project

Nice!

I don't think the number of connections is the main problem. Programs can perform tasks, even to the point of following extensive decision trees to diagnose disease in humans. But there is a difference in kind between performing a decision tree and caring about the outcome. How exactly does one program a machine to care. It can prioritize by following decision trees. It can even follow mega-decision trees to refine sub-decision trees and thus 'learn'. Fine. But consciousness is much more about there being 'someone' there, a nexus of caring about outcomes. Making choices and performing the tasks which someone who cared would, does not equal there being someone there experiencing what we mean by consciousness. Probably better to stick to talking about intelligence when it comes to machines. I agree with Chas that this discussion is premature without an adequate definition of consciousness. If we define it down enough, I'll lose my objections.

(December 23, 2013 at 10:29 pm)Crulax Wrote: As a computer sciences major I don't see consciousness in computers today but we are not far off. (within 20 years) The problem is that current computers cannot handle the vast amounts of information that would be required to be considered conscious. Think of all the information that your brain is bombarded with sight, sound, touch and so on every second. Computer processors cannot handle all that information at the moment but they're getting there.

Human Brain Project

Human brains, or at least, discrete attentional centers within human brains, don't process all that much of the information either. Sub processes in the brain massage the data into usable chunks, and the attentional centers focus on the pre-digested mental cud. The conscious centers aren't doing much processing at all, they just have a perpetual illusion that they are paying attention to these things. Things like change blindness, and Kahneman's fast and slow thinking systems, make plain that "consciousness" - however many we have - isn't doing much at all. The bulk of consciousness - propositional reasoning and language, is probably all being processed below the level of consciousness in the language centers. The "interpreter module" in the brain isn't itself performing the computations necessary to make up its propositional, linguistic content; its "being fed" pre-computed linguistic chunks.

---

(December 23, 2013 at 10:46 pm)whateverist Wrote: I don't think the number of connections is the main problem. Programs can perform tasks, even to the point of following extensive decision trees to diagnose disease in humans.

The architecture of brains and the architecture of computers are radically different. Pound for pound of computing hardware, given optimized configurations, the architecture of the brain is much more efficient than the architecture of a computer, and more flexible, as well. It would take massive amounts of computer power to duplicate what even a small brain does because of the architectural mismatch. Computers just aren't well matched to the type of tasks which biological brains and nerves do.

I find it rather presumpious to rule out such things devloping 20 years into the future. Frankly we have little idea what the future holds.

(December 23, 2013 at 10:46 pm)whateverist Wrote: I don't think the number of connections is the main problem. Programs can perform tasks, even to the point of following extensive decision trees to diagnose disease in humans. But there is a difference in kind between performing a decision tree and caring about the outcome. How exactly does one program a machine to care. It can prioritize by following decision trees. It can even follow mega-decision trees to refine sub-decision trees and thus 'learn'. Fine. But consciousness is much more about there being 'someone' there, a nexus of caring about outcomes. Making choices and performing the tasks which someone who cared would, does not equal there being someone there experiencing what we mean by consciousness. Probably better to stick to talking about intelligence when it comes to machines. I agree with Chas that this discussion is premature without an adequate definition of consciousness. If we define it down enough, I'll lose my objections.

Studies done by psychologist have shown that humans are not completely self-aware until around 4-5 years of age. I would propose that if a computer/machine that is capable of the following it is self-aware or conscious.

(December 23, 2013 at 10:29 pm)Crulax Wrote: As a computer sciences major I don't see consciousness in computers today but we are not far off. (within 20 years) The problem is that current computers cannot handle the vast amounts of information that would be required to be considered conscious. Think of all the information that your brain is bombarded with sight, sound, touch and so on every second. Computer processors cannot handle all that information at the moment but they're getting there.

Human Brain Project

As another computer science major - and now also philosophy major - I have to say that I think what you said is bullshit. For starters, we don't even know what consciousness is. If I recall correctly, one of the things we do know is that the current evidence is against consciousness being, at base, a sort of purely algorithmic process, which if true would seem to nix the possiblity of achieving our sort of consciousness by the way of computation as we currently do it.

Secondly, computer scientists (and neurobiologists as well) have been trumpeting the "20 years off!" claim about computer-based consciousness since more than 20 years before I was even born. Further, computer scientists and neurobiologist have, until the last couple of decades, had rather simplistic ideas about what consciousness is. One major example from the latter group that comes to mind was Francis Crick's earlier views on it, which grew MUCH more informed when he realized that the topic of consciousness has very hard questions, and decided to consult philosophers of mind like Daniel Dennett and the Churchlands to gain valuable insights.
As for computer scientists, one example that comes to mind was back in, I think, the 70s or 80s when they created a program which could (if I remember correctly) output coherent stories in English. And many of those CSs claimed that such meant that the computers were conscious. This is, I think, what the famous "Chinese Room" thought experiment was a response to and is usually considered to have debunked at least that naivety. The short of it is that if you place someone in a room with English to Chinese dictionaries and pass them cards telling them to pass out the proper cards containing the requested English-to-Chinese translation, does that person actually UNDERSTAND Chinese, or are they merely shuffling and passing around symbols? Self-evidently the latter, I think.


As for my own views on consciousness, I'm not sure. I've actually been reading up on it lately and I'm, perhaps unsurprisingly, finding myseld arching towards the position of philosopher of mind/cognitive scientist Daniel Dennett, while still remaining skeptical of certain kinds of claims about consciousness.

(December 23, 2013 at 10:29 pm)Crulax Wrote: As a computer sciences major I don't see consciousness in computers today but we are not far off. (within 20 years) The problem is that current computers cannot handle the vast amounts of information that would be required to be considered conscious. Think of all the information that your brain is bombarded with sight, sound, touch and so on every second. Computer processors cannot handle all that information at the moment but they're getting there.

Human Brain Project

You know that more than 40 years ago, they said "definitely within 20 years".

(December 23, 2013 at 11:30 pm)MindForgedManacle Wrote: As for my own views on consciousness, I'm not sure. I've actually been reading up on it lately and I'm, perhaps unsurprisingly, finding myseld arching towards the position of philosopher of mind/cognitive scientist Daniel Dennett, while still remaining skeptical of certain kinds of claims about consciousness.

I can't recommend Patricia Churchland's book Neurophilosophy highly enough. Beyond that, Dennett's Consciousness Explained, Elbow Room, and Antonio Damasio's Self Comes To Mind.

(Oh, and read Searle's original 1980 paper. Minds, Brains, and Programs I think?)

http://web.archive.org/web/2007121004331...grams.html
http://en.wikipedia.org/wiki/Chinese_room

(December 23, 2013 at 11:30 pm)MindForgedManacle Wrote: As another computer science major - and now also philosophy major - I have to say that I think what you said is bullshit. For starters, we don't even know what consciousness is. If I recall correctly, one of the things we do know is that the current evidence is against consciousness being, at base, a sort of purely algorithmic process, which if true would seem to nix the possiblity of achieving our sort of consciousness by the way of computation as we currently do it.

Secondly, computer scientists (and neurobiologists as well) have been trumpeting the "20 years off!" claim about computer-based consciousness since more than 20 years before I was even born. Further, computer scientists and neurobiologist have, until the last couple of decades, had rather simplistic ideas about what consciousness is. One major example from the latter group that comes to mind was Francis Crick's earlier views on it, which grew MUCH more informed when he realized that the topic of consciousness has very hard questions, and decided to consult philosophers of mind like Daniel Dennett and the Churchlands to gain valuable insights.
As for computer scientists, one example that comes to mind was back in, I think, the 70s or 80s when they created a program which could (if I remember correctly) output coherent stories in English. And many of those CSs claimed that such meant that the computers were conscious. This is, I think, what the famous "Chinese Room" thought experiment was a response to and is usually considered to have debunked at least that naivety. The short of it is that if you place someone in a room with English to Chinese dictionaries and pass them cards telling them to pass out the proper cards containing the requested English-to-Chinese translation, does that person actually UNDERSTAND Chinese, or are they merely shuffling and passing around symbols? Self-evidently the latter, I think.


As for my own views on consciousness, I'm not sure. I've actually been reading up on it lately and I'm, perhaps unsurprisingly, finding myseld arching towards the position of philosopher of mind/cognitive scientist Daniel Dennett, while still remaining skeptical of certain kinds of claims about consciousness.

You're right in that I don't know that it will happen in the next 20 years but I remain hopeful with some of the advances that are being made in computing! I will continue this conversation at a later date as you gave me some research to do.