McGinn's tirade against anthropomorphic language is also silly. It's obviously something he's trotted out from previous writings, which is okay if it were insightful. Can't he mentally translate "a neuron shouts when it sees a pattern" into "a neuron fires when a certain combination of it's input terminals are excited" or perhaps instead of "fire" which is potentially anthropomorphic itself, Kurzweil should write a script to replace "fire" with "the voltage-gated ion channels embedded in the plasma membrane will open allowing an inward flow of sodium ions, etc. etc". Please, where does it stop. Look beyond the words to what the person is actually saying.

It's similar to his criticism of Kurzweil's "law" of accelerating change. Sure, it should be called a theory not a law, big deal.

Kurzweil is a huckster, he is selling books, he's going to overstate things. Just make the mental translation law->theory in your mind and move on with your life.

McGinn meanwhile has no opinion on what is actually going to happen. He says accelerating change is "just a fortunate historical fact about the twentieth century". So is he saying it has stopped now, 13 years in to the 21st century? I don't believe it has. Or that it will stop? When will it stop McGinn, in 2015, 2029, 2045, 2387? Because this matters a lot in terms of how much computer power we end up with. He has no opinion at all what is going to happen, he only cares what we call things.

While there's a lot to criticize Re: Kurzweil and his claims, McGinn's dismissal of the pattern recognition argument is just wrong. It does stem from an obvious misconception, though.

If you've worked with artificial neural networks, it's clear that Kurzweil is talking about patterns of activation, not sensory stimuli.

McGinn's tirade is not again antropomorphic language per se, it is against homunculism. The idea that a human mind is composed by agents (homunculi) that exhibit mental properties leaves the mental properties of those agents unexplained. That's the reason he proceeds to argue agains the use of "information processing", because he seems to find that also explanatorily lacking.

> It's similar to his criticism of Kurzweil's "law" of accelerating change. Sure, it should be called a theory not a law, big deal.

It should be called a generalization, or maybe an hypothesis, it isn't even a theory. McGin's question is: is this generalization valid? Can we say? If he sees fault in Kurzweil's evidence for it, why not say it? Why would he have to provide evidence as to when Kurzweil's prediction will fail? The burden of proof is not on him.

The burden of proof is not on McGinn, but he is a smart guy, he should have an opinion and should share it. His opinion is only on what Kurzweil's calls the trend, not on the trend itself. This would be like discussing the current budget crisis, and suggesting it be called a budget debacle instead of a crisis, and offering no further insight.

The mind has mental properties because it is composed of agents (homuniculi) that exhibit a certain kind of mental, or at the very least, not-physical property (non-physical in the sense that they don't belogn to a physical theory).

For example, one could try to explain mentality in terms of modules of the brain that are information processing agents. But then, how do we explain "information processing" in those agents? Btw, those agents need not be neurons, it could also be sets of them, of something else (they don't even need to be merely parts of the brain). It's true that McGinn is not very tolerant of Kurzweil loose way of talking, but I think it's because he regards the conceptual issues as quite important. He is a philosopher after all, his standards for accepting bullshit can be quite different.

> even talk of “pattern recognition” by neurons is already far too homunculus-like for comfort: people (and animals) recognize patterns—neurons don’t. Neurons simply emit electrical impulses when caused to do so by impinging stimuli; they don’t recognize anything in the literal sense.

The paraphrases you propose for Kurzweil's way of talking seem to remove homunculism from it, but maybe only partially. I think that can be McGinn's worry, and the reason he goes agains the idea of "information processing". See:

> Even in sober neuroscience textbooks we are routinely told that bits of the brain “process information,” “send signals,” and “receive messages”—as if this were as uncontroversial as electrical and chemical processes occurring in the brain. We need to scrutinize such talk with care. Why exactly is it thought that the brain can be described in these ways? It is a collection of biological cells like any bodily organ, much like the liver or the heart, which are not apt to be described in informational terms. It can hardly be claimed that we have observed information transmission in the brain, as we have observed certain chemicals; this is a purely theoretical description of what is going on. So what is the basis for the theory?

And then:

> We have discovered that nerve fibers transmit electricity. We have not, in the same way, discovered that they transmit information. We have simply postulated this conclusion by falsely modeling neurons on persons.

> ...there is theoretical danger in such loose talk, because it fosters the illusion that we understand how the brain can give rise to the mind.

On Kurzweil's book being just pop-sience book: McGinn worries, I think, about the public having a wrong idea of what we really have in terms of our explanations of mentality because theoreticians can have a wrong idea of it. Others might think it enough to propose certain hypothesis as solutions of some sort to those problems. But with that, I've polemicized too much for now, I think.

I thought that philosophers and computer scientists actually talked to one another.

Having said that it seems like the book reviewer's central complaint was To create a mind one needs at a minimum to create consciousness.... I'm quite skeptical of this claim simply because "consciousness" is such a sketchy concept: http://www.skeptic.com/eskeptic/08-02-27/

It seems that whenever computers become able to do something that we previously used to define consciousness, we just create a new definition of what it means to be human/conscious that doesn't include that new capability.

Essentially we're continuously redefining consciousness as "that which humans can that machines can't."

My guess is that we'll keep expanding that until human consciousness is just "the version that operates in flesh and blood rather than silicon".

The book review in question is crap. I'm happy to read Kurzweil critics but claiming something is too complex to understand to does not make for legitimate criticism, in any field.

Pattern recognition is a given, how to do it is not. That is really what the book is about (hint: hierarchical hidden markov models.) As for the philosophical stuff, I read The Age of Spiritual Machines many years ago when it came out (and I was barely in high school), so there was nothing in How to Create A Mind that challenged views I've had over a decade to think about.

The whole issue of consciousness I think is minor. Which, I personally would explain consciousness as "living" awareness ( in the book Kurzweil says even the smartest people start to sound stupid when trying to describe this.)

The good news is our industry is not going to get dull anytime soon.

Perhaps I should read Kurzweil once even if just to see how much our views are aligned..

At no point did someone think a computer would wake up, sing Hallelujah, and start composing public policy when it finished playing checkers.

Consciousness is clearly: the ability to have directed agency in the world (i.e. "wanting" to do something), following an action path to create a desirable outcome, and forming opinions/judgements/actions based on prior experience ("memories"). Fun thought experiment: when does a baby become conscious and not just a bundle of predefined cry-poop-eat behaviors?

Every animal walking around in the world is consciousness (if you can "turn off agency" in the form of sedation during surgery, the thing was consciousness), but human-level intelligence consciousness only forms in our squishy brains at this point in our world line.

From a philosophical perspective it's saying that consciousness is a property of reality rather than of humans / animals and that we're merely choosing to define an essentially arbitrary subset of it that's sufficiently similar to our own as 'true consciousness'.

Saywhatnow?

consciousness is a property of reality

Again, saywhatnow? Perhaps your usage of 'reality' is throwing me off.

Consciousness is just a very high dimensional aggregation of "stuff" (e.g. synaptic connections) with a processing framework (e.g. neural architecture) for recall/storage/action based on outward perception and internal state.

We're both agreed that consciousness doesn't require squishy brains. I'm just saying human level consciousness, at this point in time, only exists in our squishy brains. We'll figure out different ways of creating computational consciousness in the future, but we aren't there yet.

Our tiny recommender systems, search engines, face detectors/recognizers, and "my first backprop neural net 101" aren't conscious.

Given that assumption, all I'm saying is that as computers progress along that scale we are currently continually moving the point on that scale at which something becomes "human level" to specifically exclude computers.

A more accurate question might be, "can you recreate the running code of the Windows OS, given an installed desktop running Windows OS". Oh, and we also have some copies of the source code--if that helps (but it is hand-written in Mandarin). :)

My current "create a consciousness" track is to get something with pattern completion completion working (NB: aspie nerds often fail this step interpersonally ("Oh, Hello!" "blank stare"), but not intellectually/professionally ("What does this thingamajigger do?" "Oh, well, that's there because historically we used a 3.2 bit kernel to operate the Arm of God and... [30 minutes later]... Why are you falling asleep?").

The thing inside your head stringing your actions together into a coherent after-the-fact story isn't pattern recognition or even pattern completion, but pattern completion completion. You have to take the context of the new world generated by your prior pattern completion and run a new internal world simulation based on that, then take action. This does leave out intrinsic or emotional motivations pulling you hither or thither, but that's a vague writeup for another time. Though, there are plenty of smart people running around in the world with seemingly no emotions. Emotional feeling doesn't seem to be a prerequisite for consciousness (or for being usable to be exploited by employers to make them 10x-1000x richer while you get 0.5x richer).

It's fun to think about at least.

I believe Kurzweil is likely correct. Pattern recognition is the essence of human intelligence.

Consider this phrase:

> The belief that London is rainy intrinsically and literally contains the propositional content that London is rainy, but no state of neurons contains that content in that way—as opposed to metaphorically or derivatively (this kind of point has been forcibly urged by John Searle for a long time).

Or this one:

> So what is the basis for the theory? The answer must surely be that the brain is causally connected to the mind and the mind contains and processes information.

This is not the language of a book review, but that of a philosophical argument - although absent any of the justification or rigor one might thereby expect.

I'm not impressed with McGill - he shows little regard for his audience's time by pontificating on the subject instead of just reviewing the damn text. Lame.

The problem isn't only about human intelligence, it is about human mentality. McGinn's point is that pattern recognition is not essential to it; it very well might be true that pattern recognition is essential to intelligence, but that means nothing to the problem of mentality unless you reduce mentality into intelligence.

Well stated.

He is just being humble.

http://en.wikipedia.org/wiki/Colin_McGinn

And besides, it's not like this is news to anyone. Psychology & Neuroscience literature have known pattern recognition is an integral part of how the brain works for decades now. If it were the secret ingredient to the human brain, someone would have written about it far before Ray Kurzweil did.

And it seems like everyone (including Kurzweil) forgot, pattern recognition is not exactly an easy thing to harness (the brain is far superior while consuming less power), so even if pattern recognition is THE SECRET to human intelligence, it doesn't really get us much closer to understanding the brain.

Do you mean that you are questioning the ability to form new pattern detectors? Do you think that isn't done, or do you just think the mechanism behind how that happens is inadequately explained?

You could see Miami and you could recognize it as similar to London. That is pattern recognition.

The fact that humans are so bad at picking random numbers and so often pick 7 does not seem so unusual if we are based on patterns. Actually, that is the sort of emergent phenomenon that seems rather obvious to me.

> rrrrrrrtrrrrrryrrrrrurrrrirrrorrpr

you can recognize patterns in it even if you have never seen, heard, or known them before.

EDIT: of course this example is maybe too simple.

the point of others and the author is that pattern recognition doesn't explain the mind because it doesn't account for memory. my point is that the term recognition obviously implies a mechanism for memory.

Is it really the case that my brain is not recognizing patterns when I remember something? I don't think this is the case, otherwise I would never realize something about a remembered object after the fact.

> He does in one place speak of dreaming as a “sequence of patterns” and he might try to say the same about thinking. But this faces obvious objections. First, even if that is true, there is no pattern recognition involved when I dream, or when I think about London and my friends and relatives there.

Seems to me the brain is taking random stimuli and piecing it together into a narrative when one dreams. Sounds like there could be some pattern recognition involved there. Also, the author has never spoken of absent friends and said something like, "Doesn't Bob look sad?" Is he remembering the recognition of sadness, or is he recognizing sadness from the memory, and how could one be sure of the difference?

Methinks the author either knows something left unexplained, or isn't being nearly as clever as he thinks.

With physical recognition (touch, temperature, smell, taste, sound, sight), all the details are instantly available to our perception and recall/memory access is easier. It's difficult to fully remember the feeling of a childhood holiday only from your memory. You get a much richer recall experience if you go back to the place/time of year you lived it. Little things will trigger your memory you didn't even know you remembered.

The review author would be less confused if "pattern recognition" was replaced with "memories linked to what you're thinking about." Saying "pattern recognition" of London doesn't make sense to non-tech people. Saying "You remember things attached to London in your mind" helps out more, right?

I like that we are kinda strange.

Also bothersome is Kurzweil's repeated but unacknowledged use of Chomsky's ideas about recursion in language.

The brain perceives by receiving electrical signals from neurons, whether from the optic nerve (which, incidentally, encodes both colour and geometry) or the nerves from the nasal cells. These are going to arrive in the brain as identical chemical 1s and 0s for the brain's sensory pattern recognition circuits to deal with (or whatever.)

Independent thought, recursive concept matching (smell -> memory of childhood, memory of childhood -> fear, adrenal response) and all the rest are a strong point of the criticism, though, since these are obviously not based on pattern recognition.

Which to my ear, belies his fundamental apprehension of geometry.