Morty--I'm sure I don't know Dawkins as well as you do--is The Selfish Gene the best place to start getting better acquainted?

Yes, that book or, depending on your state of scientific knowledge, The Blind Watchmaker. Both are excellent reads. No one is better at summing up evidence and explaining biology and its implications. He's a vivid stylist. His mode of argument is usually framed by using extended comparison. The Selfish Gene, in an appendix essay, was where he first mooted his speculations on the meme.

Thanks bro. It's on my order list. Btw, which is which, insofar as the state of my scientific knowledge is concerned?

Also, the sequester can't be unpopular since it hasn't had any effect yet

Pretty sure that's not true. Things yet to happen can be popular or unpopular. Logically, you might argue that the popularity or otherwise of those things might be misleading, since the 'true' effects are yet to be known, but that's hardly the same thing.

Anybody else a little concerned that Dennis Rodman seems to be acting as a self-appointed US ambassador to N. Korea? Isn't that a little like Mr. T as a hostage crisis negotiator?

Once people are hit with a higher bill because of the Medicare and Medicaid benefit cuts, holy hell is going to be raised.

552:

The Selfish Gene is more technical; The Blind Watchmaker more accessible if like me you hadn't swotted up on science in a few decades. TBW has probably served as a basic starting text for some philosophers of science (like Dennett), so it's more thrilling for people like us who like to opinionate at will. You sound like you're more advanced than I was at the beginning stage when I decided to do some catching up, so I'd start with the first one, The Selfish Gene. It's also a true literary feat, a brilliant exercise in argument by extended conceit (no, he doesn't just rely on that--he uses it to make his points clear to the reasonably educated and informed reader.

EDIT: And if you are interested in evolution and Darwin's theories and views on evolution, DAwkins's The Greatest Show on Earth has been called the best popular presentation, although Jerry Coyne's Why Evolution is True (he keeps it as simple as his title) is very good and very quick (it's about 130 pages long).

And if you are interested in evolution and Darwin's theories and views on evolution, DAwkins's The Greatest Show on Earth has been called the best popular presentation, although Jerry Coyne's Why Evolution is True (he keeps it as simple as his title) is very good and very quick (it's about 130 pages long).

Haven't read the latter, but Greatest Show on Earth is what I would recommend for people who are borderline under the sway of the religious dogma, but still have a moderately open mind(I know mostly an oxymoron there) It's a great introduction to evolution book.

That's true. His first two that were mentioned are really applying his theories to Darwinian implications. They are massively wonderful intellectual exercises, though it is true they demand of you more commitment than a more conventional study on just what evolution is and all about. Had The Greatest Show existed when I started reading Dawkins almost 20 years ago now, I'd probably have started there. Indeed, I think I read his slim volume, River Out of Eden, that has just come out as a bullpen warmup, just to see if I was up to the game.

And Dawkins can be read for his science, isolating his religious opinions to one side (except for something like The God Delusion of course).

The apt comparison of the last 10,000 years to the era begun with Crick and Watson et al and including petaflop computers along with sequencing the genome is of the bicycle to manned, supraorbital spacecraft

Or would be, if manned supraorbital spacecraft were still with us and not an abandoned venture from 40 years ago :)

If I learned anything from watching Jurassic Park about 100 times when my son was eight years old, it's that humans can indeed exert artificial selection on organisms (indeed Darwin and Wallace were led to conceptualize natural selection by observing artificial selection), and yes, even to design genomes – but as Publicola notes, these designer organisms are still going to have to live in a system with almost infinite numbers and permutations of variables way beyond human control. It's not flat-earth thinking to point that out; it seems to me just an acknowledgment of reality. Feral animals, weeds, and invasive species of all kinds seem to be the way of human interventions in evolution, once vigilance is dropped for a moment (and it inevitably is).

If I learned anything from watching Jurassic Park about 100 times when my son was eight years old

On a beach vacation when my kids were about that age -- probably about six and eight -- I read much of the book aloud to them. They were completely into it. Good times.

Anybody else a little concerned that Dennis Rodman seems to be acting as a self-appointed US ambassador to N. Korea? Isn't that a little like Mr. T as a hostage crisis negotiator?

When Dennis Rodman replacing your current leader would be met with massive worldwide relief and support, you know there's a problem with your country.

559:

That's a good point. I guess I could research it, to go into it more intelligently, but for right now, let me just comment that I think Darwinists like Dawkins would say that our culture not only represents the culmination of what began as biology but becomes part of that changing environment the organism is selected for in the long run. What makes prognostications along these lines chancey is that we don't have a very long run here, so it's hard to discern. But, I don't see why Singularity (or whatever) wouldn't be incorporated in natural selection. Surely some of those downloads will be more adatable than others? Maybe, as Dawkins rankly speculated in that appendix in The Selfish Gene, that's when memes, or something like that, becomes Darwinian. Dawkins and others have speculated that Darwin's theory (and other attributes of evolution) are not dependent on the particular life that we have here and now. If there's a different based life somewhere in the universe, chances are, they say, it evolved, or is evolving along Darwinian lines.

I *suspect* and it's just a gut feeling, nothing more -- that the electronics/computer exponential expansion is cresting or near to cresting. It is rapidly outstripping power efficiency, and until we have exponential growth in solar energy capture efficiency, this will get stymied. You can make small devices more power efficient, but I think we are getting near the entropy cost to store a bit. And "Ye canna fight the second lawr of thermodyamics, Captain".

Biology, Biochemistry & Genetics on the other hand is just beginning it's exponential explosion (e.g, cost of DNA sequencing is currently Supra-moore's law).

It's called exponential growth for a reason,

I'm skeptical that computing speed's exponential growth can continue indefinitely. As with most examples of exponential growth, there are physical limits -- Planck's constant, atomic scales, the speed of light, etc. -- that will take effect at some point. Quantum computing looks to be only a piece of the answer.

And skeptical you/we should be. It seems extraordinarily unlikely it can continue indefinitely, for the reasons you mention, and more.** What I look at, though, are what appear to be the likeliest constraints, when those constraints are likely to kick in, and at what point those constraints interfere with significant progress, or with the particular topic under discussion.

Supercomputers currently run easily in the petaflop (10^15/second) range. The Cray Titan peaks at around 25 petaflops, or 25 quadrillion floating point operations per second. The Chinese have committed to a 100 petaflop computer by 2015. The tech for the next jump, to exascale (10^18/second) computing, is already well-understood. The chips necessary have already been built and tested. The process by which those chips will be assembled to work at that speed has no impediments, and the Chinese plan to reach the one exascale (one quintillion FLOPS) benchmark by 2018. The Indian government recently announced it will beat that by one year. That's another plus, that multiple nations and companies are working on supercomputers. The short term and therefore most reliable projections also mean that speed is projected to increase by a factor of 40 in five years.

No one is talking about a limit on the current architectures as less than 64 exaflops, so even the most conservative views don't picture things slowing down prior to getting within a few steps of the zettaflop computer (10^21/second, or 1,000,000,000,000,000,000,000 floating point operations per second), which can more than handle most projected simulations of the human brain in real-time.

The occasional pessmistic view on how current supercomputer architectures might run dry around 64 exaflops doesn't take into account even small conceptual leaps, such as shrinking the length and width of current processors while using photolitho techniques to more than compensate by increasing the number of layers (thereby shortening the distance information needs to travel) on a chip to 32. It doesn't account for integrating memristors into supercomputers, or memristors as components of neuromorphic computer architecture, which could substantially reduce speeds required for strong AI or brain emulation.

Worst case, development can certainly slow below the trajectory described by Moore's law without significantly impeding the development of AI or brain emulators. If the doubling slowed significantly, occurring every four years, it's not really a big deal except for those of us likely to be among the last generation of humans suffering terminally from the currently irreversible disease of aging.

In short, I think the chances are fairly good that before Moore's law ceases to apply, we'll be matching or beating it for awhile. There's also a real wild card, in that some forms of (not necessarily strong) AI under development may well aid us in beating Moore's law.

------------------

Btw, fans of the xkcd comic strip who aren't familiar with its forums might get a kick out of the free floating discussions there. The second post at http://forums.xkcd.com/viewtopic.php?t=22535&p=671338 gives some solid links to some of what are currently thought to be limiting values.

As for the ultimate "limits to computation", I'd like to be around for this:

It may be possible to use a black hole as a data storage and/or computing device, if a practical mechanism for extraction of contained information can be found. Such extraction may in principle be possible (Stephen Hawking's proposed resolution to the black hole information paradox). This would achieve storage density exactly equal to the Bekenstein Bound. Professor Seth Lloyd calculated the computational abilities of an "ultimate laptop" formed by compressing a kilogram of matter into a black hole of radius 1.485 × 10?27 meters, concluding that it would only last about 10?19 seconds before evaporating due to Hawking radiation, but that during this brief time it could compute at a rate of about 5 × 1050 operations per second, ultimately performing about 10 to the 32nd operations on 10 to the 16th bits. Lloyd notes that "Interestingly, although this hypothetical computation is performed at ultra-high densities and speeds, the total number of bits available to be processed is not far from the number available to current computers operating in more familiar surroundings.

As with most examples of exponential growth, there are physical limits -- Planck's constant, atomic scales, the speed of light, etc. -- that will take effect at some point. Quantum computing looks to be only a piece of the answer.

Again, it's all relevant to what you want to accomplish. IF zettaflop computing is adequate to creating strong AI and is powerful enough to allow brain emulation, there don't seem to be any obvious constraints in the way of supercomputing continuing along the lines of Moore's law (well, House's corollary to Moore's law is more what we're talking about) and reaching zettaflop speeds.

A couple of years ago Eric Williams posted on foresight.org a nice summary of how we'll push past 64 exaflops (and on to zettaflops):

Well, first, i think some more operational definitions are in order. Let’s assume “human-level AI” means “capable of non-emotional reasoning and problem solving at the level of the average human in realtime”. Basically, what we currently understand as the functionality of the neocortex (Strong AI). I think the “in realtime” is important, because if we can run an AGI at 1/1000th the speed of a human brain (still quite a feat), we’re quite a few years from being able to interact with it.

Quite a few people picked 2030 [as the year of human level, Strong AI being developed], but i didn’t see any real reasoning behind the number other than increasing computing speeds. L Zoel touched on simulation, this seems like a good baseline for a pessimistic projection. Let’s use the Blue Brain project as our benchmark, since it’s the farthest along in true neuronal simulation (with interconnects and not simple point neurons).

Blue Brain can simulate a rat-level neocortical column (~10,000 neurons) in realtime on IBM Blue Gene/L supercomputer (36 TFLOPS). These are advanced neuronal simulations at the cellular level, including interconnects between neurons. A human neocortical column has ~50,000 neurons (varies of course). Assuming the complexity squares with increasing NCC’s (due to interconnects), 25x more computational power is required to simulate 1 neocortical column, roughly 1 petaflop, in the range of the fastest supercomputers today.

The human neocortex has between 2-5 million neocortical columns. This means that a zettaflop computer (1 million times more powerful than today’s fastest supercomputers) is required to run the blue brain simulation, in its current state, on the scale of a human neocortex.

Now, this is incredibly inefficient. We aren’t actually writing intelligence algorithms, just simulating the brain down the cellular level. A fellow from Sandia labs predicts that with a zettaflop computer, we could model the entire world’s weather patterns at a resolution of under 100m for 2 weeks. Clearly this is far beyond the scope of what 1 human brain is capable of, yet the hardware required to do both is identical. I think it speaks to the inefficiency of the simulation, and the potential for simplification of an AI model.

But even with this pessimistic outcome of AI, if the colloquial version of Moore’s Law holds, by 2030 we have the processing power to do this. Any other advances in actual AI algorithms (Jeff Hawkins’ Nupic software excels at the pattern recognition many here have mentioned, i think his HTM theory holds much promise) could speed things along. I think 2030-2050 is a sure thing if computers keep pace, and it looks like they will, to me.

Shrinking MOSFETs down to 16nm by 2016, 3d chip stacking, optical chip interconnects, self-assembling CNTFETS, graphene clock multipliers, these are all things being experimented with and tested now that don’t require any wildcard technologies (like quantum computing, single photon transistors, molecular computing, etc).

2030-2050 has my vote… [for the development of Strong AI at the level of the human brain].

Once you have Strong AI functioning at the level of a single human brain, you're almost certainly within a decade of Strong AI functioning at a level one thousand times the smartest human brain (by gaining speed, if nothing else. Add improvements in software, and the contributions the AI is making towards its own development, and we're really taking off). It does cross my mind occasionally that this kind of radical, extrahuman development explains Fermi's Paradox. On the other hand, though, nonbiological intelligence would probably construct Dyson spheres and other energy sources necessary to megascaled supercomputing; we're already capable of detecting some Dyson spheres at some distances, so their absence suggests however slightly that the evolution of consciousness doesn't take a path of nonbiologic supercession with commensurate, unintelligible aims.

-----------------------

**I'm tabling for the moment things like Dyson spheres running megascale supercomputers, and the computation possible with evaporating black holes.

Biology, Biochemistry & Genetics on the other hand is just beginning it's exponential explosion (e.g, cost of DNA sequencing is currently Supra-moore's law).

Remember that device that Dr. McCoy would wave over his patients to make a diagnosis of what was ailing them? That's no longer in the realm of inconceivable science fiction, except the device isn't waved. It's held for 20 minutes or so after a slide containing a drop or two of the patients blood is inserted into it.

@556: Thanks Morty. I'm looking forward to reading.

@565: check out the "MIT Computer Program" thread. It'll be child's play to adapt the program as software for a handheld device that through a tiny video camera reads heart rate without touching the patient.

Btw, I've assumed the device McCoy is waving is a portable MRI (among other things), with real-time, high definition reading and diagnosis, natch. That means he's packing zettaflop power at least into a salt shaker.

Maybe yottaflop power (10^24) is more like it, since HD MRIs in real time are going to have to coexist in the device with a century or three of medical literature and expert diagnostic systems also accessible in real-time.

Btw, I've assumed the device McCoy is waving is a portable MRI (among other things), with real-time, high definition reading and diagnosis, natch. That means he's packing zettaflop power at least into a salt shaker.

It has a miniature di-lithium crystal-powered battery in it :)

So what devices from the original Star Trek actually came to pass?

-Hand-held cell phones for sure.
-MRI and PET and other non-invasive imaging, though the machines are much larger.
-widescreen video
-extra-solar system space probes

Still waiting for the opportunity to travel back and forth in time and to shtup extra-terrestrial hotties though.

Clearly that last is the real point of warp drive.

I assume you could regrow organs quickly in TOS (massive repairs were done from time to time), though I don't remember specific examples, and apparently there's now a push to be able to 3D-print organs from ones DNA. Aren't we close to some form of spray-on skin in order to make repairs? I remember something like that in the series. Though in 2013 what we have is presumably generic skin or more probably a skin-like substance, and not one that tailors itself to mimic your own skin's DNA the way it surely would in 2450ish.

I assume you could regrow organs quickly in TOS

In TNG they still couldn't fix poor Geordi's eyes though...

Though in 2013 what we have is presumably generic skin or more probably a skin-like substance, and not one that tailors itself to mimic your own skin's DNA the way it surely would in 2450ish.

There's a skin patch in development that's made of the epidermal extracellular matrix that directs repopulation and repair through intrinsic cellular signals. Not quite the same thing.

-MRI and PET and other non-invasive imaging, though the machines are much larger.

That is on the list of X prizes.

If you have seen minority report, the Kinnect is making it possible for operating systems like the one used in that movie. Which dwarves the operating systems on The Next Generation.

I assume you could regrow organs quickly in TOS (massive repairs were done from time to time), though I don't remember specific examples,

In Star Trek IV they were on present day earth and McCoy gave someone a pill that regrew their kidney(?).

In Star Trek IV they were on present day earth and McCoy gave someone a pill that regrew their kidney(?).

I thought it was cancer, which is why he's all aghast about somebody doing Chemo.

Jack, The NYRB essay on Kurzweil is posted. To be clear it's about his efforts in philosophy of mind, not about some of his other claims, so one could consistently agree with McGinn and also be impressed with Kurzweil. I don't know enough about Kurzweil to have strong opinions, but it does strike me as definitionally true of genius that it manifests at times in extreme stupidity. The confidence that propels people to new discoveries also leads them to make catastrophically simple errors in other fields. That's been the experience I've had with the 3-4 people I've known whom I thought of as geniuses by contrast with the many people if hundreds of exceptionally bright but more cautious people I know. So it wouldn't shock me if Kurzweil is both embarrassing himself here and breaking startling new ground elsewhere. NYRB review

He was agast at both Chemo and Trepanning, but the woman who took a pill was merely on dialysis - she was fairly old, and here kidney failure could've been for any number of reasons.

@574: Thanks for the link Greg.

Yeah--I like Kurzweil precisely because he's willing to extend himself into areas he's not entirely expert in. We have to do that if we're going to aim at comprehensive theories of Being, but it also means we'll occasionally look like fools. The alternative is to theorize only narrowly, from subjects we've mastered.

I'm willing to cut real thinkers slack. In his Historium Animalum and Generatione Animalum Aristotle was variously of the opinion that "the female is, as it were, a mutilated male" and that "the female is more dispirited and more despondent than the male, more shameless and more lying, readier to deceive and possessing a better memory for grudges" and most famously that men had more teeth than women, but those don't mean I'm going to chuck the entirety of the Metaphysics. Where I'd diagnose kookery is in cases where someone is confronted with superior facts and persists in erroneous belief.

@571: are we going to carry around like the current medic alert bracelets small, locket-sized containers of our own stem cells that can be applied to any wound or used to repair any organ? That might be an intermediate step before medkits carry universal stem cells.

Btw, transporters seem kinda old hat. Instead you'd probably digitize your Self and send it at the speed of light (experiencing no time passing even if you send yourself 10,000 light years away) anywhere there was a receiver. You'd then have the option of downloading yourself into a handy android body*** more sensitive and durable than a human body (which, in 200 years, surely won't be much like the bodies we're in now), and take it from there. Whatever SETI looks for, I imagine one of the first things a civilization only slightly more advanced than ours will send out will be the consciousness of some of its members, and the instructions for the kit required to house those consciousnesses. Why send the ambassador's message when you can send the ambassador?

------------------------

Upthread there was reference to figuring how neurons work as essential to building brains. Developments in the field over the last few years have been extraordinary. In the wiki (tech guys have done a great job keeping wikipedia up to date on these things), under "Neuromophic", itself a good read:

In November 2011, a group of MIT researchers created the first computer chip that mimics the analog, ion-based communication in a synapse between two neurons using 400 transistors and standard CMOS manufacturing techniques.

In June 2012, Spintronic Researchers at Purdue presented a paper on design for a neuromorphic chip using lateral spin valves and memristors. They argue that the architecture they have designed works in a similar way to neurons and can therefore be used to test various ways of reproducing the brain’s processing ability. In addition, they are significantly more energy efficient than conventional chips.

Then, check out "singing neurons" and their individual voices at
www.frontiersin.org:

Within the brain, neurons all have electrical voices, each singing out in harmony with millions of others, a complex choir of information processing from which emerges the crown jewel of the human being – the conscious mind.

As described in The Blue Brain Way: Creating ‘singing’ neurons, it’s necessary to first create the voices of neurons – i.e. create electrical models that are representative of the full diversity of electrical behaviors exhibited by neurons. These are the e-types. Secondly, it is necessary to transplant these voices into the correct ‘voice boxes’ or neuron morphologies to create me-types (morpho-electrical types).

That's at: http://www.frontiersin.org/blog/The_Blue_Brain_Way_Creating_‘singing’_neurons_part_two_/191#sthash.PYXoau2J.dpuf



***If there's a civilization (including your own) capable of building receivers, surely a mobile android capable of serving as a repository for consciousness isn't too tough a task for it.

We have to do that if we're going to aim at comprehensive theories of Being, but it also means we'll occasionally look like fools. The alternative is to theorize only narrowly, from subjects we've mastered.

Why are we limited to only theorizing? We have at our disposal the scientific method, right? If Kurzweil wants to be taken seriously, he can construct a model and experiment to validate his theories. Then his critics can review and duplicate, if possible, his findings. Short of that, its all just mental masturbation.

Also, the sequester can't be unpopular since it hasn't had any effect yet; what's unpopular is the scare tactics being announced by Obama and reported uncritically by the liberal media, in which cutting an infinitesimal portion of the federal budget, will somehow cut every popular program but nothing unpopular.

If you break out the budget into general categories, there isn't a single area where cuts enjoy majority support, and there's only one area ("aid to world's needy" - a tiny item within the budget) where the cuts have even a plurality on their side.

March thread?

March thread.

Why are we limited to only theorizing? We have at our disposal the scientific method, right? If Kurzweil wants to be taken seriously, he can construct a model and experiment to validate his theories. Then his critics can review and duplicate, if possible, his findings. Short of that, its all just mental masturbation.

What on earth are you talking about?.

First, theorizing about the future inherently means theorizing about untestable hypotheses. You're therefore asserting no one should theorize about anything that can't be currently tested, which is obviously absurd. Second, even if you only theorized about what was currently on the shelf, you'd move at the pace of ants if you yourself tested every single hypothesis you were offering. That's what people developing slightly better mosquito sprays do. It's never the way any thinker in any field theorizes. The people developing cpu's as short as five years out have to theorize an enormous number of currently untestable hypotheses. If you go ten and twenty and thirty years out, which there's obviously a great need to do, the number of hypotheses you make necessarily increases, and the number of untestable hypotheses correspondingly increases.

Third, stating what the fellow with numerous patents, inventions, and successful companies to his credit, and who has just been named Director of Engineering at Google should do if he "wants to be taken seriously" tells us either that you've stooped trolling or have literally no grasp of anything that's been done in AI and computing in the last twenty years. Please: just stop.

The opening of the NYRB's review of Kurzweil's latest book is... uninspiring. Thinking that it's meaningful that someone is not a "professional" philosopher tells us about Colin McGinn, not his subject.

It's a strange review. The language is bizarre. "One this" and "one that" abounds. The anger and resentment is evident from the early going. That Kurzweil believes he's found the fundamental basis of intelligence is suspect because it's in a field where Kurzweil can 'manufacture something for a price'. But, don't "professional" philosophers offer theories "for a price", namely, their salaries? In their case, though, that's presumably a good thing. Here it garners the reviewer's suspicion.

The writer's main objection makes little sense. Kurzweil has not proposed that pattern recognition is the only function of mind, but the writer insists he has, and invents the following peculiar and irrelevant complaint.

In what way does thinking involve processing a stimulus and categorizing it? When I am thinking about London while in Miami I am not recognizing any presented stimulus as London—since I am not perceiving London with my senses. There is no perceptual recognition going on at all in thinking about an absent object. So pattern recognition cannot be the essential nature of thought. This point seems totally obvious and quite devastating, yet Kurzweil has nothing to say about it, not even acknowledging the problem.

A later objection, an extension of the above complaint, is simply a failure of reasoning:

"...there is no pattern recognition involved when I dream."

Sure there is. It's in part how you recognize faces and buildings within your dream. Pattern recognition is part of how the images (and emotions associated with them) arose and were retained in memory in the first place. It's likely that how you possess or carry or see the image involves one function of the brain, while another aspect of the brain recognizes that image. Alzeheimer's is quite obviously a disease wherein you routinely retain and carry the first part, the image, but fail to mantain the ability to recognize that image. Part of that failure involves a distintegration of the capacity for pattern recognition.

Kurzweil has to be correct in theorizing that pattern recognition is an essential component of intelligence. Additional issues are a) is pattern recognition not just essential but fundamental? b) is pursuing it as fundamental if it is not a costly diversion from a better path?

As for b), it needs to be explored, and even if it turns out to be a bit of a false lead, figuring out how it is and why it is could easily be a key step towards Strong AI.

In order to fabricate a criticism, though, McGinn takes the obvious hyperbole of the title and not only pushes it to an absurdity that has no relevance, but also pretends that perversion of the title's meaning should displace everything else Kurzweil writes. It's a pretty silly piece, and obviously so. And especially so when you get to around the fifth graf from the end of the first page where the best the writer can do is take metaphorical language literally. It's like criticizing Impressionist painting for "failing" to be realistic.

There's also an odd and unscucessful attempt at deconstructing the language of neuroscience and AI on the second page. What a mess. As the remarks on Wittgenstein demonstrate so clearly, Kurzweil's biggest sin is not talking about what the writer wants him to talk about. The last three paragraphs, rattling on about "law" as though Kurweil isn't using "law" colloquially, is a particular embarrassment.

Much more interesting is how Kurzweil might use his and Google's convergence of interests to develop intelligent algorithms based in part on what I call "deep" pattern recognition. (Structural linguists will know what I'm describing).

"Watson", IBM's Jeopardy-playing software, was successful at the narrow task its enormous resources were aimed at, but it evinced no intelligence. It did nudge information seeking programs ahead, and increased our ability to draw meaning out of certain, strictly defined phrases, but otherwise it wasn't meaningful towards creating Strong AI, any more than every chess-playing software ever developed was.

I do wonder, though, if in a decade you tie together the hundred most advanced expert systems in a hundred different fields, how much difference will there be between that combination of systems, and talking with a highly intelligent person? A great deal of effort will be made to make these expert systems user friendly. The expert system IBM developed out of the Watson program designed to help doctors make diagnoses is a billion-dollar market. The ease with which its output can be read will go a long way towards making it acceptable.

Still, none of that makes for Strong Artificial Intelligence, on the order of the way a human being with a minimal background in it might take an interest in astronomy and over several years (a few hours for a powerful computer) become expert in it.

Kurweil believes in the broadest sense that understanding and applying pattern recognition within language is the route towards creating machines that think; specifically, that we draw meaning out of the evidence we look at, and that meaning is found, in part, by both discerning patterns within evidence, and creating patterns in order to convey meaning. He's almost certainly partly right. He may not be quite as right as he thinks he is, but that's very different from being wrong.

To push much beyond where it is now, Google's search algorithms are going to have to develop the ability to draw meaning out of language by identifying patterns in advance, then applying those patterns to a given phrasing, and I don't doubt Kurzweil will be aiming at creating the kinds of algorithms that don't require software to be taught every possible, specific expression in order to draw out meaning and then respond to that meaning; then, to create meaning. .

In advance of high fidelity brain emulators allowing for downloading, I wonder how much information a human being would have to leave behind to have a shot at resurrection? Would, say, an hour of real time MRI, which might be possible by 2029, do it? Will you leave a copy of that along with instructions and enough cash to pay for the process's completion?.

One of the most interesting questions is, how will we even know when we develop Strong AI? What is the form of a Turing Test where we verify we have created a consciousness, as opposed to an entity merely asserting consciousness? That's a problem we may never completely solve.