Talk:Technological singularity/Archive 4

Hal 9000, Monoliths and Space Odyssey saga
In my opinion these should be mentioned here as they're probably among the most important early conceptions on the topic, especially the incomprehensible monoliths.
 * If you're talking about Clarke wouldn't Childhood's End be a much better example? BTW sign your posts Tarcieri 09:21, 22 August 2006 (UTC)

Fictional depictions
I was thinking that Philip K. Dick should be mentioned in the fiction section. The Blade Runner story, as an obvious example, is primarily about the conflict of AI Simulacra with ordinary humans (and also the question of sentience vs. sapience). But in addition to that he wrote dozens of stories involving AI and intelligence amplification long before it was a popular subject (or even a defined subject at all). Is it just that no one is familiar with these stories or are they deliberately not mentioned? -- abfackeln 19:12, 11 June 2006 (UTC)

Pseudoscience
The term "Technological Singluarity" keeps being added to (and subsequently deleted from) the article on Pseudoscience. There is a discussion going on right now on the Pseudoscience talk page on whether or not to keep the term in or not. In my opinion it should not be listed as Pseudoscience, (perhaps Protoscience, but definitely not Pseudoscience). All comments on the other talk page are welcome. (Cardsplayer4life 18:21, 12 April 2006 (UTC))


 * Proto. It is susceptible to disproof (ie. if over next hundred year progress declines in either acceleration or velocity). --maru  (talk)  contribs 22:02, 12 April 2006 (UTC)

Singularity Timetable
—Preceding unsigned comment added by 67.150.216.115 (talk • contribs)
 * Singularity Timetable [link removed] predicts a Technological Singularity to happen in 2012:
 * 2006 -- True AI
 * 2007 -- AI Landrush
 * 2009 -- Human-Level AI
 * 2011 -- Cybernetic Economy
 * 2012 -- Superintelligent AI
 * 2012 -- Joint Stewardship of Earth
 * 2012 -- Technological Singularity


 * I am wary of anything that has 'blog' in it's name. True AI this year? Who is the author of that site?--Piotr Konieczny aka Prokonsul Piotrus Talk 18:42, 7 May 2006 (UTC)


 * Almost certainly Mentifex. --maru  (talk)  contribs 19:18, 7 May 2006 (UTC)
 * ?--Piotr Konieczny aka Prokonsul Piotrus Talk 20:51, 7 May 2006 (UTC)


 * Who knows. Google him, and you'll find plenty of material. In case you missed it, the most important one to read would be Google hit #3, The Arthur T. Murray/Mentifex FAQ. --maru  (talk)  contribs 22:30, 7 May 2006 (UTC)


 * Maru, small world, isn't it? Back when I met you in the Supershadow article, I never expected you to also be interested in the Technological Singularity. What drew you to become interested in this topic in the first place?


 * I got interested in it because of the prospects of the Singularity eliminating poverty, cleaning up the planet, instantaneously replicating anything we wanted, inventing brand new things we never would've thought of in many years, and generally giving us a much greater quality of life. I'm hoping to see Sierra Leone reach an HDI of .905 and the US reach .9995 due to the Singularity. --Shultz IV 04:24, 8 May 2006 (UTC)


 * Ah, hey, Shultz. I haven't heard too much about you lately (which is a good thing!) But it's not so much a small world as it is the restricted scope of our respective interests, and my proflicness (last I checked, I have over 2000 articles on my watchlist, and in my ~year here have at some point or other editted >6000 distinct pages). The Birthday Paradox is also relevant as well.
 * As for why am I interested? Simply put, I'm pretty sure it has to be wrong, but I haven't been able to come up with any evidence or any chains of reasoning that can satisfy me (much less anyone else) as to why it is wrong. So I have to provisionally accept it as true, and so it behooves me to follow the area. Besides, the SL4 mailing list makes for pretty interesting reading. --maru   (talk)  contribs 06:07, 8 May 2006 (UTC)

This timetable is ludicrous. The Vingian singularity can't be accurately predicted as it relies on AI technology that not only is not yet invented but for which we really have no idea as to whether it could ever been invented. It could happen next year or it could happen in fifty years or it could never happen. Tdewey 00:34, 30 October 2006 (UTC)
 * Vingh also proposed Intelligence enhancement as a possible path to a singularity. This is certainly possible although the limits are unknown, maybe not enough to trigger a singularity. Or simulating / copying an existing intelligence (ie a person) without necessarily understanding it could also lead to AI-like entities without the need for the Artificial creation. Anyway, certainly not accuratly predictable. 203.46.224.202 01:22, 6 November 2006 (UTC)

Unabomber
Must the Unabomber's writings be used? -- Comment unsigned


 * I second the sentiment, and would like to remove the Unabomber or at least reduce his inclusion to a passing mention. His prevalence in the article now is likely due to edits I made several years ago (when I held neo-Luddism in higher regard) that overstate the connection between his writings and the Singularity. I've read the "Unabomber Manifesto" several times and while it certainly contains speculations regarding the future of technology and how it's bad for humans, it has no particular relevance to either accelerating change or the Vingean Singularity. Could other editors of this article please comment so that we might reach a stronger consensus on whether or not he can be removed? -- Schaefer (Talk) 18:11, 30 July 2006 (UTC)


 * Over two weeks have passed without any objections. I am removing the Kaczynski paragraph. -- Schaefer (Talk) 06:15, 17 August 2006 (UTC)

Article needs Criticisms
I'm not familiar enough with this topic to edit this page myself, but I would appreciate a dedicated section on the scientific criticisms of this idea. This is limited to one brief and vague para at the end of "K's Law of Accelerating Returns" - in particular, the last statement could use a source. The section on Neo-Luddite Views seems just to discuss opponents as anti-technology, wheras there must be tech thinkers that don't believe in Singularity per se. Plus these opponents are described as fighting or fearful of it, which implies it's inevitable or even definitively possible. It would enrich this page to include points of debate on its plausibility. Thanks! Neonaomix 21:22, 20 May 2006 (UTC)

I wrote this a long time ago, but it was deleted: 71.199.123.24 21:53, 27 July 2006 (UTC)

Weasel critics
I don't know what to do about the guy who keeps entering the single line "A common criticism is that the singularity is unrealistic and unlikely to occur." It has been deleted and replaced a number of times now. -- abfackeln 08:53, 13 October 2006 (UTC)

And why has it been deleted? Isn't that a valid criticism? -- unsigned
 * I didn't delete it but I am going to guess it was deleted to avoid weasel words because it did not state any sources or evidence. Take the next statement in that section, for example: It starts weasely enough with "Some criticize Kurzweil's choices of specific past events to support his theory." but then goes on to explain some of the who and why to support this statement, which is why it hasn't been deleted. -- abfackeln 04:00, 16 October 2006 (UTC)

The possibility of an intelligence creating something more intelligent then itself
Technological singularity assumes that it is possible for an intelligence (human or otherwise) to create something more intelligent then itself. First by humans creating such an AI, and later by this AI creating something even more intelligent.

However this is far from proven. For 40 or so years reaserchers have been predicting AI's and nothing has even come close. It may in fact be impossible for humans to create an AI that is even close to as inteligent as a human, and even more difficult for this AI to create an even better AI.

It is important to define inteligence here. For the purpose of this theory intelligence is best defined by creativity: creating something new that has never existed before, and is not merely a new combination of old ideas. Even for humans this is a rare occurance (true breakthroughts are exceedingly rare), and it may be impossible to create a machine that can do it.

All AI's currently envisioned are deterministic in programming. Meaning: given enough time a programmer can prededict exactly what the output of the AI is. If that is the case then it is hard to imagine how such a machine can create something beyond what the programmer can envision. This begs the question: perhaps it's possible to create a non-deterministic AI. However no ideas (beyond a random number generator, and unlimited permutations of random programming) have been proposed on how to do it.

Random-AI's are distinct from evolutionary AI's, (which have been created). Evolutionary AI's have never created anything new - they only stumble toward an already pre-defined goal. To actually come up with a new goal appears to be impossible unless they are run undirected, in which case they are no longer evolutionary (in the survival of the fittest sense) and have to try every single possible thing in order to come up with anything. The computational requirments for such an AI are many orders of magnitude beyond anything positied for the future, and are probably beyond what even a galactic quantum computer can do.

(Even trying every possible 1KB program would require 12048 permutations, assuming 100 possible useful commands per byte. Please see Large_number for more information about the impossibility of doing this.)

It may be that a technological sigularity never happens: computers get faster, but not smarter. -- Unsigned


 * Minor correction: It's 1001024 (assuming 100 useful commands per byte as you specified), not 12048. The latter simplifies to 1. -- Schaefer (Talk) 23:38, 27 July 2006 (UTC)


 * This whole argument is pointless. If you're trying to say that it will take a breakthrough to achieve such a thing as human-level AI, I don't think anyone doubts that it would indeed be quite a breakthrough -- but to claim its impossible because it hasn't happened yet makes no sense. Though I find it fascinating that you simultaneously overestimate human intelligence (in saying that it would be impossible to reproduce artificially) while also underestimating it to a similar degree (that it could never produce such a thing). -- abfackeln 08:50, 13 October 2006 (UTC)

Distinguish between increasing general intelligence of mankind and specific intelligence of individual humans/AI
I think the initial poster in the above section fails to distinguish between the three types of intelligence enhancements postulated by the various folks on this subjet.

1. General Intelligence. Technological, biological and social advancements that increase the society-wide or species-wide intelligence of mankind. Examples include (but are certainly not limited to) the invention of agriculture (allowing permanent human settlements as well as better nutrition), writing (allowing knowledge to be passed from one generation to the next in a secure fashion), medicine, mathematics, democracy (allowing free distribution of knowledge), the printing press, computers, the internet and wikipedia. Possible future advances that might trigger the singularity include the development of hard AI accessible by the average person (see for example Pohl's Sigfrid character in the Heechee novel Gateway).

2. Specific human intelligence. Technological, biological and social advancements that increase specific individual's intelligence. Examples might include medical and nutritional advances that permit improved in-utero development of an individual baby's brain such as ingesting appropriate amounts of folic acid. Possible future advances might include genetic modification of human DNA or methods of better integrating the human mind with non-AI or AI computers.

3. Specific machine intellgience. Technological advances that create either hard-AI computers or soft-AI machines capable of integrating with the human mind.

Given these definitions I think we can state that we have improved our intelligence. The general human intelligence now is greater than 500 years ago or even 50 years. Similiarly, the average specific intelligence is greater than 500 years ago or even 50 years ago. See cite to Brain Gain article below; See also http://pespmc1.vub.ac.be/FLYNNEFF.html showing that average IQ for developed nations has increased by between 10 - 30 points per generation. See wikiarticle Flynn effect. Of course, while we have improved soft-AI technology, we do seem to be no closer to creating hard-AI now than we were 40 years ago.

Given the above I'm not sure that I agee that the creation of self-replicating or self-improving Hard-AI is either a necessary or sufficient condition for the technological singularity though I do agree with Vinge/Goode that some-type of method of intelligence explosion is necessary. Tdewey 22:16, 29 October 2006 (UTC)

God and the Singularity
I think a small section discussing God and the Singularity should be added. Maybe just a quick summary? Tdewey 22:16, 29 October 2006 (UTC)

First figure contradicts Kurzweil's thesis
I'm surprised that nobody has questioned the use of the first figure in this article. In it, we see a straight line on a log-log plot; such lines are not indicative of exponential trends (as suggested in the caption), and therefore appear at first sight to contradict Kurzweil's thesis of accelerating returns. The math savvy may recognize that straight lines on log-log plots are characteristic of power law relationships. How does one reconcile this with the exponential relationship predicted by Kurzweil?

If nothing else, the caption to the first figure ought to be corrected. It would imply that the graph is exponential, which it is not! Besides the Wiki entry, I'd also be interested to hear anybody's take on how this graph does not contradict Kurzweil's thesis. (Incidentally, as of May 13, 2006, Kurzweil is still using this graph in his talks; he briefly displayed it at the Singularity Summit at Stanford.) --chodges

I was going to point that exact same thing -- any cretin having studied calculus should realize in seconds that a polynomial function plotted log-log will give a curve approaching a straight line very quickly. This means that the log-log curve of any monotonous function that is *NOT* exponential or worse will have a nice quasi-linear portion that you can exhibit by cropping the display range. Anyone who pretends that this graph shows exponential behaviour should be condemned to cleaning toilets for the rest of his life. Not only that, but the graph doesn't mean anything anyway, as the choice of data points is totally arbitrary. Carl Sagan, a technological event !?! Hello? See for instance. Now (as pointed in the Wikipedia article), the idea of a singularity, which is quite interesting, is nothing new and is not due to that Kurzweil guy. I don't see why we should continue to associate that concept with him. Does he sell Singularity T-shirts and mugs ? --Congruence 22:04, 23 May 2006 (UTC)
 * Aren't you ignoring the fact that the data on the vertical axis of the graph represents time BETWEEN events and not the time of the events themselves? Rbarreira 15:42, 27 May 2006 (UTC)


 * That may increment the polynomial degree, but it still doesn't make an exponential. --Congruence 19:54, 27 May 2006 (UTC)


 * Well, first of all I should remark that you didn't understand the graph at all. You said "Carl Sagan, a technological event !?! Hello?". If you had understood the graph, you would have noticed that the symbol for Carl Sagan (or any other symbol) appears multiple times, being the important events as collected by Carl Sagan. Carl Sagan is not an event, which shows that you were indeed quite attentive in your analysis of the graph. As for your other objection, it's just plain wrong. Go to this site and use some spreadsheet software to plot one of the datasets into a graph and see what the curve is like if you assume time between events. Rbarreira 22:31, 27 May 2006 (UTC)


 * I somewhat doubt that Kurzweil would make such a simple mistake. Although I learned math in another language and would have to check the translations to be sure I understand the presumed error, even if you are right I don't think that invalidates the graph. Exponential or power law, the result is similar (for our needs). Last, let me remind you of WP:V: if Kurzweil sais it's exponential, and we can verify this, this is good enough for the caption. If another proper reputable source points out he made an error, then we can adjust it.--Piotr Konieczny aka Prokonsul Piotrus Talk 00:57, 28 May 2006 (UTC)

Well, I stand corrected. I actually did some math to see if Rbarreira was right. [For those interested, assume that $$n = n_0 \exp(kt)$$, where n is something like number of inventions, and t is time. This is exponential technological advancement. In this model, the time until next invention at any point t, is just
 * $$\Delta \tau (t) = dt/dn=(n_0 k)^{-1}\exp(-kt)$$.

If you plot this on a log-log scale, then you do indeed get a straight line like the debated figure.] So Kurzweil's plot is consistent with an exponential trend -- BUT -- to be fair, I still maintain that the way Kurzweil plots this is very deceptive. The reason that he didn't plot it on a normal semilog plot (like all the others he normally shows), is that the scatter would look huge. By compressing both axes into log scale, it makes it look as if the correlation is better than it really is, and this is a rather big no-no for serious academics. I think many peer-reviewers would sneer at showing a plot in such a way. Just my two cents.

Also, I think Piotrus has warped the spirit of verifiability. Simply because a well-known person makes a claim doesn't mean it should be portrayed as fact in Wikipedia. An intellectually honest way of maintaining both verifiability and NPOV is to report something like "Kurzweil claims such a trend is exponential," rather than stating flatly that said trend is exponential and proving it by saying (sic) "Kurzweil sais it's exponential, and we can verify this." --chodges 19:47, 6 June 2006 (UTC)

OK to clarify a bit. I was saying that a linear-looking log-log plot doesn't IMPLY exponentiality. It doesn't IMPLY non-exponentiality neither. In fact, it tells very little. I'm not saying that the relationship depicted in the graph is not exponential. I was saying that the graph doesn't tell us if it goes one way or another. Now we'd like a proper graph to assess whether there is indeed an exponential relationship, and then we can go and criticize the data points. --Congruence 12:22, 11 June 2006 (UTC)


 * Isn't that backward? Shouldn't one get the datapoints, then decide whether they fit any any exponential curve, and then plot that curve to see whether it is at all interesting? --maru  (talk)  contribs 17:59, 11 June 2006 (UTC)

Movies
Don't forget The Forbidden Plant released in 1956. The Krell who merged with the computer. Morbius from Earth who used the computer. (see Wikipedia http://en.wikipedia.org/wiki/Forbidden_Planet)

Singularity thinking just an artifact of human memory?
AI researcher Juergen Schmidhuber recently published a paper where he mentions the singularity (he calls it Omega, referring to Teilhard de Chardin's Omega point, 1916). For Omega = 2040 he says the series Omega - 2^n human lifetimes (n<10; one lifetime = 80 years) matches the timeline of the most important events in human history. That by itself seems remarkable. But then he goes on to question such lists, suggesting they just reflect a general rule for "both the individual memory of single humans and the collective memory of entire societies and their history books: constant amounts of memory space get allocated to exponentially larger, adjacent time intervals further and further into the past (...) Maybe that's why there has never been a shortage of prophets predicting that the end is near -  the important events according to one's own view of the past always seem to accelerate exponentially." Here are the links: http://www.idsia.ch/~juergen/history.html and arxiv: http://arxiv.org/abs/cs.AI/0606081 Science History 14:15, 26 June 2006 (UTC)

Exponential and Hyperbola
There seems to be a mistake that is systematically reproduced throughout the whole "technological singularity" thread, whereby hyperbolic growth is denoted as exponetial. Note, e.g., that exponential growth does NOT lead to "singularity" in a strict mathematical sense of this word.

The exponetial growth is described by the following differential equation:

dX/dt = kX,

and does not imply any singular points.

The hyperbolic growth is described by the following differential equation:

dX/dt = kX^2

The solution of this differential equation looks as follows:

Xt = C/(t0 - t),

where Xt is the value of X in time t, C = 1/k, and t0 ("critical t")corresponds just to the singular point ("singularity") when the value of X becomes infinite. Note that the curve generated by this equation is nothing else but a hyperbola.

Note also that this is a hyperbola (not exponential) that looks as a straight line in a double logarithmic scale (an exponential looks as a straight line in a single logarithmic scale). Hence, I would advise to replace systematically "exponential" with "hyperbola/hyperbolic" (or "power-law") throughout the whole "technological singularity" thread. —Preceding unsigned comment added by Athkalani (talk • contribs)

Law of accelerating returns does not equal the singularity hypothesis
The technological singularity in its original formulation by Vernor Vinge had nothing to do with hyperbolic growth or the law of accelerating returns, however interesting they are in their own right. The singularity hypothesis is simple: At some point in the future, science and technology may lead to the creation of smarter-than-human intelligence that can recursively self-improve, a turning point in human history -- I.J. Good's "intelligence explosion." The singularity has nothing to do inherently with hyperbolic growth. The singularity is solely about the power of intelligence and what happens when that power is expanded beyond the range where our species' cognition has resided for the past ~40,000 years. Smarter-than-human intelligence could result during a period of modest technology acceleration, expansive technology acceleration, technological deceleration, or technological stagnation. The singularity hypothesis is not about the law of accelerating returns, Moore's Law, or hyperbolic growth, and does not require that any particular technology predictions hold true. Everyone needs to stop making that mistake.

Editors: Please remove this mistake from the Wikipedia article e.g. in the "Pop-Culture" section where it's claimed falsely that static analysis can refute the singularity hypotehsis; the latter has nothing to do with the former. -- The preceding message was not signed


 * I couldn't agree more. I'm glad to see I'm not the only one concerned about this. I'm slowly trying to reorganize the article to make clearer the distinction between the Singularity as seen by Vinge and the Singularity as seen by Kurzweil. -- Schaefer (Talk) 11:54, 19 July 2006 (UTC)


 * I don't agree with this formulation. This conflates the Singularity (that point after which future-time is unknowable and unpredictable) with the methodology that brings about the Singularity (the intelligence explosion).  I do agree that the methodology (strong AI, human-computer interfaces, whatever) that starts the chain reaction leading to the intelligence explosion could start at any time and is not conneted to technological acceleration (though hyperbolic technological growth leading to the singularity is the expected result from the intelligence explosion). Tdewey 06:15, 30 October 2006 (UTC)

Don't forget to mention that Vinge invented the notion!!
Guys, this is a super article - but don't forget to mention that Vinge invented the concept and coined the phrase!!

Kurzweil is withouth doubt the current *popularizer and investigator* of the singularity - indeed he, Kurzweil, constantly mentions that Vinge was the bloke who thought it up, and coined the phrase.

Vinge is like the Wright Brothers; Kurzweil is like the modern airlines! Vinge is like Deisel or Benz; Kurzweil is like Konigsegg.

It's silly to sort of mention or describe the singularity without mentioning Vinge in the same breath.

(Which indeed, is what Kurzweil does every time.. "the singularity, which Vinge came up with and I am popularizing.." sort of thing.)

Also ........ something to consider .. .where the article reads:

"The concept, put forth primarily by mathematician Vernor Vinge and inventor and futurist Ray Kurzweil, " .....

I'm not really sure that that is accurate. Only really KURZWEIL now "puts forward", explores, writes about, sells books about, the singularity. Vinge really doesn't have, in a way, that much interest any more. He seems to have moved on to other things.

Kurzweil is, without doubt, the PRE-EMINENT SINGULARITY POPULARIZER of our day. Vinge is not, in our day, a singularity popularizer .. he's just the guy who invented the notion and coined the phrase a few decades ago.

So FOR TWO REASONS it's somewhat confusing to say:

"The concept, put forth primarily by mathematician Vernor Vinge and inventor and futurist Ray Kurzweil, "

(Reason One) -- people might accidentally think that Kurzweil had something to do with CREATING the concept - which would be utterly ridiculous.

(Reason Two) -- conversely, people might think Vinge is 'putting forth' stuff currently about the Singularity. I can't really see that he is; Kurzweil utterly dominates this field and is putting out all the interesting new information and thinking about the field.

(The only real involvement of Vinge currently in .. singularity strudies ... seems to be that Kurzweil trots Vinge out now and again to say "and this is the guy, Vernorn Vinge, who came up with the concept of the singularity and named it!" in a radio debate to help sell one of Kurzweil's books or promote his website or whatever.)

Again -- to make an analogy -- if I said "The history of the automobile. Automobiles are primarily put forth by Dr. Deisel, Dr. Benz, and Ford Focus." ... that would be ridiculous ... it would make you think that "the ford focus" had some part in the invention of the automobile. In contrast, of course, the ford focus is a best selling car now, and is totally, utterly and completely unrelated to the ?invention of the automobile.

Style -- Sentences far too long
Just incidentally, whoever's primarily writing this article - your sentences are far too long! It makes it sound pseudointellectual, like someone trying to write in a "technical journal" style! Chop 'em up.

Second intro paragraph

 * The concept, put forth primarily by mathematician Vernor Vinge and inventor and futurist Ray Kurzweil, predicts a drastic increase in the rate of scientific and technological progress following the liberation of consciousness from the confines of human biology, allowing it not only to scale past the computational capacity of the human brain but also to interact directly with computer networks. Furthermore, progress inside of the posthuman/AI culture would quickly accelerate to the point that it would be incomprehensible to normal humans. [...]

These lines are overly specific. As far as I know, Vinge does not predict human intelligence augmentation and brain-computer interfaces will precede superhuman AI, or even occur at all. To say that the concept of the Singularity predicts humans will gain augmented intelligence, interface with computers directly, and form something that we would recognize as "posthuman culture" is roughly analogous to saying the concept of global warming predicts that Alaskan oysters will suffer from increased bacterial infections due to the rising water temperatures. If temperatures rise drastically and the ice caps melt but the oysters unexpectedly stay perfectly healthy, it's still global warming. Likewise, if some superhuman AI were to come along, solve scientific problems far beyond the ability of human scientists, and develop sufficient nanotechnology to reorganize every molecule of the Solar System as it sees fit, all without the aid of cybernetically augmented humans, it's still the Singularity.

Some futurists probably do predict humans will be liberated from the confines of their biology, will interact directly with computer networks, and will develop a posthuman/AI culture. I think Kurzweil does, but I haven't read The Singularity Is Near so I can't say for sure what his present forecasts are. Either way, these are only predictions of what the world might be like if the Singularity were to occur, not necessary conditions for its arrival. -- Schaefer (Talk) 13:51, 18 July 2006 (UTC)


 * Greg Bear's Eon novel set in his The Way (Greg Bear)  universe presents a society where post-humans live either within computers or in artificial bodies.  Tdewey 14:47, 30 October 2006 (UTC)

First paragraph
The first paragraph as it stands now:


 * In futures studies, a technological singularity represents an "event horizon" in the predictability of human technological development. Past this event horizon, following the creation of strong artificial intelligence or the amplification of human intelligence, existing models of the future cease to give reliable or accurate answers. Futurists predict that after the Singularity, posthumans and/or strong AI will replace humans as the dominating force in science and technology, rendering human-specific social models obsolete.

I have a few issues with this:
 * 1) The article opens with a metaphor. It should begin with a clear definition.
 * 2) Even if we excuse the metaphor, the first sentence is not a good definition. The Singularity is not merely something unpredictable. There's nothing novel about our inability to predict the future&mdash;we've been failing miserably at predicting the future since the time of Nostradamus.
 * 3) The opening is not NPOV, since it implies the Singularity will happen: "past this event horizon [...] existing models of the future cease to give reliable or accurate answers." There certainly are people who think they can predict the future just fine, including futures that contain smarter-than-human intelligences.

I'd like to submit the following new intro sentence, which I think says the same thing as the existing paragraph, but with more neutrality and in fewer words:


 * In futures studies, a technological singularity (often the Singularity) is a predicted future event or period usually characterized by rapid technological progress, unpredictability to pre-Singularity humans, and the presence of smarter-than-human minds, whether they be augmented "posthumans" or artificial intelligences.

-- Schaefer (Talk)

Incidentally, the intro paragraph has just been quoted in full on the front page of Slashdot. -- Schaefer (Talk) 04:57, 24 July 2006 (UTC)
 * Well, I may be a bit biased because I wrote most of the previous intro paragraph, but don't you think the present intro is a bit cluttered and unclear? Tarcieri 09:59, 24 July 2006 (UTC)


 * Do you mean the introduction in article now, or the opening sentence I proposed right above your message? I agree the introduction is cluttered and unclear, which is why I would like to change it. -- Schaefer (Talk) 22:05, 24 July 2006 (UTC)


 * Excuse the confusion, I meant how it existed at the time Tarcieri 08:41, 30 July 2006 (UTC)

And I'd like to note that the present intro paragraph fails to cover Vinge addressing both IA and AI. Also, it's grown rather long... Tarcieri 08:41, 30 July 2006 (UTC)
 * So I went ahead expounded upon Vinge's ideas. Too techie now? Tarcieri 08:47, 30 July 2006 (UTC)
 * Nope, not at all. I think your changes are an improvement. -- Schaefer (Talk) 18:02, 30 July 2006 (UTC)

Long time ago I proposed a lead based on most popular definitions: perhaps now that we have more users interested in that issue you'd like to revisit that proposal.--Piotr Konieczny aka Prokonsul Piotrus 14:35, 30 July 2006 (UTC)


 * I must oppose your proposed definition. It's overly Kurzweilian, in that it assumes a "smooth takeoff". Requiring rapid "societal, scientific and economic change" implies steady acceleration of technological advancement involving all of humanity, which is what you get if Vinge's Singularity doesn't happen. Accelerating progress is only relevant to the Singularity of Good and Vinge in that it supports the proposition that super-human intelligence may soon be technological feasible. Once you have created such intelligence, the old rules break down, because you have a mind that can repeatedly self-augment itself to the point of superintelligence (what Vinge calls "Powers" in his novels), and the availability of technological means becomes so disproportionate that the goal systems of the superintelligence have more sway over the future of the world than the human societies that have been inhabiting it thus far. If this event happens tomorrow, it's still the Singularity.
 * The problem, as I see it, is that we're really working with two completely incompatible meanings of the word Singularity. There's the intelligence explosion (Good and Vinge), and there's accelerating change (Kurzweil). The former was the original meaning of "the Singularity", but the latter is rapidly becoming more popular. The ideas are such disparate concepts that if you need to provide a definition that includes both of them, you get a sentence so vague that it does nothing to enhance the reader's understanding of either concept. In the interest of factual accuracy, I've implemented a fix that I'm not really happy with. It starts with as non-vague a sentence as I thought I could get away with, and then clarifies in the following two paragraphs the two presently competing meanings of the word. I'm open to suggestions on how to better handle this. -- Schaefer (Talk) 17:59, 30 July 2006 (UTC)

I'd agree with this and would state it in the opening sentence. Kurzweil's singularity is non-transcendent and results in a predictable future. Vinge's singularity is transcendent and results in a non-predictable future.

I have issues with the current opening. Sorry mate. 1. The technological singularity is a theoretical limitation on the prognosticative abilities of future studies not a predicted future event. That is, after all, why we're calling it a black hole. 2. Per the section following, the opening isn't nearly skeptical enough, given that after 40 solid years of trying we have no evidence that hard-AI is possibe, let alone the singularity. 3. Per the comments above, the opening doesn't distinguish between the two main types of singularity (Kurzeilian and Vingian) early enough. I would suggest something along the following lines.


 * A technological singularity (often the Singularity) is a hypothetical future point which is defined by almost infinite technological progress in an unprecedentedly short time. The nature and speed of the technological change is such that there is no methodology available to future studies now or at any time preceeding the event that can accurately or adequately predict the effect of this technological change on humankind. There are two schools of thought as to the nature of, and events leading to, the Singularity.  However since both schools of thought are describing an event that "passeth all human understanding" the actual technological singularity, should it ever occur, would undoubtedly be far different than either conception.

-- Tdewey 19:50, 29 October 2006 (UTC)

A few objections:


 * "Almost infinite" makes no sense. All finite numbers are infinitely far from infinity. I understand the intended meaning, but the phrase "almost infinite" is a non-encyclopedic hyperbole.
 * Unpredictability is not a defining characteristic of the Singularity, at least not as used by Good, Ulam, or Vinge. Where is this definition coming from? At best, it's a non-defining characteristic. Even with Kurzweil, the view is "Egads, we don't know what this Singularity thing is going to look like!" not "Let's give a name to our ignorance about the future. We shall call it the Singularity." Vinge discusses "tailoring" the Singularity to our desires, and all of the work of the Singularity Institute is based on effecting a positive Singularity, both of which make little sense if the Singularity is by definition incomprehensible.
 * The line "passeth all human understanding" is sarcastic and condescending. Doesn't belong in an encyclopedia article.
 * The last sentence makes the very bold an unattributed claim that everyone's predictions of the Singularity are wrong, inferring from the above unpredictability requirement. If people make predictions about the Singularity, they probably aren't using that definition. Nobody argues, "There will come a point in the future, the Singularity, such that all predictions about it and the events that follow will be wrong. I predict that following this Singularity, there will be..."

This being said, I'm all in favor of making an earlier distinction between the Vingean / Kurzweilian Singularity predictions (immediate takeoff following superintelligence vs. accelerating technological progress trends). -- Schaefer (talk) 20:31, 31 October 2006 (UTC)


 * "Passeth all human understanding" is a quote -- but no worries. Anyway Vinge is on record as saying that the lack of predictability (he terms it the "wall" or "prediction horizon") is part of what he means when he uses the term singularity.  . Also with respect to your comments below in the def of singularity section -- you're quite right -- Vinge also states that he did not use the term singularity with reference to anything going infinite.  Tdewey 01:38, 1 November 2006 (UTC)
 * Here's a chat with Vinge and Kurzweil -- both seem to use the black hole as unknowable/unpredictable metaphor. Tdewey 03:51, 1 November 2006 (UTC)
 * While re-reading Vinge's Marooned in Realtime came across the same definition (Chapter 11). The character Della Lu describes the Singularity as "a place where extrapolation breaks down and new models must be applied. And those new models are beyond our intelligence."  I think collectively my point is made and the definition and first paragraph needs to be updated. I also think the idea of lack of extrapolation (prediction) is the defining difference between a Vingian and Kurzweilian singularity Tdewey 22:22, 11 November 2006 (UTC)

I'd like to propose a replacement for the pre-TOC lead that is much shorter and suggests a break from existing predictions without unattributed speculation or implied acceptance of the Singularity:


 * A technological singularity (often the Singularity) is an enormous increase in technological progress featured in many futurological predictions and science fiction stories. Precise definitions vary, but the Singularity is most often characterized as a rapid, unprecedented departure from existing predictive models resulting from the creation of intelligences significantly smarter than present humans. The event, which I. J. Good first described in the 1960s as an "intelligence explosion", was greatly popularized in the 1980s by author Vernor Vinge and subsequently by futurist Raymond Kurzweil. Some transhumanists and artificial intelligence researchers advocate the Singularity as a feasible and worthy goal, while critics question whether the Singularity is desirable or even possible.

I think issues regarding the origin of the term (whether it was by analogy with black holes) are best left out of the introduction. -- Schaefer (talk) 00:43, 12 November 2006 (UTC)

disclaimer, where is it?
There should be a disclaimer at the beginning of the article stating the evolutional path of this article topic "Technological singularity" as originating as a fictional concept, as there is no evidence in human history of either our ability to create smarter than us technology or our inability to understand what we create with our intelligence. In fact there is evidence that our persistant intentional ignorance and lack of applying what we know that indicates reaching such machine intelligence is not a primary concern, but rather our learning to apply what we already know. I.E. what the world wants why is it not happening? We have the knowledge and the resources, but who's choice? Plausable deniability? Intentional ignorance is a powerful tool. And this doesn't even touch on abstraction physics and how it simplifies computing as the decimal system simplified math. The world is not flat, there be no dragons beyond the horizon, but the horizon is always there, to fictionalize upon. Occums razor.


 * Absurd. Computers have indeed solved problems too difficult for humans; I refer you to the many solutions provided by genetic algorithms and other evolutionary approaches, to say nothing of the Four color theorem's proof, or even the whole bloody field of modelling something in which the requisite computing power is far beyond human ability. --maru  (talk)  contribs 04:06, 24 July 2006 (UTC)


 * There is to a degree a fantastiche "Chariots of the Gods" character to this article. I agree that a disclaimer ... certainly more skepticism ... needs to be imbedded in the article.Of course machines have solved very difficult problems, but 1. they are a special class of problems (scalable to a kind of general, patterned, mental equivalence? HOW?), whose solutions were 2. anticipated and programmed by existing human intelligence. I've always enjoyed computers -- and programming them -- but: the machines invented nothing... they just cranked very hard like the slavish simulacra they are. Bravo, and bravo to the mere unsung (in your symphony) humans who did all the lab work; but: the article gives me no evidence that "strong AI" is any closer than when it was announced - with hubris - decades ago. We "meat machines" did pretty well with modeling, didn't we? I'm not aware of any algorithmic or heuristic breakthroughs that would eventuate this posited event horizon. Nor of any alien fallout to provide the necessary new wisdom. Neither did I detect any such evidence when I heard Kurzweil talk.Your use of the terms "Absurd" and "bloody" is mere ad hominem hand-waving, - further evidence of zealousness untroubled with mere evidence -  and your examples contribute no evidence of a "transcending breakthrough" whatsoever. We've enjoyed merely the speeding up of what we humans have done, since the renaissance liberated reason, for our better health and welfare: apply our intelligence to gather more intelligence. To paraphrase the old lady: Where's the beef? Finally, "new" paradigms are not better paradigms, they're merely novel. A carnival, like a shopping mall, can be loaded with novelties; they're fun, but hardly transcendent. -- Twang July 24, 2006.


 * I never claimed that we had had any "transcending breakthrough"; my point was that humans had indeed created machines which really were, perhaps in a bunch of rather narrow fields like chess and various scientific problems and suchlike, but neverthless genuinely better than humans (which undercut the first anon's argument). Even if you object that it is still "really" the humans doing it, that doesn't stick for evolutionary approaches where the goal but not the solution is specified by the human. And as for Strong AI... there is a saying, "It's not AI when a computer can do it." And I think you'd better look up ad hominem before attacking my literary style. --maru  (talk)  contribs 13:06, 24 July 2006 (UTC)

You are wrong, they are not better, only faster! The four color problem could be proven by a human if you spent enough time. Millions of years maybe, but all the computer did is try lots and lots of variations on the same thing. It never came up with anything new. 71.199.123.24 21:49, 27 July 2006 (UTC)

Many things originate as fictional concepts, but that doesn't make them fictional per se. Perhaps we should apply that disclaimer to the article on spacecraft? Throughout history scientists have worked towards ideas of their own that had no factual basis or tangible evidence, only theories and conjecture. Germs were discovered before they could be seen. Evidence in human history isn't required to develop something entirely new - that's the very opposite of progress.

In any case there is plenty of evidence to support the notion that intelligence can evolve from non-intelligence. If it can happen in a mucky primordial environment, it can probably happen a lot faster in computers. Fucube 21:58, 4 September 2006 (UTC)

I agree a disclaimer is needed. The arguments that we've created computers that have solved problems that humans cannot is specious. Mankind has long created things that can do other things that humans themselves cannot. This things are generally called "machines." The technological singularity (at least as envisioned by Vinge) is completely speculative and this should be noted somewhere. Indeed the fact that the article notes that the technological singularity has been callsed the "rapture for nerds" is evidence enough that we are entering the religous -- purely faith-based -- realm at this point. Additionally the Luddite reference should be dropped it comes off as an attack on the critics of the singularity. Tdewey 19:33, 29 October 2006 (UTC)

Why singularity? The concept has nothing to do with the meaning of the term.
In analysis a singularity is a value for which an otherwise continuous or differentiable function ceases to be defined. In curve theory, a curve has a singularity whenever its tangent vector has magnitude zero or is undefined. In ordinary differential equations, a singularity is a point where one of the functions defining the differential equation has a singularity, which usually affects the behavior of the phenomena being modelled by the equation.

A removable singularity can disappear, along with all of its effects, by a simple substitution. For example, $$f(x) = \frac{\sin(x)}{x}$$ has a singularity at $$x = 0$$, but the simple substitution: $$f(0):=0$$ removes the singularity.

A pole is a singularity that is the equivalent of dividing by zero once. It is well-understood, and relatively simple to handle in analysis. The singularity disappears when we multiply the function by $$x$$. A multiple pole of order $$n$$ is the equivalent of dividing by zero $$n$$ times. We can remove it by multiplying the function by $$x^n$$. An essential singularity is a singularity that cannot be removed by any of the techniques described above. Riemann proved that in the neighborhood of an essential singularity, a complex-valued function takes on every possible complex value except possibly two values. That is chaos-- the confusion described by the technology singularity is trite in comparison.

In physics, a singularity in a vector field forms a source, or a sink, or a combination of many of both or either. A massive body is a source of gravitational force. A bar magnet has a source of magnetic radiation at one end and a sink of magnetic radiation at the other. The singularities in a vector field go a long way towards determining the nature of the field. Singularities exist in all fields other than free space.

A continuous accumulation of technology such as that which the article describes is continuous, i.e. nonsingular. Therefore, calling such an accumulation a singularity seems like a non sequitur.

Incidentally, there is a simple pole in the mathematics of Terrence McKenna's Timewave. Thus if the singularity described in the article is that kind of singularity, then so be it, but at least mention that fact.

--Moly 15:21, 24 July 2006 (UTC)


 * Erm, isn't the technological singularity the point where invention/acquisition of technology is asymptotic with time? That is, where "our" understanding of the universe is suddenly sufficiently complete to render all "problems" soluble by technological means.  Whether such a thing exists is debatable, but it sounds like the singularity you describe.  And, anyway, understanding is really about information, and it's not clear that this behaves like the matter or energy that provide the examples of singularities that you describe.  But, at least as a metaphor, it works for me (maybe we need to state that?).  Cheers, --Plumbago 15:49, 24 July 2006 (UTC)

Definitions of "singularity" listed in the American Heritage Dictionary, fourth edition:
 * 1) The quality or condition of being singular.
 * 2) A trait marking one as distinct from others; a peculiarity.
 * 3) Something uncommon or unusual. [emphasis mine]
 * 4) Astrophysics. A point in space-time at which gravitational forces cause matter to have infinite density and infinitesimal volume, and space and time to become infinitely distorted.
 * 5) Mathematics. A point at which the derivative does not exist for a given function but every neighborhood of which contains points for which the derivative exists. Also called singular point.

Why is it so utterly fascinating that the technological singularity doesn't involve a function with a vertical assymptote? Why is it not equally bothersome that the technological singularity has nothing to do with infinite densities or the distortion of spacetime? Complex organisms have nothing to do with complex numbers from mathematics. The neighborhood I live in has nothing to do with the neighborhood of a point in a function. This isn't the first time that a word has been used for two different things. -- Schaefer (Talk)


 * It's termed a Singularity because like a black hole it is a point that we can't see into or beyond. Or as Vinge put it in The Coming Technological Singularity "It is a point where our old models must be discarded and a new reality rules."  Tdewey 01:37, 31 October 2006 (UTC)


 * I've heard this explanation before, and it's the one that the Singularity Institute uses to describe the term's origin, but I'm yet to see any real evidence that this is the analogy Vinge was shooting for. He calls it "a point where our old models must be discarded", yes, but this is perfectly consistent with the regular old-fashioned meaning of singularity as a synonym for peculiarity or uniqueness. I'm not aware of anywhere where he (or any earlier user of the term) explicitly mentions black holes. If you are, please share. -- Schaefer (Talk) 04:58, 31 October 2006 (UTC)


 * Fair enough--I probably picked it up from the institute. Interestingly I just came across this reference which might satisfy the math guy above.   Since I'm not a math major humour me -- if I understand the guy correctly he argues that Vinge is calling it a singularity because one definition of a singularity is a continuous function that becomes infinite.  When the singularity hits instead of the accumulation of finite knowledge over a finite time we would have the accumulation of infinite knowledge over a finite time.  Tdewey 16:51, 31 October 2006 (UTC)


 * The author of that page drastically mischaracterizes Vinge's argument. Vinge is clear that the Singularity is a consequence of superhuman intelligence, and only indirectly a consequence of Moore's law and technological growth. He invokes accelerating progress as evidence that superhuman intelligence will arrive soon, which is all he needs it for. Daniel Dennett once described the evolution of pre-DNA chemical replicators as a step ladder that could be thrown away when DNA gained monopoly status as the Earth's best replicating molecule. Vinge uses Moore's law similarly, as a disposable step ladder for getting from human intelligence to the Singularity. When the author you linked to writes about the doubling period reaching zero and generating an infinite amount of knowledge (whatever that means) in a finite amount of time, he/she's missing the point entirely. -- Schaefer (talk) 19:20, 31 October 2006 (UTC)
 * I agree with this but with refrence to my earlier comments 4 paragraphs up Vinge has stated that he used the term "by metaphor with the use of the term in general relativity." Tdewey 01:37, 1 November 2006 (UTC)

Fictional works
The fiction section of this article is getting ridiculous. Not only do most of the works contain no mention whatsoever of the Singularity, many of them were written/filmed before I. J. Good even proposed the Singularity. Claiming that, say, Arthur C. Clarke's "Childhood's End" involves the Singularity (It doesn't. As I recall, it turns into a big fantasy involving telepathic abilities.) is obvious original research. Now, if someone notable writes an article titled "Singularity Themes in Early Science Fiction", by all means, cite their opinions. But these unattributed assertions of Singularity depictions in fiction have to go.

One problem in the section is irrelevance: "One of the earliest examples of smarter-than-human AI in film is Colossus: The Forbin Project. In the 1969 film, a U.S. defense supercomputer becomes self-aware and unilaterally imposes peace on humanity." This is true, but not worthy of inclusion. Not every mention of smarter-than-human AI is a mention of the Singularity. Robots and AIs are common in sci-fi, and rarely are they depicted as being just plain stupid in comparison to humans. The Singularity is founded upon a belief that smarter-than-human intelligences will be able to improve their own minds better than their human designers, and once improved as such, will be able to make better improvements still, and so on. The Singularity is not the belief that things smarter than humans will be... um... smart... and, like, able to do really hard math and stuff.

Many of the examples provided (The Matrix, The Terminator) don't even have examples of AIs doing anything particularly clever (at least that I can recall), let alone anything superhumanly clever. How hard would it really be to rewrite The Matrix so that the villains are just humans with better kung-fu-downloading programs that let them dodge bullets? If that were the plot of the movie, would it strain your suspension of disbelief to think that humans could ever be as smart as the Agents? In the same way that it would strain it if a team comprising a tortoise and a chimpanzee outsmarted a global army of modern humans dedicated to killing them? Not only do the AIs fail to exhibit recursive self-improvement capabilities, they aren't even obviously smarter-than-human.

The other problem is verifiability. "Ken MacLeod describes the Singularity as "the Rapture for nerds" in his 1998 novel The Cassini Division." This is a fact. It's an uncited fact, and not as easy to verify as it should be, but it's a fact. I can, at the very least, Google around and get some confirmation that this is true. Other examples are harder to verify: "Some earlier science fiction works such as Arthur Clarke's Childhood's End, Isaac Asimov's The Last Question, and John W. Campbell's The Last Evolution also feature technological singularities." I, personally, can attest that Childhood's End has nothing to do with the Singularity, but only because I've read it. Why should you believe me over the article, or vice versa? There's no way to evaluate this claim other than to go out and read every book/story mentioned in the sentence, in their entireties, for yourself. Only then can you form an opinion on its truth. The statement is not verifiable.

The solution to this is simple: We need to stop comparing works with the Singularity, and restrict ourselves to reporting pre-existing comparisons published by notable commentators. We're encyclopedists, not literary critics. -- Schaefer (Talk) 00:55, 28 July 2006 (UTC)


 * The topic could be discussed in a book by book manner right here if different opinions are present. What is a pre-existing comparison by a notable commentator? I know    what you mean, but in the end, even if the commentator was Asimov himself, wasn't he just another biased human being? Wolflow 11:09, 29 August 2006 (UTC)


 * I don't see what you're getting at. You don't have to be neutral to be notable. Asimov is notable. Random Wikipedia editors aren't. -- Schaefer (Talk) 16:46, 29 August 2006 (UTC)

I.J.Good
Is I.J.Good really so important that he deserves mention as one whose writings a school of Singularity thought centers around? Sure, Vinge mentions him in his original essay a couple of times, but I've never heard him mentioned in any other Singularity discussion, and Vinge doesn't seem to have based his essay on Good's writings, either. Besides, the "ad infinitum" improvement in the description of a Vingean Singularity seems a bit silly - in my understanding the Vingean Singularity concentrates merely on the fact that humans will be surpassed by machines, while saying nothing about "infinite" improvement. -- Xuenay 23:22, 4 August 2006 (UTC)Xuenay

I've changed the intro around to remove the overemphasis on Good over Vinge, and removed the ad infinitum line. -- Schaefer (Talk) 20:28, 5 August 2006 (UTC)

Internet Singularity
I was browsing digg tonight and found a story on social computing (ugly article, needs a lot of work which i started) and they mentioned internet singularity. Such a notion, as defined at http://www.escholarlypub.com/digitalkoans/2006/01/28/gary-flakes-internet-singularity/ seems to be a sub-section of this article. I can't quite tell, though, whether to try and squeeze it in here or to give it its own article. I'm leaning towards giving it its own article and adding it to the disambiguation page for singularity, any thoughts? Mdanziger 07:07, 27 August 2006 (UTC)
 * I am not sure if this is notable enough, no hits on Google Print or Scholar.--Piotr Konieczny aka Prokonsul Piotrus 20:53, 28 August 2006 (UTC)
 * It is also a very sloppy use of "singularity". "Things are going to snowball" doesn't automatically translate into the kind of infinite singularity that this article is talking about. --Kaz 22:54, 30 August 2006 (UTC)

This seems to be very much missing from the main page. The fact that *humans* can start the singularity by using better software to organize themselves into clusters. Are humans today more intelligent than those thousands of years ago? No, but we invented mathematics and other scientific "thinking instruments" to reprogram our minds, and we have become better thinking machines. We might as well have turned into cyborgs without realizing it because it's "only" mental reprogramming. If humans can interact better with each other using internet tools (e.g. Wikipedia itself) that is *the same thing* and it is also comparable to an intelligent computer reprogramming itself in order to become more intelligent. The fact that this happens at human timescales which are a lot slower than digital reprogramming and the fact that it's about clusters of humans using computers as communication tools instead of it being about a single supercomputer does not matter at all, once we have a gradient, however slow it might be initially, the result is a runaway snowball effect. BTW I have tried to put his on the main page before, because it's so absolutely self evident to me, but it was reverted because of "apparent original research". What's up with this stupid regulation? Can't wikipedians think for themselves, have we become like those instances Lawrence Lessig warns us about? --anonymous user, 12 September 2006


 * You're free to have original ideas, but Wikipedia is not the place to introduce them. Write down your ideas, submit them to respectable publications, and if they get printed maybe they'll be referenced in a Wikipedia article. -- Schaefer (Talk) 23:14, 12 September 2006 (UTC)


 * See my comments above -- we need to more clearly define and deliminate the types of human/AI intelligence enhancement we're talking about. First of all humans are (with respect to the developed countries on average) more intelligent now than they were 20,000 or 2000 or even 50 years ago due to society wide improvements in education and nutrition and (more recently) due to the advent of television, video-games and other early-learning tools.  There is a good summary of both evidence and criticism of the evidence in the article "Brain Gain", Phyllida Brown, New Scientist, 2 March 2002 (also accessible on the web but that's probably a copyright violation). Tdewey 19:19, 29 October 2006 (UTC)

Razor
The razor picture is not valid. Furthermore, The Economist has been proved wrong by Kurzweil before. --Anon.


 * I haven't been involved in the recent edits on the processor speeds but I'd like to point out that while clock speeds have not increase recently, this is due to the switch of adding multiple cores to a processor. It is less expensive for them at the moment to increase performance by including multiple cores on the processor. This still follows Moore's law in processing speeds. Computers are still getting faster. If tomorrow they came out with a computer that ran twice as fast at only 1Ghz would you criticize Kurzweil for being wrong in this regard. Clock speed is not the be all to computing speed. It become sort of a gimmick. --Morphh (talk) 03:08, 15 October 2006 (UTC)


 * I'm not much but a causal observer here, but multicore strikes me as a quantitative improvement in processor technology rather than the qualitative improvements the exponential curves demand. But that aside, I still think the curves have currently broken down: have multicores really continued the 18 month halving or whatever the time period was that we observed previously? It doesn't look so. --Gwern (contribs) 17:02, 15 October 2006 (UTC)


 * Also don't forget the jump to 64bit processors. Multi-core and 64bit... these are architectural changes that take a certain amount of time to implement.  There have also been large improvements in hard disk (speed, storage, physical size), caching, & memory speeds.  Also consider how software and new techonogiles have played a part in this trend.  Grid computing in databases and applicatioins, new bus interfaces that let you tie computers together, blade systems - I can now take four two processor boxes and turn it into a redundent 8 processor system at a much lower cost.  Point is, Kurzweil never claimed that processor clock speeds would continue to increase.  He claimed computing power would follow the growth curve.  You can't take one small aspect of computers (processor clock speed for the here and now) and use this to justify a historical slowdown.  History will show if there was an actual lag but I don't think we can make that judgment today based solely on processor clock speed. Morphh (talk) 22:08, 15 October 2006 (UTC)


 * I think a better view (or at least an easier view) of the rate of change of computer performance is available by looking at the past and current performance on the top500 supercomputer site. Looking at their graphs the rate of change has been near-constant. [] Tdewey 06:40, 30 October 2006 (UTC)


 * We might also want to consider what has been achived in the lab. Just because we don't produce the advancement due to mass production costs does not mean we have not achived the gain.  We may skip certain advances as each advance adds a certain level of cost that may not be worth the effort or ready for prime time. Morphh (talk) 11:51, 30 October 2006 (UTC)


 * Seems reasonable. Look at the MDGRAPE-3 which hit a petaflop over the summer but has only limited general use -- it's essentially a lab computer. Tdewey 01:43, 1 November 2006 (UTC)


 * Ekhm. The top500 graph is logarythmic (1,10,100) :) Thus in linear scale it is indeed exponential.--Piotr Konieczny aka Prokonsul Piotrus 03:09, 2 November 2006 (UTC)

Greg Bear's short story Blood Music is squishy singularity
One of my favorite memes is the use of the Singularity as a solution to the Fermi paradox (www.faughnan.com/setifail.html). In this context I think Greg Bear's original short story Blood Music (not the novel) deserves pride of place. Published when he was a teenager in 1982 it explained the "great silence" as the result of an inescapable "squishy singularity" that was the end point for all worlds ... ref: http://faughnan.com/setifail.html#[3]


 * Greg Bear was quite a young man in 1982 but far from being a teenager. He turned 31 that year. Metamagician3000 00:52, 18 September 2006 (UTC)


 * I agree and added Fermi Paradox as a link to the see also section. Tdewey 04:21, 1 November 2006 (UTC)

Just a suggestion:

Stanislaw Lem's book "Imaginary Magnitudes" (written in 1985, I believe) contains a long essay titles "GOLEM XIV" about a super intelligent computer that bootstaps itself into ever higher levels of intelligence. I think mentioning this essay would add to the discussion on this page of Good's 1965 quotation.

-Jon Cohen

Omega Point
I'm confused as to why editors of this article insist on removing links to the WP article on Omega Point in the See Also section. Is Omega Point a completely unrelated concept that must not exist on this page? I await your reasoning. -- Unsigned


 * The link that I and several other editors have been repeatedly removing most recently is not to the WP article, but to an external website. A wikilink to Omega point is currently in the article's See also section and has been for some time. -- Schaefer (talk) 13:24, 7 November 2006 (UTC)

Robots do our work for us, so what do we do?
Crudely speaking, robots do our work for us, so what do we do? Do we have a page about this, because I'd like to link to it here, else can someone help me create this page? I know sci-fi books must have addressed this issue. Thanks, Peregrinefisher 06:17, 10 December 2006 (UTC)