En medborgare och matematiker ger synpunkter på samhällsfrågor, litteratur och vetenskap.
tisdag 29 oktober 2013
Det enklaste sättet är att använda definitionen
onsdag 23 oktober 2013
No nonsense - my reply to David Sumpter
- (1) the possibility and likelihood of a future intelligence explosion is not a scientific topic,
- (2) a future intelligence explosion is very unlikely.
- We make models, we make predictions and test them against data, we revise the model and move forward. You can be a Bayesian or a frequentist or Popperian, emphasise deductive or inductive reasoning, or whatever, but this is what you do if you are a scientist. Scientific discoveries are often radical and surprising but they always rely on the loop of reasoning coupled with observation.
- C1 = "an intelligent explosion is likely to happen around 2100".
- C2 = "under the IPCC AR4 emission scenario A2, global average temperature in 2100 will likely exceed the pre-industrial level by at least 3°C".
- C3 = "a newborn human who is immersed head-to-toe for 30 minutes in a tank of fuming nitric acid will not survive".
- If you are a Stone Age historical thinker called on to predict the future in a comprehensive report for your chief tribal planner, you must project the invention of the wheel or you will miss pretty much all of the action. Now, if you can prophesy the invention of the wheel, you already know what a wheel looks like, and thus you already know how to build a wheel, so you are already on your way.
tisdag 22 oktober 2013
Guest post by David Sumpter: Why "intelligence explosion" and many other futurist arguments are nonsense
söndag 20 oktober 2013
Om svenskkyrklig tankelättja
- utsett en [ny] form av andlig synd som orsak att fördela skam, skuld och låghet – den intellektuella synden att hävda att något i verkligheten förhåller sig på ett visst sätt snarare än på ett annat, att definiera sina begrepp och systematisera sina utsagor, att utesluta orimligheter, att ställa följdfrågor till fluffiga påståenden, att inte nöja sig med tomma hyllningar till ”ambivalensen” och ”det mångfacetterade”, utan i stället undra vad man ska ha de substanslösa påståendena till och hur de kan analyseras.
Begår man dessa fel har man enligt det nya syndaregistret brister i sin mentala utrustning. Man lider av svartvitt tänkande, man är fyrkantig, tvärsäker och har en torftig livstolkning. Liksom med den forna synden ges inga argument, utan man går direkt på personens okänslighet inför att saker ”inte behöver vara antingen eller utan kan vara både och”.
[...]
Dunbolsterteologerna som har makten i Svenska kyrkan tycks ha som ideal att ju mer viktlöst och konturlöst ett tankegods är, desto mer fullfjädrat. Om någon, troende eller icketroende, efterlyser motsägelsefria resonemang är de där och hötter med fingret. Livet, säger de, består av motsägelser och oförklarliga underligheter, så varför skulle inte yttranden om livet göra detsamma?
De ägnar decennier åt att förstå Gud men om någon avkräver dem ett svar med någorlunda precision på vad de kommit fram till och ställer ännu en fråga efter att den första luftpastejen presenterats, är denna av den förkrympta sort som ”vill ha tvärsäkra svar på allt”.
[...]
I teologisk dunbolsterdiskurs ska man inte använda språket för att förmedla en klar tanke, det som brukar vara en av dess viktigare funktioner. En klar tanke är nämligen en begränsande tanke, en tanke som föreslår att något är på ett visst sätt och som därför riskerar att utesluta andra möjligheter eller möjligheten att allt kan vara hur som helst. Det är ju helt riktigt att vad som helst kan beskrivas hur som helst och ur vilket perspektiv som helst, men det är inte samma sak som att det är eller kan vara hur som helst i den värld vi känner.
De tvärsäkert fyrkantiga som fått för sig att yttranden om verklighetens beskaffenhet, teologiska eller bara allmänt kunskapssökande, handlar om att få reda på något, som därför intresserar sig för beståndsdelar och inte bara stumt förundras över helheten, säger sig dunbolsterteologerna vara ”väldigt rädda för”.
- Att beskylla Antje Jackelén för tankelättja är Lena Anderssons roligaste skämt hittills.
- Märkligt tycker jag, att när vi får en ÄB som har ett skarpare huvud än de flesta av oss, och som är respekterad i vetenskapliga kretsar långt utanför teologin, så får hon kritik för att vara luddig! — ...tankelättja och Antje? Hur går det ihop?
torsdag 17 oktober 2013
Norge har en ny regering
tisdag 15 oktober 2013
Reading the Hanson-Yudkowsky debate
- Hanson and Yudkowsky are both driven by an eager desire to understand each other's arguments and to pinpoint the source of their disagreement. This drives them to bring a good deal of meta into their discussion, which I think (in this case) is mostly a good thing. The discussion gets an extra nerve from the facts that they both aspire to be (as best as they can) rational Bayesian agents and that they are both well-acquainted with Robert Aumann's Agreeing to Disagree theorem, which states that whenever two rational Bayesian agents have common knowledge of each other's estimates of the probability of some event, their estimates must in fact agree.2 Hence, as long as Hanson's and Yudkowsky's disagreement persists, this is a sign that at least one of them is irrational. Especially Yudkowsky tends to obsess over this.
- All scientific reasoning involves both induction and deduction, although the proportions can vary quite a bit. Concerning the difference between Hanson's and Yudkowsky's respective styles of thinking, what strikes me most is that Hanson relies much more on induction,3 compared to Yudkowsky who is much more willing to engage in deductive reasoning. When Hanson reasons about the future, he prefers to find some empirical trend in the past and to extrapolate it into the future. Yudkowsky engages much more in mechanistic explanations of various phenomena, and combining then in order to deduce other (and hitherto unseen) phenomena.4 (This difference in thinking styles is close or perhaps even identical to what Yudkowsky calls the "outside" versus "inside" views.)
- Yudkowsky gave, in his very old (1996 - he was a teenager then) paper Staring into the Singularity, a beautiful illustration based on Moore's law of the idea that a self-improving AI might take off towards superintelligence very fast:
- If computing speeds double every two years,
what happens when computer-based AIs are doing the research?
Computing speed doubles every two years.
Computing speed doubles every two years of work.
Computing speed doubles every two subjective years of work.Two years after Artificial Intelligences reach human equivalence, their speed doubles. One year later, their speed doubles again.
Six months - three months - 1.5 months ... Singularity.
- In Chapter 18 (Surprised by Brains), Yudkowsky begins...
- Imagine two agents who've never seen an intelligence - including, somehow, themselves - but who've seen the rest of the universe up until now, arguing about what these newfangled "humans" with their "language" might be able to do
- They made fun of Galileo, and he was right.
They make fun of me, therefore I am right. - The central concept that leads Yudkowsky to predict an intelligence explosion is the new positive feedback introduced by recursive self-improvement. But it isn't really new, says Hanson, recalling (in Chapter 2: Engelbart as UberTool?) the case of Douglas Engelbart, his 1962 paper Augmenting Human Intellect: A Conceptual Framework, and his project to create computer tools (many of which are commonplace today) that will improve the power and efficiency of human cognition. Take word processing as an example. Writing is a non-negligible part of R&D, so if we get an efficient word processor, we will get (at least a bit) better at R&D, so we can then device an even better word processor, and so on. The challenge here to Yudkowsky is this: Why hasn't the invention of the word processor triggered an intelligence explosion, and why is the word processor case different from the self-improving AI feedback loop?
An answer to the last question might be that the writing part of the R&D process is not really all that crucial to the R&D process, taking up maybe just 2% of the time involved, as opposed to the stuff going on in the AI's brain which makes up maybe 90% of the R&D work. In the word processor case, no more than 2% improvement is possible, and after each iteration the percentage decreases, quickly fizzling out to undetectable levels. But is there really a big qualitative difference between 0.02 and 0.9 here? Won't the 90% part of the R&D taking place inside the AI's brain similarly fizzle out after a number of iterations of the feedback loop, with other factors (external logistics) taking on the role as dominant bottlenecks? Perhaps not, if the improved AI brain figures out ways to improve the external logistics as well. But then again, why doesn't that same argument apply to word processing? I think this is an interesting criticism from Hanson, and I'm not sure how conclusively Yudkowsky has answered it.
fredag 11 oktober 2013
En anspråksfull bok om matematik och evolutionsbiologi
- Groundbreaking mathematician Gregory Chaitin gives us the first book to posit that we can prove how Darwin's theory of evolution works on a mathematical level.
For years it has been received wisdom among most scientists that, just as Darwin claimed, all of the Earth's life-forms evolved by blind chance. But does Darwin's theory function on a purely mathematical level? Has there been enough time for evolution to produce the remarkable biological diversity we see around us? It's a question no one has answered - in fact, no one has even attempted to answer it until now.
lördag 5 oktober 2013
Om de två kulturerna
- Världen behöver människor som kombinerar naturvetenskap, samhällsvetenskap och humaniora. Kursen vill uppmuntra dig att bli en sådan människa.
Naturvetenskap+ är en modern gymnasiekurs (Naturvetenskaplig specialisering) som behandlar evolutionsbiologi, kognitions- och socialpsykologi, logik, statistik, sannolikhet, retorik, argumentationsanalys, vetenskapshistoria och -teori, olika vetenskapliga metoder, filosofi och etik, samhälle, media och klimat. Med mer.
- Olle Häggström tror att gymnasieskolans kurs har precis rätt tänk. För att klara utmaningar som mänskligheten står inför måste naturvetenskap kombineras med filosofi, samhällsvetenskap och humaniora.
- Det skulle vara mycket värt om Häggström eller någon annan av samma kaliber fördjupar diskussionen och författar en mera klargörande text om hur det egentligen ligger till mellan Vetenskap och Humaniora inte bara inom den akademiska världen utan vad ett närmande mellan dessa discipliner skulle kunna innebära för samhället i stort.
- In his commentary on my essay “Science is Not your Enemy,” Leon Wieseltier writes, “It is not for science to say whether science belongs in morality and politics and art.” I reply: It is not for Leon Wieseltier to say where science belongs. Good ideas can come from any source, and they must be evaluated on their cogency, not on the occupational clique of the people who originated them.
Wieseltier’s insistence that science should stay inside a box he has built for it and leave the weighty questions to philosophy is based on a fallacy. Yes, certain propositions are empirical, others logical or conceptual or normative; they should not be confused. But propositions are not academic disciplines. Science is not a listing of empirical facts, nor has philosophy ever confined itself to the non-empirical.
Why should either discipline stay inside Wieseltier’s sterile rooms? Does morality have nothing to do with the facts of human well-being, or with the source of human moral intuitions? Does political theory have nothing to learn from a better understanding of people’s inclinations to cooperate, aggress, hoard, share, work, empathize, or submit to authority? Is art really independent of language, perception, memory, emotion? If not, and if scientists have made discoveries about these faculties which go beyond received wisdom, why isn’t it for them to say that these ideas belong in any sophisticated discussion of these topics?
- Wieseltier doubts my sincerity when I note that the benefits of a consilience between the sciences and humanities go both ways. He bizarrely translates my observation that “the sciences [can] challenge their theories with the natural experiments and ecologically valid phenomena that have been so richly characterized by humanists” as “scientists think well and humanists write well." So let me explain the observation with a few examples of two-way traffic just from my own research. Theories of the mental representation of the visual field must accommodate the fact that the reproduction of linear perspective in painting is not cognitively natural but was a late invention in the history of art. A major theory of auditory scene analysis receives important confirmation from the phenomenon of virtual polyphony in music. Theories of mental imagery must account for the observations by analytical philosophers that images lack the geometric detail of visual percepts, and that even with such detail they would be unsuited to represent abstract concepts. Research in psycholinguistics depends heavily on philological scholarship on the history of words and grammatical constructions in English and other languages. Theories of cognitive categorization begin with contrasting views on the nature of concepts from Aristotle and Wittgenstein. The science of human aggression has learned immeasurably from the history of crime, war, genocide, criminal punishment, and religious and cultural attitudes toward violence. Examples could be multiplied from the research of others.
tisdag 1 oktober 2013
Michael Chorost har en dator i sitt huvud
- These technologies are only for people with serious medical conditions. Warwick consistently neglects this important qualification. Upon having used his implant to pilot an electric wheelchair he comments, "I told everyone that this would ultimately mean, in the future, we should be able to drive a car around by picking up signals directly from the brain, and change direction just by thinking about it, right, left, and so on." For disabled people, this is an exciting prospect. But the rest of us already do drive with signals from our brains, picked up and executed perfectly by our arms and legs.
Warwick might respond that having direct nervous control over an automobile would let us drive it better, since we wouldn't have to move around our heavy meat arms and legs. But then issues of practicality and safety come into play. (s 176-177)
- Dear reader with organic ears, you do not percieve the world as it really is, and you never will. But be glad: a truly faithful rendition of the auditory world would either be flat and shallow, or unbearably painful. (s 150-151)