--- Day changed Sat Dec 06 2014
00:27 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has joined ##hplusroadmap
01:34 < ebowden> paperbot: http://www.sciencedirect.com/science/article/pii/S0168010214002119
01:34 < paperbot> http://diyhpl.us/~bryan/papers2/paperbot/%0A%20Disruption%20of%20striatal-enriched%20protein%20tyrosine%20phosphatase%20%28STEP%29%20function%20in%20neuropsychiatric%20disorders%0A%20.pdf
01:46 -!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has joined ##hplusroadmap
01:49 -!- ebowden_ [~ebowden@CPE-121-223-157-226.lns2.bat.bigpond.net.au] has joined ##hplusroadmap
01:49 -!- ebowden [~ebowden@CPE-121-223-157-226.lns2.bat.bigpond.net.au] has quit [Read error: Connection reset by peer]
02:22 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
02:54 -!- eudoxia [~eudoxia@r186-50-231-92.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap
03:24 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has quit [Ping timeout: 264 seconds]
03:55 -!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has quit [Ping timeout: 250 seconds]
03:55 -!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has joined ##hplusroadmap
04:20 -!- _0bitcount [~big-byte@81.61.34.185.dyn.user.ono.com] has joined ##hplusroadmap
04:21 -!- eudoxia [~eudoxia@r186-50-231-92.dialup.adsl.anteldata.net.uy] has quit [Quit: leaving]
05:03 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-qltimvzfifgmqtbh] has joined ##hplusroadmap
06:01 -!- eudoxia [~eudoxia@r186-50-231-92.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap
06:24 -!- Zinglon [~Zinglon@ip565f6f48.direct-adsl.nl] has joined ##hplusroadmap
06:37 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 264 seconds]
06:41 -!- Vutral [~ss@mirbsd/special/Vutral] has quit [Ping timeout: 260 seconds]
06:45 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
06:53 -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap
07:09 < kanzure> fenn: universal psychometrics http://users.dsic.upv.es/~flip/papers/TR-upsycho2012.pdf
07:17 -!- eudoxia [~eudoxia@r186-50-231-92.dialup.adsl.anteldata.net.uy] has quit [Quit: wow such lunchtime]
07:31 -!- ebowden_ [~ebowden@CPE-121-223-157-226.lns2.bat.bigpond.net.au] has quit [Remote host closed the connection]
07:35 < kanzure> hehe "While the results show that some improvement can be shown because of the enriched human context, animal abilities are limited by their genes. What about machines? Things are more complex here. Theoretically, it is possible to construct a machine such that it is dumb until an appropriate “training” signal is received, when it changes into another state where it becomes intelligent. In fact, for any universal Turing machine (UTM) ...
07:35 < kanzure> ... there is an input —a program— such that the machine becomes any other machine, e.g., a machine with any degree of actual intelligence. So we could loosely say that any UTM has maximal potential intelligence (and the same applies for any cognitive ability). However, if we construct a second machine such that this second state is accessed much more easily (without the need of a very specific input), we can intuitively say that the ...
07:35 < kanzure> ... second machine has more potential intelligence."
07:51 -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has quit [Remote host closed the connection]
08:08 -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has joined ##hplusroadmap
08:09 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has quit [Ping timeout: 245 seconds]
08:11 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has joined ##hplusroadmap
08:27 -!- yashgaroth [~ffffff@2606:6000:cb85:6a00:1822:2af0:a16b:35ec] has joined ##hplusroadmap
09:21 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Ping timeout: 260 seconds]
09:25 -!- _0bitcount [~big-byte@81.61.34.185.dyn.user.ono.com] has quit [Quit: Leaving]
09:28 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
09:49 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Ping timeout: 258 seconds]
09:56 -!- juri_ [~juri@vpn166.sdf.org] has quit [Ping timeout: 265 seconds]
10:02 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap
10:08 -!- juri__ [~juri@vpn166.sdf.org] has joined ##hplusroadmap
10:22 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Ping timeout: 260 seconds]
10:58 -!- juri__ [~juri@vpn166.sdf.org] has quit [Ping timeout: 264 seconds]
11:09 -!- Burnin8 [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has quit [Read error: Connection reset by peer]
11:09 -!- Burnin8 [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has joined ##hplusroadmap
11:10 -!- _0bitcount [~big-byte@81.61.34.185.dyn.user.ono.com] has joined ##hplusroadmap
11:35 -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:4dd9:ea8:d228:4bbc] has joined ##hplusroadmap
11:35 -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:4dd9:ea8:d228:4bbc] has quit [Changing host]
11:35 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap
12:07 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Quit: ...unyaaa ~~~]
12:12 < kanzure> jrayhawk: this thing argues that human cognitive ability is sexually selected because it indicates disease resistance http://phthiraptera.info/Publications/47267.pdf
12:13 < kanzure> i'm not really sure if that's strong enough though
12:13 < kanzure> i would imagine that not being dead would be a good enough indicator
12:27 < bbrittain> paperbot: http://www.sciencedirect.com/science/article/pii/S0304420313001242
12:28 < paperbot> http://diyhpl.us/~bryan/papers2/paperbot/%0A%20Biogenic%20halocarbons%20in%20young%20Arctic%20sea%20ice%20and%20frost%20flowers%0A%20.pdf
12:30 < bbrittain> :/
12:31 < bbrittain> new plan, never graduate, keep my school paper access
12:31 < bbrittain> nmz787: ^
12:33 -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:4dd9:ea8:d228:4bbc] has joined ##hplusroadmap
12:33 -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:4dd9:ea8:d228:4bbc] has quit [Changing host]
12:33 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap
12:41 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Ex-Chat]
12:53 -!- poppingtonic [~poppingto@unaffiliated/poppingtonic] has joined ##hplusroadmap
12:58 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
13:05 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has joined ##hplusroadmap
13:06 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Quit: ...unyaaa ~~~]
13:08 < kanzure> that is a good plan
13:17 -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:8590:deb4:7ee7:a03] has joined ##hplusroadmap
13:17 -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:8590:deb4:7ee7:a03] has quit [Changing host]
13:17 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap
13:29 -!- drewbot [~cinch@ec2-54-145-76-87.compute-1.amazonaws.com] has quit [Remote host closed the connection]
13:29 -!- drewbot [~cinch@ec2-54-163-179-128.compute-1.amazonaws.com] has joined ##hplusroadmap
13:30 < kanzure> "Artificial selection on relative brain size in the guppy reveals costs and benefits of evolving a larger brain" http://users.dsic.upv.es/~flip/papers/MINDS-MACHINES-potential.pdf
13:33 < kanzure> whoops, how about http://www.iee.unibe.ch/unibe/philnat/biology/zoologie/content/e7493/e7854/e355359/e373691/KotrschalCurrentBiol2013.pdf
13:33 < poppingtonic> .title
13:33 < yoleaux> poppingtonic: Sorry, that doesn't appear to be an HTML page.
13:33 < poppingtonic> oops
13:35 < kanzure> "Relative brain size was already 9% larger in the upward- compared to the downward-selected lines after two generations of selection"
13:35 < kanzure> hehe just two generations
13:37 -!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has quit [Quit: leaving]
13:37 -!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has joined ##hplusroadmap
13:38 -!- bombuzal [~bombuzal@unaffiliated/bombuzal] has joined ##hplusroadmap
13:45 -!- bombuzal [~bombuzal@unaffiliated/bombuzal] has quit [Quit: Leaving]
13:45 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Ex-Chat]
13:50 -!- namespace [~user@184.12.107.90] has joined ##hplusroadmap
14:03 < poppingtonic> i'm gonna have to read these...
14:04 < kanzure> also try http://diyhpl.us/~bryan/papers2/neuro/Human-specific%20transcriptional%20networks%20in%20the%20brain.pdf
14:04 -!- yashgaroth [~ffffff@2606:6000:cb85:6a00:1822:2af0:a16b:35ec] has quit [Quit: Leaving]
14:09 < jrayhawk> "in fact, IQ-heritability studies measure
14:09 < jrayhawk> the heritability of disease resistance to some ex-
14:09 < jrayhawk> tent"
14:09 < jrayhawk> that was the only insightful thing in this entire paper
14:15 < kanzure> can you specify how shit their hypothesis was?
14:16 < kanzure> i'm really strugging to postulate credible selection mechanisms for human cognitive abilities in evolutionary history
14:16 < kanzure> *struggling
14:41 -!- strangewarp_ [~strangewa@c-76-25-206-3.hsd1.co.comcast.net] has quit [Quit: austerity chic brand destruction]
14:46 -!- pete4242 [~smuxi@boole.london.hackspace.org.uk] has joined ##hplusroadmap
14:48 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 245 seconds]
14:53 -!- _0bitcount [~big-byte@81.61.34.185.dyn.user.ono.com] has quit [Quit: Leaving]
14:56 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
14:56 -!- strangewarp [~strangewa@c-76-25-206-3.hsd1.co.comcast.net] has joined ##hplusroadmap
14:57 -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has quit [Remote host closed the connection]
15:01 -!- ebowden [~ebowden@CPE-121-223-157-226.lns2.bat.bigpond.net.au] has joined ##hplusroadmap
15:03 -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has joined ##hplusroadmap
15:16 < andytoshi> kanzure: leading idea is politics no?
15:16 < andytoshi> outsmart everyone to be the harem master
15:18 < kanzure> but where did that come from?
15:18 < kanzure> here are the mechanisms that wikipedia proposes: http://en.wikipedia.org/wiki/Evolution_of_human_intelligence#Models
15:20 < kanzure> "He argues that the manifestations of intelligence such as language, music and art did not evolve because of their utilitarian value to the survival of ancient hominids. Rather, intelligence may have been a fitness indicator." (lame)
15:20 < andytoshi> that's lame
15:20 < namespace> Lame.
15:20 < andytoshi> intelligence is a fully general advantage
15:20 < kanzure> "Intelligence as a disease resistance signal"
15:20 -!- poppingtonic [~poppingto@unaffiliated/poppingtonic] has quit [Read error: Connection reset by peer]
15:20 < namespace> "Intelligence is a fitness indicator" is like saying "Being able to grow arms is a fitness indicator"
15:21 < kanzure> "The group benefits of intelligence have apparent utility in increasing the survival potential of a group"
15:21 < kanzure> yeah so those were the best ideas known to man
15:21 < kanzure> for why human cognitive ability was selected for
15:21 < kanzure> pretty pathetic
15:21 < namespace> Well to be fair.
15:21 < namespace> Those are so bad that we're looking at something on the order of "Nobody really thought about it."
15:21 < namespace> Rather than "Somebody seriously thought about it and they're an idiot."
15:22 < kanzure> these have references, so they thought about it enough to put words onto paper
15:22 < namespace> D:
15:22 < namespace> -1 Faith in humanity
15:23 < kanzure> andytoshi: i was thinking about some mechanism earlier today related to "bride price", where groups of humans bought phenotypes/genotypes with resources
15:24 < kanzure> andytoshi: and maybe cognitive abilities are selected by that sort of group-to-group genetic trade over time
15:24 < kanzure> however, this is not a thorough explanation and it does not feel right to me yet
15:26 < andytoshi> kanzure: well, some of this did happen in the middle ages ... i think this is why white women are hotter than white men?
15:26 < andytoshi> because they would "buy" brides by being royalty?
15:27 < kanzure> women are more attractive because you and i are heterosexual
15:27 < andytoshi> hmm, that might be it
15:27 < kanzure> nice try
15:27 < andytoshi> ;)
15:27 < namespace> kanzure: :P
15:27 < kanzure> "bride price" was happening long before the middle ages
15:27 < namespace> I was about to say "They are?"
15:28 < kanzure> code of hammurabi from ~5000 years ago has rules written in stone about buying mates
15:29 < andytoshi> ok, so, back to intelligence ... istm that no matter what you are doing, it is better to do it more intelligently
15:30 < kanzure> you may be interested in this paper sometime: http://users.dsic.upv.es/~flip/papers/TR-upsycho2012.pdf
15:30 < kanzure> (universal psychometrics for turing machines)
15:31 < kanzure> general intelligence certainly confers lots of survival benefit compared to not having a brain
15:32 < andytoshi> i'm reading the wiki article .. it actually surprises me that intelligence is not a slam dunk
15:32 < kanzure> instead of waiting around for millions of years to get some other kind of adaptation in a species, adaptive behavior in a single lifetime is much better
15:32 < kanzure> what do you mean intelligence as a slam dunk?
15:33 < andytoshi> i mean, it is so beneficial that it should appear in any species with extra resources to support it
15:33 < andytoshi> and since it leads to having extra resources you get a feedback loop
15:33 < kanzure> well, brains in general do appear in many species
15:34 < andytoshi> hmm, yeah, i guess that's evidence that i'm right in a really limited way
15:34 < namespace> General intelligence will not help you much as a fish?
15:34 < kanzure> many properties of the human brain are highly conserved in chimpanzees and gorilla, which are ancestors from more than a million years back
15:35 < kanzure> and other more basic brain parts are conserved in mammals in general too
15:35 < namespace> I mean from what we've seen with humans, general intelligence seems to have a very long setup time compared to a canned instincts/brain sort of deal, so if you have relatively short lifespans its not worth the investment, in fact it's a net negative.
15:36 < andytoshi> kanzure: so this gives credence to the "politics" theory (which on the wiki page i think is dunbar's claim)
15:36 < kanzure> just because it's beneficial does not mean that natural selection will stumbe into it
15:36 < kanzure> *stumble
15:37 < namespace> Like part of why humans work is that we live a relatively long time.
15:37 < andytoshi> kanzure: sure, but i would suggest that if it did, it'd latch on unconditionally..
15:37 < andytoshi> namespace has some good arguments why i'm wrong
15:37 < kanzure> politics is an okay direction but i'd have to see more,
15:37 < kanzure> dunbar's version is pretty lame
15:38 -!- CharlieNobody [~CharlieNo@97-85-246-63.static.stls.mo.charter.com] has joined ##hplusroadmap
15:38 < andytoshi> long setup time, big brains make childbirth hard (and babies come out still useless), the brain is a massive resource hog..
15:38 < kanzure> he posits some direct proportionality to brain volume
15:38 < kanzure> everyone focuses on brain size for some reason
15:38 < andytoshi> oh bleh
15:38 < namespace> kanzure: I wasn't saying that.
15:38 < namespace> kanzure: I just said that it takes a very long time to go from baby to adult human, maybe this is a necessary component?
15:38 < kanzure> there's definitely high resource costs to human brains
15:39 < namespace> Like if you're a fish, you don't have time to learn how not to get eaten, you need to know that from day one.
15:39 < namespace> The ocean is too dangerous for it to be worth learning better evasive tactics with general intelligence.
15:39 < kanzure> selection mechanisms are things like, "in human evolutionary history, aliens showed up on the planet and administered general intelligence tests, and then had selective mating"
15:40 < namespace> kanzure: Hmm, have we dug up any ancient SAT's yet? :P
15:40 -!- Baube [~Baube@64.229.103.26] has joined ##hplusroadmap
15:40 < kanzure> right, that's a highly implausible one
15:40 -!- pete4242 [~smuxi@boole.london.hackspace.org.uk] has quit [Remote host closed the connection]
15:41 < kanzure> "Dunbar theorized that "this limit is a direct function of relative neocortex size, and that this in turn limits group size ... the limit imposed by neocortical processing capacity is simply on the number of individuals with whom a stable inter-personal relationship can be maintained.""
15:41 < kanzure> volume of neocortex -_-
15:41 < namespace> Not to mention that, you have octopuses, which are relatively intelligent as far as seacritter go, but they haven't taken over the world or anything because they're stuck underwater with prehensile limbs.
15:41 < namespace> There's a lot of important hardware that needs to be comorbid with intelligence for world domination.
15:42 < kanzure> an octopus does not have general intelligence as far as we know
15:42 < kanzure> so i don't know why you are mentioning them
15:42 < namespace> Right.
15:42 < namespace> They're the most intelligent non-mammal IIRC?
15:42 < namespace> Was going with my fish theme.
15:43 < kanzure> i don't know what you're taking about at all
15:43 < kanzure> *talking
15:44 < namespace> andytoshi had a hypothesis that general intelligence is such a benefit that we should expect to see it in everything capable of supporting it. I'm pointing out that even though octopi are relatively intelligent, it doesn't actually matter if they get more intelligent because they don't really have enough ability to manipulate their environment for it to matter.
15:44 < namespace> I didn't really explain it very well I'm sorry.
15:45 < kanzure> oh, well andytoshi's hypothesis is wrong because obviously only one species so far has general intelligence
15:45 < kanzure> anthropics, yo
15:45 < namespace> Yup.
15:45 < namespace> If we lived in the universe where it was true, that is what we would observe.
15:45 < namespace> At the same time.
15:45 < kanzure> so nothing about manipulating the environment or whatever
15:45 < namespace> Knowing it's wrong isn't really enough in some cases, knowing *why* something is wrong can be important information.
15:45 < kanzure> parapalegics are still intelligent
15:49 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has quit [Ping timeout: 255 seconds]
15:50 < namespace> Parapalegics are such as a function of their environment, if you take parapalegic eggs or sperm and make babies with them you will get healthy children.
15:50 < andytoshi> there was a lesswrong article about the politics thing, i can't seem to find it, it should have some good references..
15:51 < andytoshi> something about one monkey betraying another after pretending to be a servant and eliezier was like "a human would've seen that coming a mile away [and not gotten killed]"
15:52 < kanzure> monkeys are just as old, or even older, than we are
15:53 < kanzure> the ability to buy genotypes/phenotypes is something that hasn't happened anywhere else in evolutionary history
15:54 < kanzure> although maybe that only works for the past 70,000 to 100,000 years of human evolution
15:54 < kanzure> before which hominid remains are not seen with valuables or posessions as much
15:55 < kanzure> oh, 500k years actually. hmm. that's more likely.
15:55 < kanzure> http://en.wikipedia.org/wiki/Behavioral_modernity
15:55 < kanzure> oh, that's 50k. uh...
15:55  * namespace idly wonders if there's enough information in the universe left to determine this to our satisfaction
15:56 < kanzure> uh sure there is
15:56 < namespace> Yes of course what am I even saying.
15:56 < kanzure> all you have to do is come up with some plausible mechanisms and then think hard about them
15:56 < kanzure> then you can test on monkeys
15:56 < namespace> I doubt we'll actually get to the testing on monkeys stage.
15:56 < namespace> Because ethics.
15:57 < kanzure> yes, raising monkeys is so fucking unethical
15:57 < namespace> Do you think an ethics review board would okay it if the explicit goal was to get human level intelligence out of it?
15:57 < namespace> I'm not saying I agree with that, but I doubt they would.
15:58 < kanzure> an ethics review board is not the crown jewel of being ethical -_-
15:58 < namespace> When I said 'ethics' it was shorthand for such.
15:59 -!- Zinglon [~Zinglon@ip565f6f48.direct-adsl.nl] has quit [Ping timeout: 250 seconds]
15:59 < kanzure> besides, you hardly know the details of my idea, so from my point of view you're the one being unethical here
16:00 < namespace> Um what?
16:00 < kanzure> you heard me :)
16:00 < namespace> I did.
16:00 < namespace> And what you said seems to be relying on assumptions that are not true.
16:00 < kanzure> i believe it is unethical to appeal to authority (especially "ethics" authorities)
16:00 < kanzure> especially for ideas that haven't even been elaborated
16:00 < kanzure> "authorities"
16:01 < namespace> That statement was in the same vein as if you'd said you had a cure for aging and I'd said:
16:01 < namespace> "Oh I bet the FDA will skewer you over it."
16:01 < namespace> And then you go:
16:01 < namespace> "APPEAL TO AUTHORITY"
16:01 < namespace> And I go:
16:01 < namespace> "Um that's not what I-"
16:01 < namespace> "UNETHICAL"
16:01 < kanzure> yes and?
16:02 < namespace> I'm saying that you're attacking me for no good reason?
16:02 < kanzure> in general i tend to attack bad ideas
16:03 < namespace> Cynical observations about how society will treat things are now Bad Ideas?
16:04 < kanzure> no, it is a bad idea for you to suggest some threat vector that is totally superfluous at this stage of idea development
16:04 < namespace> Well okay if you insist.
16:05 -!- poppingtonic [~poppingto@unaffiliated/poppingtonic] has joined ##hplusroadmap
16:05 < kanzure> also, it would be super unfortunate if the fda started sending nastygrams and lawsuit notices over just talking about nutrition
16:05 < kanzure> or if ethics review boards were to have jurisdiction over ideation
16:06 < kanzure> namespace: http://web.archive.org/web/20130709183013/http://www.maxmore.com/extprn3.htm
16:06 < namespace> kanzure: Um, you're starting to take what I said into wholly new territory from where I'd started. Which was that I'm not sure we can in practice test on monkeys, which I do not think is a bad observation to make?
16:06 < kanzure> namespace: http://diyhpl.us/wiki/declaration
16:06 < kanzure> anyone can own a monkey man
16:07 < namespace> ...
16:07 < namespace> How does this monkey man feel about this?
16:07 < kanzure> i meant to put a comma after monkey
16:07 < kanzure> and i did not
16:08 < namespace> *the monkey man
16:11 < kanzure> andytoshi: so one idea is that if there is a good psychometric test for general intelligence, or a good definition, then you could just run a genetic algorithm on some fancypants hardware for a while and see what happens
16:12 < kanzure> andytoshi: alternatively, if you knew what the selection mechanisms may have been in human history, then you could replicate those sorts of effects on software programs
16:12 -!- rk[1] [~rak@opensource.cse.ohio-state.edu] has quit [Remote host closed the connection]
16:12 < kanzure> you could even have things approximating gene trading (software trading) between different agents in your pool
16:13 < andytoshi> those both sound really computationally hard because humans have been manipulating their environment since forever
16:14 < kanzure> the above paper about universal psychometric testing for universa turing machines has some suggestions for the testing aspect
16:14 < kanzure> but the downside is that you would need to make a bunch of tests using their suggested format (it's related to kolmogorov complexity)
16:16 < kanzure> maybe you could arbitrarily generate tests automatically
16:18 < kanzure> er, this is clearly the "iterative" approach, i don't know why i just suggested that
16:18 < kanzure> i retract my test generation statement
16:18 < kanzure> except to the extent that you pick "good" tests, and then just have a collection of "good" tests
16:24 < namespace> Burnin8 wants us to move the convo into #lesswrong to get the crap out of there
16:24 < namespace> I'm not sure this is actually a good idea, personally.
16:25 < Burnin8> depends on whether you want more participants or not
16:25  * namespace was done with it, personally
16:25 < namespace> I got the impression I was giving Kanzure the impression I'm dumb, so I stopped.
16:37 -!- eudoxia [~eudoxia@r190-134-51-29.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap
16:38 -!- eudoxia [~eudoxia@r190-134-51-29.dialup.adsl.anteldata.net.uy] has quit [Client Quit]
16:40 -!- Baube [~Baube@64.229.103.26] has quit [Ping timeout: 244 seconds]
16:49 < poppingtonic> kanzure: why does Dunbar's theory, esp the part about neocortical volume, seem implausible?
16:50 < kanzure> at minimum, a correlation is not an explanation
16:51 < kanzure> absolute brain volume is reduced in children and they function just fine
16:51 < kanzure> and tiny people...
16:51 < kanzure> it also seems unlikely that just meeting lots of people is a good reason why over time more cognitive abilities were selected for
16:52 < kanzure> considering that many other animals live in very large herds and groups and do not have the human range of cognitive abilities at the moment
16:52 < kanzure> i'm sure there's other things broken there
17:32 < fenn> "this hypothesis can explain why humans enjoy wasting most of their intellectual capabilities for totally useless purposes"
17:34 < fenn> it's not that the brain is particularly vulnerable to infections, but that the immune system and brain are both huge metabolic energy users, and in the evolutionary environment (or even much of the modern world) the body can't support both brain development and fending off infection at the same time
17:34 < fenn> this is probably why the flynn effect exists
17:36 < andytoshi> fenn: what are you quating?
17:37 < andytoshi> quoting
17:37 -!- Urchin[emacs] [~user@unaffiliated/urchin] has joined ##hplusroadmap
17:51 < kanzure> immunity signaling seems insufficiently strong for a reason for the particular changes in human cognition
18:08 < kanzure> i wonder if there's parental selection for intelligence in potential mates for offspring
18:08 < kanzure> and maybe parents were in charge of accepting bribes
18:09 < kanzure> intelligence really isn't that high on the list of priorities compared to say sexual attributes that might more realistically confer advantages to raising resource-hungry brain development
18:09 < kanzure> *list of priorities [in mate selection]
18:19 < fenn> quoting "the rise of non-adaptive intelligence in humans under pathogen pressure"
18:20 < fenn> kanzure said "maybe cognitive abilities are selected by that sort of group-to-group genetic trade over time"
18:20 < fenn> not sure what "genetic trade" means but the ashkenazi jews showed strong selection for mathematical abilities during the middle ages because of the taboo/exile status as money changers
18:21 < kanzure> genetic trade is referring to "bride price" practices
18:21 < kanzure> of buying females for mating from other groups/tribes/clans/whatever
18:21 < kanzure> buying would usually involve "money", jewels, rocks, stones, livestock, surplus food, tools
18:22 < kanzure> it seems to be called marriage/dowry now, but i don't know what history people call it
18:22 < fenn> half the time dowry goes in the opposite direction i expect it to
18:23 < kanzure> my "trade" thing isn't a fully formed idea though
18:24 < kanzure> somehow being clever means you get more resources and stuff to collect... and then you buy whatever you want.. but how does this select for cognitive ability at all?
18:24 < kanzure> later in life you're selecting suitors who bring you good valuables?
18:24 < fenn> money is a fitness indicator
18:24 < fenn> or rather, symbols of money like fancy clothes and gold plated cars
18:24 < kanzure> right... but i'm not ready to say it's indicative of cognitive ability...
18:25 < kanzure> maybe it's indicative of cognitive ability 60,000 years ago
18:25 < fenn> that's all that matters
18:25 < kanzure> would it be?
18:25 < fenn> evolution is pretty slow most of the time
18:25 < kanzure> the past 100k years did some pretty fast stuff
18:25 < kanzure> so it had to be some sort of socially-mediated selection
18:26 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has joined ##hplusroadmap
18:26 < fenn> why did it have to be social
18:26 < fenn> how many stories are there about young brides and princes
18:26 < fenn> prince = guy with money
18:26 -!- AdrianG [~User@unaffiliated/amphetamine] has joined ##hplusroadmap
18:27 < kanzure> well usually you have populations that are pretty bad until there's some mutation that happens to help
18:27 < kanzure> with brains you don't need to wait around that long
18:28 < kanzure> and if brains are getting better at mate selection or something, then i would call that social
18:28 < kanzure> okay, maybe it should be cognitive selection
18:29 < kanzure> except i would expect to see evidence of mate selection based on cognitive abilities if that was the case, and there's only very obscure examples of that in modern society
18:29 < fenn> you mean cultural mores cause people to select mates with money?
18:29 < kanzure> no
18:29 < kanzure> oh, so what i meant about that money thing
18:29 < kanzure> was specifically the cultural practice where the prince buys the daughter from the other family
18:29 < kanzure> code of hammurabi says a lot about this practice
18:29 < kanzure> sounds like it was very widespread and universal
18:30 < fenn> in modern india the bride pays the prince's family for the privilege of marrying him (i think)
18:30 < kanzure> errr... processing.
18:30 < fenn> i always get confused about this
18:30 -!- gene_hacker [~chatzilla@c-50-137-46-240.hsd1.or.comcast.net] has quit [Ping timeout: 250 seconds]
18:31 < fenn> "A dowry is the transfer of parental property to a daughter at her marriage"
18:32 < kanzure> right.. i was arguing about this with gwern
18:32 < kanzure> that exact sentence from wikipedia
18:32 < kanzure> "bride price" makes much more sense to me
18:32 < fenn> gah this article rapidly devolves to academic "so and so believes..."
18:33 < fenn> "bride price" makes more sense but it doesn't relflect reality
18:34 -!- juri__ [~juri@vpn166.sdf.org] has joined ##hplusroadmap
18:35 < fenn> oh boy "6 percent of North American indigenous cultures practised reciprocal exchange, involving the giving of gifts between both the bride and groom's families."
18:36 < fenn> "property brought to the marriage by the bride is called a dowry. Property made over to the bride's family at the time of the wedding is a bride price. This property does not pass to the bride herself."
18:36 < heath> .title https://github.com/np1/pafy
18:36 < yoleaux> np1/pafy · GitHub
18:36 < heath> Python library to download YouTube content and retrieve metadata
18:37 < fenn> so i think bride price and dowry are opposites
18:37 < kanzure> i vote to just get rid of dowry and start over
18:37 < kanzure> also the 6% thing is a little strange- is that modern-day only, or what
18:37 < kanzure> i mean, everyone still buys diamond rings
18:37 < kanzure> presumably diamond ring buying is an extension of past behavior
18:38 < fenn> indigenous cultures = Native Americans
18:38 < kanzure> sure
18:38 < fenn> their culture has been so thoroughly fucked over it's hard to say anything meaningful about modern day practices
18:39 < kanzure> maybe the practice of bride price just got corrupted over time
18:39 < kanzure> and turned into other weird things that nobody knows why they bother with
18:39 < fenn> because it's more useful to have money at the start of a marriage than when your parents die
18:40 < kanzure> "here's a sheep if you let your daughter marry me"
18:40 < kanzure> this is bride price, right?
18:40 < fenn> yes
18:40 < fenn> it gets weird because the husband usually controls the finances of the new couple; so any gifts to "the bride" actually end up going to the husband
18:41 < kanzure> right..
18:41 < kanzure> sorta makes sense though that you would want your immediate descendants to have similar resources as you have
18:42 < kanzure> none of this seems like selecting for cognitive abilities would be a side effect, in any obvious sort of way
18:42 < kanzure> hehe maybe for ability to understand dowry :)
18:43 < fenn> it's only difficult for you to understand because it's not part of your culture
18:43 < kanzure> i think wikipedia is just poorly written here
18:43 < kanzure> "parental property" come on
18:43 < fenn> eh?
18:44 < kanzure> "parent's property" if that's what's meant
18:44 < fenn> ok here's the one that fucks everything up: "[in India] Dowry is a payment of cash or gifts from the bride's family to the bridegroom's family upon marriage."
18:45 < kanzure> so another possibility is that there was no single selective mechanism that most strongly influenced the evolution of cognitive abilities in humans
18:45 < kanzure> and instead it was 400 competing things
18:46 < kanzure> but, there were always competing reasons that moar cognitive abilities would have been useful
18:46 < kanzure> even for other species
18:46 -!- juri__ [~juri@vpn166.sdf.org] has quit [Ping timeout: 245 seconds]
18:47 < kanzure> so that is why i think there is probably at least one rather plausible, strong mechanism that must have been in effect
18:51 < fenn> this is why i tend to go with "aquatic diet enables big brains"
18:52 < kanzure> still not convinced about size of brain organs
18:52 < fenn> oh jeez really
18:53 < fenn> birds and bees aside, intelligence and brain size are totally correlated
18:53 < kanzure> i am convinced that the size of brains has changed
18:53 < kanzure> i'm also willing to go with correlation, sure
18:53 < kanzure> but i am more concerned about cognitive abilities
18:54 < fenn> given that big brains are metabolically (and catabolically) expensive, if it weren't important i'm 100% certain evolution would have pruned it away by now
18:54 < kanzure> even non-particularly-cognitive-able-as-human brains are kept around
18:55 < fenn> only stuff that's not neocortex
18:55 < fenn> fine, "aquatic diet enables big neocortex"
18:55 < fenn> because a bigger brainstem confers no advantage
18:56 < kanzure> "bigger's better because texas"
18:56 < kanzure> "nuber of neurons"
18:56 < kanzure> *number
18:56 < fenn> yeah, more neurons, more memory, more chance of recognizing familiar situations
18:56 < kanzure> why are you fixated on number of neurons instead of any of the other differences in brain biology to say chimpanzees
18:56 < kanzure> such as uh, the one about myelination
18:57 < fenn> because i don't know about the other differences really
18:57 < fenn> neuroscience is really hard to navigate
18:57 < fenn> also things with bigger brains seem to be more intelligent
18:58 < fenn> "why are you so fixated on the one thing that seems to correlate with what we're interested in!?"
18:58 < kanzure> hmm i thought i had a link to a paper with a table of neuron and brain differences between human and chimpanzees
18:59 < fenn> http://fennetic.net/irc/human_chimpanzee_brain_differences.png
19:00 < kanzure> hey that's the one
19:00 < fenn> "A. a. 3 times more neurons" right at the top of the list
19:02 < fenn> so, hard drives are pretty good at storing data but they have terrible latency and bandwidth vs storage size
19:03 < fenn> if you think of the neocortex as an associative memory, intelligence is the ability to remember similar situations
19:03 < kanzure> "Prominent brain shrinkage at old age"
19:06 < fenn> that's overstated
19:07 < fenn> a) chimpanzees don't live to old age because of poor quality of life, b) the shrinkage is gradual over decades
19:08 < fenn> standards of living, whatever
19:08 < kanzure> "A high density of X-linked genes for general cognitive ability: a run-away process shaping human evolution?" https://www.genetikum.de/images/PDF/Diverses/TIG-High-density-on-X-linked-genes.pdf
19:08 < kanzure> not specifically related to this question
19:09 < fenn> do educated people usually know about x-inactivation? somehow i never learned about this until recently
19:13  * fenn senses a conspiracy of feminist reptilians from sirius A
19:15 < kanzure> nope wasn't aware of that
19:15 < kanzure> seems important to me
19:21 < kanzure> i wonder how many generations of selective breeding would be required to get humans down to lower brain volumes
19:21 < kanzure> that might help the problem of mind uploading
19:21 < kanzure> or uh brain scanning i mean
19:21 < kanzure> when there's less brain to scan
19:24 -!- juri__ [~juri@vpn166.sdf.org] has joined ##hplusroadmap
19:31 < fenn> "obviously only one species so far has general intelligence" i think this is wrong
19:32 < fenn> the reason humans are so much more effective is we have intelligence, language, and hands
19:32 < fenn> any one of which is not as useful alone
19:33 -!- AdrianG [~User@unaffiliated/amphetamine] has left ##hplusroadmap []
19:34 < kanzure> if humans are not the only species with general intelligence then only the conserved brain properties matter
19:36 < fenn> this combination allows you to teach someone else how to build something
19:36 < kanzure> you don't need language for that
19:38 < Urchin[emacs]> kanzure: developing better brain scanning would probably be faster
19:38 < kanzure> there's just so much of it though
19:39 < fenn> somehow the use of language allows preservation of knowledge across longer spans of time
19:40 -!- mode/##hplusroadmap [+o kanzure] by ChanServ
19:40 -!- mode/##hplusroadmap [-q adriang!*@*] by kanzure
19:40 -!- mode/##hplusroadmap [-q AdrianG!*@*] by kanzure
19:40 -!- mode/##hplusroadmap [-o kanzure] by kanzure
19:41 < fenn> why was he shushed?
19:41 < kanzure> all sorts of reasons
19:41 < kanzure> 18:23 < AdrianG> superkuh: r u aspergian
19:41 < kanzure> 18:25 <@kanzure> AdrianG: can't you just leave
19:41 < kanzure> 18:25 < AdrianG> no :<
19:41 < kanzure> 18:25 < AdrianG> im addicted to this chat.
19:42 < kanzure> 18:25 < AdrianG> why do you hate me.
19:43 < fenn> language allows the incorporation of new tricks into the noome or neo-genome or whatever we decide to call it
19:43 < kanzure> cognitome :(
19:43 < fenn> yuck
19:43 < fenn> .wik noosphere
19:43 < yoleaux> "The noosphere (/ˈnoʊ.əsfɪər/; sometimes noösphere) is the sphere of human thought. The word derives from the Greek νοῦς (nous "mind") and σφαῖρα (sphaira "sphere"), in lexical analogy to "atmosphere" and "biosphere". It was introduced by Pierre Teilhard de Chardin in 1922 in his Cosmogenesis." — http://en.wikipedia.org/wiki/Noosphere
19:44 < fenn> cogni-* implies some kind of active processing occurring
19:44 < fenn> noo-* is just there in storage
19:45 < kanzure> so one idea i was thinking about was applying directed evolution to artificial general intelligence projects
19:45 < fenn> "In colloquial British English, nous also denotes "good sense", which is close to one everyday meaning it had in Ancient Greece."
19:45 < kanzure> hell, 50% of the past million years was spent sleeping
19:45 < kanzure> that definitely doesn't need to be replicated
19:45 < fenn> au contraire
19:45 < fenn> even flies sleep, there's a good reason for it
19:45 < kanzure> says the supersleeper
19:46 < kanzure> even flies have brains
19:46 < Urchin[emacs]> de Chardin - the guy who thought that humans are a part of God's reproductive system, set to create a new universe
19:46 < fenn> i'd rather sleep less, sure, but that doesn't mean i can wish it away
19:47 < fenn> Urchin[emacs]: yeah this was part of my childhood
19:47 < kanzure> well, i don't know how to implement sleep in code, so let's skip that
19:47 < fenn> garbage collection
19:47 < kanzure> there's evidence for other non-garbage-collection-related activities though
19:48 < Urchin[emacs]> fenn: mine too, I just haven't heard of it in a long while
19:48 < fenn> dreams have something to do with transfering memories to long term memory
19:48 < kanzure> also, idea-trade and gene-trade and reproduction get fucked up in the ai domain
19:48 < kanzure> i don't see a good reason for a difference between "idea-trade" and "gene-trade"
19:50 < fenn> one is related to prediction and the other is related to function
19:50 < fenn> you could trade algorithms without having a use in mind for them
19:50 < fenn> but a gene is directed toward some specific goal
19:51 < fenn> or function
19:51 < fenn> s/goal/function/
19:51 < kanzure> even humans execute "ideas" they come across
19:51 < kanzure> they just inhibit any behavior that would result from that idea or something, unless they evaluated it to be salient to their interests etc
19:51 < fenn> when you think about evolving AI architectures you have to separate the data structure from the data, or all hell breaks loose
19:51 < kanzure> why?
19:52 < fenn> because self modifying code is ridiculous
19:52 < fenn> it just is
19:52 < fenn> it's unpredictable
19:52 < kanzure> so?
19:52 < fenn> that's why self-modifying code has been outlawed (by OS authors) on most computing platforms
19:53 < fenn> https://en.wikipedia.org/wiki/Self-modifying_code#Self-referential_machine_learning_systems
19:54 < kanzure> "Traditional machine learning systems have a fixed, pre-programmed learning algorithm to adjust their parameters. However, since the 1980s Jürgen Schmidhuber has published several self-modifying systems with the ability to change their own learning algorithm. They avoid the danger of catastrophic self-rewrites by making sure that self-modifications will survive only if they are useful according to a user-given fitness, error or reward ...
19:54 < kanzure> ... function."
19:54 < kanzure> that all sounds totally trivial and boring
19:54 < kanzure> so maybe you choose to isolate some components, maybe not, maybe there's some breakout exploit. meh.
19:55 < fenn> that just changes the algorithm, not the data structure itself
19:56 < kanzure> most possible changes to any code will make it crash, and that's fine
19:56 -!- CharlieNobody [~CharlieNo@97-85-246-63.static.stls.mo.charter.com] has left ##hplusroadmap ["Leaving"]
19:56 < kanzure> this is not some code permutation engine
19:56 < fenn> it's bad for the individual agent to have a high probability of accidentally committing suicide
19:57 < kanzure> i don't think i'm suggesting otherwise
19:57 < fenn> there's some kind of information theory problem here about simulating a system in perfect detail using the system itself but without increasing the size of the system
19:58 < fenn> map and territory recursion
19:58 < kanzure> whatever
19:58 < kanzure> i am not concerned about self-modifying code
19:59 < fenn> if you can't simulate the effect of your modification it's like jumping off a cliff and hoping you survive
19:59 < catern> code is data yo
19:59 < catern> data is code
19:59  * catern sips his lisp
19:59 < fenn> catern: your appearance was not unexpected
19:59 < kanzure> catern: https://www.youtube.com/watch?v=rdj6deraQ6k&index=84&list=PL85F050BEFA28E044
19:59 -!- yashgaroth [~ffffff@2606:6000:cb85:6a00:4108:5bf7:9f03:e1c0] has joined ##hplusroadmap
20:00 < fenn> .title
20:00 < yoleaux> Abelson Stole The Precious Course - MIT Version - YouTube
20:00 < catern> a classic
20:01 < fenn> anyway forking on modification is totally fine and that's what biology does
20:01 < kanzure> meh
20:01 < fenn> cancer, viruses notwithstanding
20:02 < kanzure> if you want to emulate biology then just emulate biology..
20:02 < kanzure> which, i do, but that's a different topic i think
20:02 < fenn> you wanted to know the difference between gene trade and idea trade...
20:03 < kanzure> i don't think there has to be a difference
20:03 < fenn> an idea can become a gene (data can become code) but you'd be foolish to stuff it into your running self
20:03 < kanzure> all genes can be received through ideas
20:03 < kanzure> by which i mean uh data
20:03 < kanzure> sensory input
20:06 < kanzure> i was hoping that by figuring out which particular selective mechanisms were in effect influencing cognitive abilities in humans that the mechanism could be repeated for critterdrug
20:06 < kanzure> my bad i mean telepathic-critterdrug
20:07 < fenn> most alife stuff is far too simplistic imho
20:07 < fenn> "we'll just use a 32 bit genome" derp
20:08 < fenn> there was one thing with simulated sensory input based on a simulated environment that looked vaguely interesting
20:10 < fenn> anything like that needs a lot of time in the real world to be useful though
20:10 < kanzure> i wonder how far back the conserved brain stuff goes. hrm.
20:10 < fenn> otherwise it will just optimize for the quirks of your particular simulation
20:11 < kanzure> mice have most of the right parts
20:11 < fenn> vertebrates
20:11 < kanzure> "According to research the cerebrum first developed about 200 million years ago, having a unique highly convoluted surface that is called the neocortex"
20:11 < fenn> .wik coelacanth
20:11 < yoleaux> "The coelacanths (i/ˈsiːləkænθ/ SEE-lə-kanth) constitute a now rare order of fish that includes two extant species in the genus Latimeria: the West Indian Ocean coelacanth (Latimeria chalumnae) and the Indonesian coelacanth (Latimeria menadoensis)." — http://en.wikipedia.org/wiki/Coelacanth
20:13 < kanzure> goldfish might work, what's in their brain
20:14 < kanzure> http://en.wikipedia.org/wiki/Evolution_of_the_brain#Genetic_factors_contributing_to_modern_evolution "Lahn's earlier studies displayed that Microcephalin experienced rapid evolution along the primate lineage which eventually led to the emergence of Homo sapiens. After the emergence of humans, Microcephalin seems to have shown a slower evolution rate. On the contrary, ASPM showed its most rapid evolution in the later years of human ...
20:14 < kanzure> ... evolution once the divergence between chimpanzees and humans had already occurred.[9]"
20:15 < kanzure> "Over the past 20,000 years, the average volume of the human male brain has decreased from 1,500 cubic centimeters to 1,350 cc, losing a chunk the size of a tennis ball. The female brain has shrunk by about the same proportion. "
20:16 < kanzure> from http://discovermagazine.com/2010/sep/25-modern-humans-smart-why-brain-shrinking (i haven't evaluated this please don't shoot me)
20:16 < fenn> is that because people are living longer? (brain shrinks with age)
20:16 < fenn> i thought it was the opposite anyway
20:17 < fenn> 20k years coincides with agriculture
20:18 < fenn> lack of aquatic diet would be a major strain on a system that evolved with plenty of omega-3 available
20:19 < Urchin[emacs]> loss of senses?
20:19 < kanzure> "Another popular theory attributes the decrease to the advent of agriculture, which, paradoxically, had the initial effect of worsening nutrition. Quite simply, the first farmers were not very successful at eking out a living from the land, and their grain-heavy diet was deficient in protein and vitamins—critical for fueling growth of the body and brain. In response to chronic malnutrition, our body and brain might have shrunk. Many ...
20:19 < kanzure> ... anthropologists are skeptical of that explanation, however. The reason: The agricultural revolution did not arrive in Australia or southern Africa until almost contemporary times, yet brain size has declined since the Stone Age in those places, too."
20:20 < fenn> "protein and vitamins" duhhhh no
20:20 < kanzure> "The observation led the researchers to a radical conclusion: As complex societies emerged, the brain became smaller because people did not have to be as smart to stay alive. As Geary explains, individuals who would not have been able to survive by their wits alone could scrape by with the help of others—supported, as it were, by the first social safety nets."
20:21 < fenn> the idiocracy hypothesis
20:21 -!- HEx1 [~HEx@hexwab.plus.com] has quit [Ping timeout: 250 seconds]
20:21 < kanzure> "Hawks spent last summer measuring skulls of Europeans dating from the Bronze Age, 4,000 years ago, to medieval times. Over that period the land became even more densely packed with people and, just as the Missouri team’s model predicts, the brain shrank more quickly than did overall body size, causing EQ values to fall. In short, Hawks documented the same trend as Geary and Bailey did in their older sample of fossils; in fact, the ...
20:21 < kanzure> ... pattern he detected is even more pronounced. “Since the Bronze Age, the brain shrank a lot more than you would expect based on the decrease in body size,” Hawks reports. “For a brain as small as that found in the average European male today, the body would have to shrink to the size of a pygmy” to maintain proportional scaling."
20:24 < fenn> "others believe that the reduction in brain size is proof that we have tamed ourselves, just as we domesticated sheep"
20:24 < fenn> wake up sheeple!
20:24 < kanzure> "“When you select against aggression, you get some surprising traits that come along with it,” Wrangham says. “My suspicion is that the easiest way for natural selection to reduce aggressiveness is to favor those individuals whose brains develop relatively slowly in relation to their bodies.” When fully grown, such an animal does not display as much aggression because it has a more juvenile brain, which tends to be less ...
20:24 < kanzure> ... aggressive than that of an adult. “This is a very easy target for natural selection,” Wrangham argues, because it probably does not depend on numerous mutations but rather on the tweaking of one or two regulatory genes that determine the timing of a whole cascade of developmental events. For that reason, he says, “it happens consistently.” The result, he believes, is an adult possessing a suite of juvenile characteristics, ...
20:24 < kanzure> ... including a very different temperament."
20:25 < kanzure> "So what breeding effect might have sent humans down the same path? Wrangham offers a blunt response: capital punishment. “Over the last 100,000 years,” he theorizes, “language became sufficiently sophisticated that when you had some bully who was a repeat offender, people got together and said, ‘We’ve got to do something about Joe.’ And they would make a calm, deliberate decision to kill Joe or expel him from the group—the ...
20:25 < kanzure> ... functional equivalent of executing him.” Anthropological records on hunter-gatherers suggest that capital punishment has been a regular feature of our species, according to Wrangham. In two recent and well-documented studies of New Guinea groups following ancient tribal custom, the ultimate punishment appears to be meted out to at least 10 percent of the young men in each generation."
20:25 < fenn> this doesn't square with the fact that conscientious objectors have higher IQ
20:25 < kanzure> uh... okay. that's an interesting one i didn't consider.
20:26 < kanzure> killing off the bottom 10% all the time
20:26 < kanzure> geeze
20:26 < fenn> the pak protector hypothesis
20:28 < fenn> or the top 10% depending on how you look at it
20:28 < fenn> .wik harrison bergeron
20:28 < yoleaux> ""Harrison Bergeron" is a satirical and dystopian science-fiction short story written by Kurt Vonnegut and first published in October 1961. Originally published in The Magazine of Fantasy and Science Fiction, the story was republished in the author's Welcome to the Monkey House collection in 1968." — http://en.wikipedia.org/wiki/Harrison_Bergeron
20:28 -!- HEx1 [~HEx@hexwab.plus.com] has joined ##hplusroadmap
20:29 < kanzure> killing the top 10% seems like a way to not get cognitive abilities
20:29 < fenn> hence the brain shrinkage
20:29 < kanzure> "When anthropologist Richard Jantz of the University of Tennessee measured the craniums of Americans of European and African descent from colonial times up to the late 20th century, he found that brain volume was once again moving upward"
20:29 < kanzure> "Hawks, for instance, says the explanation is “mostly nutrition.” Jantz agrees but still thinks the trend has “an evolutionary component because the forces of natural selection have changed so radically in the last 200 years.” His theory: In earlier periods, when famine was more common, people with unusually large brains would have been at greater peril of starving to death because of gray matter’s prodigious energy ...
20:29 < kanzure> ... requirements. But with the unprecedented abundance of food in more recent times, those selective forces have relaxed, reducing the evolutionary cost of a large brain."
20:30 < kanzure> hmm.
20:30 < fenn> that makes a lot more sense to me than the "let's kill joe" hypothesis
20:32 < fenn> my understanding of new guinea killing is that it's a cycle of blood feuds and revenge
20:32 < fenn> you're required to kill whoever killed someone who killed someone because someone killed someone... unti nobody remembers why they are even fighting
20:33 < fenn> it's not like there is some council of elders who forces everyone to take some SAT test
20:34 < kanzure> they were talking about aggression
20:34 < kanzure> not intelligence
20:34 < fenn> and the most aggressive tend to be more successful in that society
20:34 < fenn> contrary to the hypothesis' prediction that group violence would lead to domestication
20:35 < fenn> because it's inter-group violence not intra-group violence
20:36 < kanzure> mutually-beneficial trade surely had some impact somewhere
20:36 < fenn> sure, trade and aggression are not exclusive
20:36 < kanzure> no was idle thought
20:46 < kanzure> "no more than a million lines of code" should be a universal law for all software projects
20:46 < kanzure> this may encourage more software projects though
20:46 < kanzure> which might be okay
20:46 < fenn> not necessarily a bad thing
20:47 < fenn> unix philosophy
20:47 < fenn> standardized interfaces
20:47 < kanzure> i agree that there's a much in common with operating system and kernel design
20:47 < catern> BSDs have more than a million lines of code
20:47 < kanzure> with ai things
20:48 < fenn> i have no idea about kernel design
20:48 < kanzure> memory allocation, resource utilization, management
20:48 < kanzure> isolation
20:48 < catern> because they have all their stuff in the same repo, coreutils, libc, etc.
20:48 < kanzure> i don't know, read tanenbaum you lazy bum
20:48 < kanzure> or read linux source code, that's more productive
20:48 < fenn> doesn't he say something like 'microkernels are the only thing that could possibly work"
20:49 < kanzure> honestly i forget
20:49 < kanzure> i was reading it when i was 13 which was my year of "pretend to learn everything"
20:49 < fenn> https://en.wikipedia.org/wiki/Tanenbaum%E2%80%93Torvalds_debate
20:50 < kanzure> hah that's such a unix professor photo
20:51 < fenn> linus even thought microkernels were better
20:51 < fenn> i guess with kernel modules we have something like a microkernel in practice
20:59 < fenn> "The whole [MINIX] build, kernel + user-mode drivers and all the user-mode servers (125 compilations in all) takes about 5-10 seconds."
21:00 < fenn> "It is a minimal but functional UNIX system with X, bash, pdksh, zsh, cc, gcc, perl, python, awk, emacs, vi, pine, ssh, ftp, the GNU tools and over 400 other programs."
21:01 < fenn> this was written in 2005 ish
21:03 < fenn> sometimes i forget that UNIX ran on 5MHz computers
21:04 < fenn> The PDP-7 had a ~1-2 Mhz CPU
21:05 < kragen> it wasn't until they bought a PDP-11 that they started writing Unix though
21:05 < fenn> "In 1969, Ken Thompson wrote the first UNIX system in assembly language on a PDP-7, then named Unics"
21:06  * fenn mumbles something about "no philosophy"
21:06 < kragen> I'm not sure that really counts :)
21:07 < kragen> I mean it had the same name as UNIX (modulo a spelling change) but it was different code in an incompatible language with incompatible interfaces
21:07 < fenn> it wasn't unti 1973 that they rewrote it in C
21:07 < kragen> but around 1971 they rewrote it in PDP-11 assembly
21:07 < fenn> so what
21:08 < kragen> that was still the time period when time_t incremented 60 times a second and the filesystem was nonhierarchical though
21:08 < kragen> and instead of exiting, your program had to exec sh
21:09 < kragen> because sh didn't fork to run your program
21:09 < fenn> i guess i'm just surprised that there isn't something unix-equivalent for AVR class microcontrollers
21:09 < kragen> and they didn't have stderr yet.  in fact it was a while before they had pipes
21:09 < kragen> AVR microcontrollers have a lot less RAM than the PDP-7
21:10 < kragen> less than half
21:10 < kragen> in many cases a lot less than that
21:11 < fenn> i'm not seeing stats on how much ram PDP-11 had
21:11 < kragen> the PDP-7 shipped with 4096 18-bit words
21:12 < kragen> (which should give you some clue of how far from C's worldview it was)
21:12 < kragen> but I think the main obstacle is actually I/O devices rather than RAM — for the price of a decent keyboard, you can also have a 32-bit microprocessor running at hundreds or thousands of MHz
21:13 < fenn> huh
21:13 < kragen> however if you're interested
21:13 < fenn> for the price of an arduino you can have an android tablet with touchscreen, wifi, GPS, battery..
21:13 < kragen> you can?
21:13 < kragen> isn't an Arduino like US$10?
21:14 < fenn> something like that
21:14 < fenn> er, no the arduino is like $30
21:14 < lichen> 10 if you build it yourself
21:14 < fenn> $2 if you build it yourself
21:14 < kragen> the KnightOS people are putting together a Unix-like system for Z80
21:14 < lichen> 30 dollar tablet is still rather unusual
21:14 < kragen> because TI calculators are still based on Z80s
21:14 < kanzure> an android tablet is like $9
21:15 < fenn> kanzure are you sure about that
21:15 < kanzure> you're my source
21:15 < kragen> and they have enough RAM to support writing programs in high-level languages, which Arduinos don't really
21:15 < kragen> also
21:15 < kragen> monitor and keyboard and battery
21:15 < fenn> the thing on aliexpress was a scam, not an actual viable commercial product
21:15 < kragen> they cost more than Android tablets though
21:15 < kragen> there are actual viable commercial Android tablets for US$50 though
21:15 < kragen> still
21:16 < kragen> an Arduino Nano costs about US$10
21:16 < kragen> rossum.posterous had some interesting stuff about hacking cellphone displays
21:16 < kragen> they are very cheap indeed, and much more widely available than Arduinos
21:17 < kragen> (which makes me curse their benighted lack of programmability)
21:17 < kanzure> are there any software neural networks that have been benchmarked as "better" than the real deal?
21:17 < kanzure> like some simple neural task that a little worm does or something
21:17 < fenn> better?
21:18 < kanzure> faster response times?
21:18 < fenn> more wormy?
21:18 < kragen> probably "better" in this case would be "more robust to crisis situations"
21:19 < kanzure> "does not get served its own ass by the biological version"
21:19 < kanzure> "is not a pathetic joke compared to biological implementation"
21:19 < fenn> software latency is more related to i/o cruft in the x86 peripheral design from 1980's
21:19 < fenn> but that's still on the order of a millisecond or two
21:19 < kragen> like, higher likelihood of the worm surviving when it gets swallowed by a larger animal, or it snows, or the puddle dries up
21:19 < fenn> what's a worm's reaction time?
21:19 < kragen> .g worm reaction time
21:19 < yoleaux> http://www.cis.upenn.edu/~cse331/lectures/CSE331-16.pdf
21:20 < kanzure> where are the openworm people argh
21:20 < fenn> .g dear lazyweb where are the openworm people
21:20 < yoleaux> http://www.openworm.org/get_involved.html
21:20 < kragen> false hit at upenn
21:21 < kragen> fenn: I've tried persuading the KnightOS people that maybe CP/M compilers and assemblers and things could be useful
21:21 < kragen> but they point out that the TI's memory layout is incompatible
21:22 < kragen> I think they'd probably be more convinced by a working demonstration than by some random dude bloviating about it
21:22 < kragen> they're hoping to get an interpretive language up and running soon
21:22 < kragen> "kpy"
21:22 < fenn> oh yeah people still need to buy TI-8X for their math class because it sucks and it also has graphing capability
21:23 < kragen> yes
21:23 < kragen> there was a related mp3 player called s1mp3 a few years back
21:23 < kragen> Z80 with a giant glob of DSP on the side
21:23 < fenn> i remember running a grayscale wolfenstein clone on mine
21:23 < kragen> but nobody buys MP3 players any more
21:23 < fenn> was going to build an IR transciever to chat in class but nobody else cared
21:24 < kragen> like seriously I was just in a bad neighborhood where an old man robbed a young woman in front of me
21:24 < kragen> and there was this monstrously fat guy sitting out on the sidewalk with his family with a hand-lettered sign selling sausage sandwiches
21:24 < kragen> cooked there on the sidewalk on a charcoal grill
21:25 < fenn> and, uh, what?
21:25  * fenn is not following this conversation all of a sudden
21:25 < kragen> and he was playing his music on a Samsung Android smartphone with a cracked faceplate
21:25 < kragen> plugged into his stereo
21:25 < kragen> that's what I mean by "nobody buys MP3 players any more"
21:27 < fenn> i thought you were talking about some mp3 software for the ti-8x calculators
21:27 < kragen> so there aren't that many consumer products where you have an 8-bit micro with a screen any more
21:27 < kragen> no, I was talking about the s1mp3
21:27 < kragen> maybe some feature phones
21:27 < kragen> but those are probably using cortex-m0 these days
21:28 < kragen> I dissected a DVD player last year
21:28 < kragen> turned out its main monster ASIC was an 8051
21:28 < kragen> with like 200 pins hooked up to DSP glommed onto the side of the 8051
21:28 < kragen> and a megabyte or two of program-counter address space
21:28 < kragen> but 8051 instruction set
21:29 < kragen> but it was an old one
21:29 < fenn> not at all surprised
21:29 < kragen> to me that's a platform it would be interesting to have a reasonable OS for
21:29 < fenn> why is that interesting
21:30 < fenn> you might as well write an OS for your hard drive's controller
21:30 < kragen> yeah, hard drive controllers are also pretty interesting
21:30 < kragen> but they don't have video output
21:31 < fenn> false economy detected
21:31 < kragen> it's not about economy
21:31 < kragen> it's about autonomy
21:31 < fenn> you want autonomy, don't buy a DVD player!
21:31 < fenn> wtf
21:31 < kragen> what am I going to generate my Bitcoin keys on?
21:31 < fenn> an abacus
21:32 < kragen> clearly
21:32 < fenn> you can write an os for an arbitrarily obscure platform
21:32 < kragen> you can waste an arbitrarily large amount of time doing it too
21:34 < fenn> unix for AVR is interesting because you can do "internet of things" program changes without having to recompile and flash new binary code to the device
21:34 < kragen> I note that even Contiki doesn't support the 8051
21:35 < kragen> well, if that's what you want, use Contoki
21:35 < kragen> Contiki
21:35 < kragen> it does support the AVR
21:35 < kragen> as well as every other common "internet of things" CPU
21:35 < kragen> plus the 6502
21:36 < fenn> also it would be nice to have a DIY processor capable of running modern software, or at least software designed with modern concepts
21:36 < kragen> what do you mean?
21:36 < fenn> where DIY = silicon fab in a garage
21:37 < kragen> you could probably do a 6502 or a MuP21 in your garage
21:37 < fenn> where do you get 30kB of RAM from
21:37 < kragen> (actually TinyOS also does a bunch of program upgrade things, which might be an option too)
21:37 < kragen> yeah, that's probably going to be a problem
21:38 < fenn> i dont really mean bootloader tricks
21:38 < kragen> I wonder if you could use focused ion beam etching instead of masks
21:38 < fenn> with FIB you can just etch a zillion transistors and have a big modern cpu
21:38 < fenn> i'm talking about lithographic masks in the micron range
21:38 < kragen> can you do FIB in your garage?
21:38 < fenn> not unless you have a FIB to begin with
21:39 < fenn> but yes
21:39 < kragen> what's hard about building one?
21:39 < kragen> (sort of like the injection-molding to FDM shift: if you only need one of something, making masks/molds for it may not be economical)
21:39 < fenn> i don't know; nothing seems particularly difficult about it
21:39 < kragen> I have this very naïve idea about FIB
21:39 < kragen> which is that it's basically a scanning electron microscope with nuclei instead of electrons
21:40 < kragen> on the other hand, Jeri didn't build her SEM herself
21:40 < kragen> she bought it
21:40 < fenn> it takes a certain skillset to build a SEM that doesn't seem to exist in the sort of people that would want to build a computer chip
21:41 < fenn> also there's no reason to build when you could just buy one
21:41  * fenn shrugs
21:42 < fenn> there's also digital mask projection with DLP
21:42 < fenn> you don't need to actually print a mask to do lithography
21:42 < kragen> that's an interesting idea
21:42 < kragen> are there people doing that at micron scale?
21:42 < fenn> yes
21:43 < fenn> the whole setup is a lot cheaper and more compact than FIB
21:43 < kragen> neat, who? what are they making?
21:43 < fenn> test patterns
21:43 < kragen> yeah, it seems like it would be (if it works at all)
21:43 < kragen> test patterns? so they don't have it working very well yet?
21:43 < fenn> there's just no huge push for it in the first place
21:45 < fenn> how the fuck does kanzure know what's a category and what's just a random file that happens to be a directory
21:45 < fenn> kragen: some papers in here http://diyhpl.us/~bryan/papers2/optics/photolithography/
21:51 < fenn> the contiki GUI is ridiculous
21:52 < fenn> any serious interface could be realized with javascript and json
21:52 < dingo> or curses :)
21:52 < fenn> not really a GUI but fine
21:53 < dingo> http://jeffquast.com/blessed-weather.png
21:53 < dingo> who says that isn't graphical :)
21:53 < fenn> reimplementing font kerning just seems like a waste of resources on such a resource constrained system (the intended use case)
21:54 < fenn> dingo: why not just use existing icons
21:59 < fenn> none of these microcontroller things say anything about cryptography
21:59 < fenn> that would seem to be important for something controlling physical hardware and connected to a network
22:00 < fenn> or even just a thing connected to a network (a vector for further attacks on something else)
22:04 < fenn> protocol buffers would probably work well for interfaces to humans and machines
22:04 < fenn> i wonder how much space a .js file to read a protocol buffer stream would take up
22:12 < dingo> reminds me of http://phrack.org/issues/60/14.html#article
22:15 < fenn> i am sitting at a traffic light; six days later i have a working implementation to make the light change color
22:15 < fenn> mission accomplished?
22:15 < fenn> bytebuffer.js is <5kB with comments and full length variable names
22:16 < fenn> this seems reasonable to store on in microcontroller flash
22:17 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-qltimvzfifgmqtbh] has quit [Quit: Connection closed for inactivity]
22:17 < fenn> also you can probably do some crude compression
22:18 < fenn> with gzip it's 1.5kB
22:20 < fenn> this can't be right
22:21 -!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has quit [Quit: Lost terminal]
22:26 < fenn> the point of all the being that it's faster to spew out binary data than json
22:27 < fenn> endpoint of said data being a human looking at a webpage displaying data from an embedded sensor
22:31 < nmz787> fenn: some microcontrollers have crypto engines in them
22:34 < nmz787> http://www.nxp.com/documents/user_manual/UM10503.pdf
22:37 < nmz787> see from pg 51 and pg 93
22:37 < fenn> BeRTOS has an implementation of MD2  :\
22:37 < nmz787> AES 128-bit there
22:38 < fenn> i don't doubt that there exists unnumerable hardware crypto engines
22:38 < fenn> but i'd want something simple and auditable
22:38 < fenn> did you seriously just link to a 1400 page datasheet
22:39 < nmz787> yeah, it's pretty much what you have to refer to for turning features on
22:40 < nmz787> you could fuzz it
22:40 < nmz787> I can't actually tell if they're selling the version with that crypto engine
22:40 < fenn> fuzzing doesn't tell you if there are backdoors
22:40 < nmz787> you could decap and slice and view
22:41 < nmz787> but that would take equip
22:41 < fenn> how would you process the images, assuming you could get them
22:41 -!- yashgaroth [~ffffff@2606:6000:cb85:6a00:4108:5bf7:9f03:e1c0] has quit [Quit: Leaving]
22:41 < nmz787> you'd need elemental map data or maybe crystallographic orientation
22:42 < nmz787> that you could determine the doping with
22:42 < fenn> you can guess that
22:42 < fenn> but you'd still need some kind of OCR to get a netlist
22:42 < nmz787> yeah so then you'd need to create some rtl or something, vhdl, from that
22:42 < fenn> right
22:42 < fenn> easier to just do it in software
22:42 < nmz787> yeah with good contrast you might automate it
22:43 < nmz787> or the 'right' cv algo
22:43 < fenn> i mean, easier to do the whole cryptosystem in software so you can audit the code instead of messing around with acid and chips and microscopes and computer vision programs
22:43 < nmz787> they do everything but the actually layer-to-layer mapping automatically/assisted today
22:44 < nmz787> yeah but since it is closed source you need to back-annotate from the physical chip
22:44 < fenn> what is closed source
22:45 < nmz787> e.g. (that linked processor)
22:45 < nmz787> but anything pretty much
22:45 < nmz787> closed or licensed away
22:47 < fenn> i think that AES-128 stuff is just to protect the object code from peeking
22:48 < fenn> or 'trusted computing' stuff
22:48 < fenn> since it's only talking about valid boot images
22:49 < fenn> oo it passes the NIST random number generator standard (EXCELLENT!!!)
22:49 < nmz787> seemed to say you could use it via an API and a software key, or a randomly generated key
22:49 < nmz787> in addition to the hardware key for disk transactions
22:50 < nmz787> but yeah I couldn't tell if it was just for boot, or for all disk transactions
22:50 < nmz787> yay I just got a GIMP Python plugin to save layers as BMPs!
22:51 < nmz787> now I can convert them to FIB patters
22:51 < fenn> why did you settle on gimp instead of inkscape or just raw svg?
22:53 < nmz787> GIMP is easier to use than inkscape for me... I remember having problems defining hard edges with it when exporting to SVG (I think it was dithering, maybe?) (or maybe that was on export to BMP from SVG)
22:53 < nmz787> but GIMP layers seem pretty reasonable, they can remain paths or whatever
22:53 < nmz787> and I believe there is a way to generate them via python programmatically
22:54 < nmz787> the FIB pattern is essentially a sparse BMP
22:54 < fenn> oh. svg uses a center-aligned stroke so a 1px line ends up being 2px wide at 50% intensity
22:54 < fenn> you have to shift everything by half a pixel
22:54 < nmz787> so I will just look through the 1024x1024 image and add pixels that are >0
22:55 < nmz787> err, 4096x4096
22:55 < nmz787> with 1024x1024 pix avail to write
22:56 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has quit [Ping timeout: 245 seconds]
22:56 < fenn> hrm i guess NSA-backdoored encryption is better than no encryption
22:57 < fenn> until everyone else figures out the backdoor
22:58  * fenn sulks and watches cartoons instead
22:58 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 240 seconds]
23:01 -!- Qfwfq [~WashIrvin@unaffiliated/washirving] has joined ##hplusroadmap
23:01 -!- JayDugger [~jwdugger@pool-173-57-55-138.dllstx.fios.verizon.net] has joined ##hplusroadmap
23:11 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
23:15 -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Max SendQ exceeded]
23:17 -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap
23:25 < fenn> .title http://www.meetearl.com/
23:26 < yoleaux> Earl - Backcountry Survival Tablet
23:26 < fenn> it's an e-ink solar-powered waterproof gps walkie-talkie thing
23:45 -!- juri__ [~juri@vpn166.sdf.org] has quit [Ping timeout: 250 seconds]
23:47 -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has quit [Remote host closed the connection]