2008-07-24.log

--- Day changed Thu Jul 24 2008
ybitoh yeah, back 00:17
* ybit makes a note not to annoy others with his status on IRC00:17
kanzurehm?00:20
ybitit´s annoying for others and for myself 00:36
-!- kanzure [i=bryan@66.112.232.74] has quit ["Leaving."]00:39
-!- kanzure [i=bryan@66.112.232.74] has joined #hplusroadmap00:46
-!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has joined #hplusroadmap02:26
kanzurehttp://heybryan.org/egan.html quantification of the problem03:03
-!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has quit [Read error: 110 (Connection timed out)]03:43
-!- Splicer [n=p@h124n1c1o261.bredband.skanova.com] has quit []05:07
-!- jm [n=jm@p57B9FD0C.dip.t-dialin.net] has joined #hplusroadmap06:21
-!- Democritus [n=ard@c-68-49-79-239.hsd1.md.comcast.net] has quit [Read error: 110 (Connection timed out)]08:56
kanzurehttp://p2pfoundation.net/Category:Design09:04
kanzurefrom my inbox today, someone from that site contacted me09:04
kanzureyep, I'm impressed09:20
ybithttp://www.lumenlab.com/12:36
ybitlink courtesy of Phreedom12:36
ybithttp://www.lumenlab.com/robotics/ - all the parts only cost around $50012:37
fennhype13:11
ybitah, less than $200013:16
ybitfenn: why do you say that?13:16
fennmore importantly, it's made up of rods like a printer13:16
fenntry machining some aluminum with a printer carriage and see what happens13:17
fennit might work on wood, but slowly13:18
fennalso i dont see any electronics to run the servos13:21
fennits like they just threw a bunch of stuff in a box without seeing if it would actually work13:22
fennhmm. diaspora is starting to read like a lecture on 5-dimensional physics17:42
kanzureI first encountered it when I was 1217:45
kanzuredidn't like it much17:45
kanzurewhile the idea is "of course" and nice that he tried, eh17:45
fenn"if there are known effects, doesn't this render the whole damn map pointless? The map is not the territory, but if you "know" the territory, then why the hell have the map?"17:50
fennthey didnt know the results of all possible combinations (points on the map)17:50
fenni got the impression that some parts of the map were more thoroughly sampled than others17:51
kanzureso perhaps it's to the point where it's "roughly mapped"17:53
kanzureI'm pretty sure that we somewhat have Egan's map with Google Mouse, but there needs to be some way to record something about the brains that the SNP data produces :-)17:53
kanzureif this ends up being just some natural language the brain supposedly inputs into the system, so be it :-( (mostly useless)17:53
fennhonestly the google mouse looks like a bunch of colored squiggles to me17:54
kanzurethe idea or implementation17:54
kanzurebasically Google Mouse can also be the ABA Brain Explorer (my implementation) so it's 3D and showing cortices and so on17:54
kanzuredoes your comment still apply in that context ?17:54
fennit doesn't show the relationship between physical structure and behavior/functionality17:55
kanzurecorrect17:56
fennthen what does it show?17:56
kanzureit was going to be a 'scaffold' where I could somehow constrain my research17:56
kanzurei.e., tag some possible avenues for research17:56
kanzureif there's a statistical correlation between some genes and some perceived functionality, then that would be relevant to search for, no?17:56
fennright17:56
kanzurebut there's still no way to do this functionality mapping17:57
kanzureargh17:57
kanzureworst case scenario it's just random measurements from lab testing17:57
fennbut a realtionship between some gene and functionality doesn't have any bearing on your map (haha)17:57
kanzure"ok, does the brain respond to theta waves ... does it respond to amphetamine ... make the following measurements ..."17:57
kanzurethat's true17:57
kanzureepistasis and such17:57
kanzureI was hoping that could be sorted out by the massive statistical aspect of the thing ... if you have a few SNP datasets and measurements from each of the brains, that could get useful, no?17:58
* kanzure doesn't know if 1k is enough to sort out that many damn variables17:58
kanzureand /me admits how nasty a hack this is17:58
kanzureI was going to email Greg and yell at him about it17:59
kanzurebut I don't know if he'll have any ideas17:59
fennyou could correlate genotype with phenotype (the google map) and use that to infer a relationship between gene and physical shape, which would link to gene/function studies (i think i need to rephrase this sentence)17:59
kanzurewell I thought the issue with that would be multiply confluencing genes17:59
kanzurepolymerase without helicase just isn't the same17:59
fennbut it's not n! combinations either18:00
kanzureheh18:00
kanzureoh, right18:00
kanzurethe lab above me was doing giant correlational studies with yeast18:00
kanzurethe same could be done for brain circuits18:00
kanzurei.e., Marcotte was isolating the genetic regulatory networks18:00
kanzurehttp://reactome.org/18:00
kanzureso he had the entire yeast knockout library frozen in a few fridges upstairs18:01
kanzureso how does one describe phenotype anyway?18:02
fennthat's some interesting javascript18:02
kanzureI remember in my first biology class ever, the first week, I was assigned the task of writing down descriptions of little critters in jars18:02
kanzureprobably filled with formaldehyde or something18:02
kanzureformalin, whatever18:02
kanzureanyway, being the way that I am,18:02
kanzureI neglected to write down "arm, leg" and such because I just thought that it wasn't all that scientific approach to phenotypic descriptions18:02
kanzureso I went hardcore with "projection from the 5th millimeter of the upper extremity"18:03
kanzureof course I failed miserably, but the idea applies here18:03
fennwell, insects dont really have arms18:03
kanzure:)18:04
kanzureexactly18:04
fennjust a bunch of differently shaped plates18:04
kanzurethese things aren't really 'legs' even though they have correlated functions18:04
fennarm or leg is really a functional description18:04
kanzurewhich seems to be more .. ethology, perhaps18:04
fennreactome.org is really impressive18:06
kanzureI agree.18:06
kanzurehaven't gotten around to digesting the whole thing yet18:07
fennheh there's a lot of nodes to digest18:07
kanzurewell, I refer to the framework as well18:09
kanzurehow they are doing this18:09
kanzureso the phenotypic description of wtf the brain is doing18:11
kanzurelet's magically assume we have a way to describe the function of the brain18:11
kanzurecould this information be routed back into the map in a constructive way ?18:11
fennyes, with physical "knockout" studies (lesions)18:13
fennbut i dont see a whole lot of value in that18:13
kanzurewasn't this what enki-2 was saying .. "since we can't say it, let's just specify everything that isn't right, and do it in prolog"18:14
kanzurethere were some objections to that method in a logic/computational context for something else18:14
kanzurepsychometrics are out of the question I think18:14
fennaroo? i mean burn some cells at a specific location and see what the functional effect is18:14
kanzureor molecular lesions btw18:14
fennhrm. molecular things are so much easier to directly correlate to genes18:15
fennpsychometrics is the same as your magical functional description18:16
kanzurepsychometrics uses that layer of grammar and natural language though18:16
kanzureit fits, yes, but18:16
fennnot entirely.. there's a lot of abstractions, like what does IQ represent18:17
fenn"it represents G, the general intelligence factor!"18:17
fennwhoopee18:17
kanzurevariables dependent on the organism's function during a very particular context for "IQ testing"18:17
fennbut they're aiming for a context independent measurement18:18
kanzurelet's not use IQ, that has a lot of baggage18:19
kanzurecan you make the point with another variable/abstraction?18:19
fennok, i dont know any other psychometric tests18:19
kanzureStroop attentional task18:19
kanzureit's where you read the color of a word that represents another color18:19
kanzure"red" but it's in blue18:19
fennok18:19
kanzureand the measurement is something like hit to miss ratio18:19
fennseems like its measuring some kind of cross talk18:19
fennis that a function?18:20
kanzurehm?18:20
kanzureis what a function, crosstalk?18:20
fennyeah18:20
kanzuresure, why not18:20
kanzurethat would be good to have on a map, if it exists18:20
kanzuredunno if the stroop test is measuring it ... that would be better suited for fMRI or something18:20
fennsemantic cross talk, not necessarily electrical18:21
kanzuredefinitely not necessarily electrical18:21
kanzuremaybe my framing is off18:22
kanzureI've been considering some other possible scenarios18:22
kanzurefor instance, just mapping *people* on to the map18:22
kanzureand saying "go look at kanzure for this"18:22
kanzurepointers/refs18:22
fennsure, each person is a combination of many functions in varying degrees18:22
kanzuredoesn't really help do anything though ...18:22
fennyou could decompose them into basis functions with enough samples18:23
kanzureyeah, how many?18:23
fenna lot18:23
kanzure"a lot" is a *shit* load, probably18:23
kanzureconsider the trillions of synapses18:23
kanzureand 20,000 brain-related genes18:23
fennwell, you dont have to get a 100% complete map18:23
fennits like a hypercube, most of the volume is in the corners18:24
kanzureokay, so let's talk about nematodes instead18:24
kanzure100 neurons, right?18:24
fennbut as long as you map the central region you have a good idea what a hypercube looks like18:24
fennwhat functionality does a nematode brain provide?18:24
kanzuresomeone in the torrent-vid that nsh hated yesterday, has been layering neurons on top of silicon circuit paths and testing the neurons individual with different wirings and connections .. using the surface as a scaffold, then listening to their conversations18:25
kanzuredunno ... surely it's doing something18:25
fennthat's a good idea18:25
fennbetter than poking at them with syringes :P18:25
kanzuredifferent silicon patterns for different networks18:25
kanzureyeah, realy18:25
kanzure*really18:25
fennits sorts like an inverse nanobot18:26
fenninstead of nanobots swarming into the rigid structure of the brain, you have brain cells swarm into the rigid structure of a microchip18:26
kanzurelayered the paths with some sort of food for the neurons to grow on18:27
kanzureperhaps it was hormone or something18:27
kanzureout of all of the things that you can say about the brain, what's the most general?18:28
kanzure(yet still correct/sayable)18:28
fennum, pass18:28
kanzureheh18:28
kanzureI'm still on the phenotype question18:28
kanzureclearly the brain is passing information about itself to others18:28
kanzurethus we have our 'memes' hypothesis18:29
kanzureideas/thoughts and the like.18:29
fennthat has nothing to do with nematodes18:29
kanzurebut it's by no means rigid18:29
kanzurehrm18:29
fenni was thinking something like 'responds to its environment' or 'responds to stimulus'18:29
fennbut then i got lost in semantics18:29
kanzureany observation that we record of its response is psychometric testing, no?18:30
kanzurelike "it wiggles (only) left"18:30
kanzureand then what's "wiggle" ?18:31
kanzurehurray for semantics18:31
fennbrain is much faster than most single-cell control mechanisms, i think that's why it evolved18:32
fenn"something's biting my tail, get it off me!"18:32
fennor for controlling circulatory system pumps where chemical diffusion isn't fast enough18:33
fennactually they do use sodium/potassium diffusion but it's phase locked18:34
kanzureultimately we do have the instruments to measure variables about brains, particularly look at patch clamp or the neurons-on-a-chip stuff,18:35
kanzurebtu we can't ask 20,000 users to slice open their heads, build these instruments, and make the measurements18:35
fennbah i hate google18:37
kanzureneurofeedback is interesting because it allows the user to isolate an internal variable ... many people claim to be able to control their pulse quite easily with the proper equipment, and then some training and soon enough they don't have to use the equipment any more.18:37
fennsure why not18:37
kanzureI was thinking of the natural language / psychology problem here with semantis18:38
kanzureand I figure that it's like trying to draw somethng through a damn pin hole18:38
fennah so you have an objective quantity to measure with an instrument18:38
kanzurebut having that neurofeedback component isolates an internal variable and can bring it to the forefront of attention18:38
kanzureyes18:38
fennso hopefully some people will be able to do this and others won't, and that means something?18:39
kanzureor the range of numbers they can generate perhaps?18:39
kanzurewould that be useful?18:39
* kanzure is afraid of the possibility of a lack of interesting measurements18:39
fenndunno, maybe if the isolated variable has some effect on magical semantic 'functionality'18:40
kanzures/generate/control/18:40
fennwhat are you going for exactly?18:40
fennend-goal18:40
kanzureextraction of phenotype-variables18:40
kanzureand their values18:40
fennwhy?18:40
kanzureto map on top of Google Mouse/earth18:40
fennwhy?18:40
fenn:)18:41
kanzuremostly to constrain my own research to figure out what areas to attack attention at18:41
kanzurethat was what it originally started out as18:41
kanzureI have all of this information, but it's pretty useless18:41
fennsay you finally invented the magical attention pill, what would you do with it?18:41
kanzureuse it18:42
kanzureit's not a pill btw, it's more like a health monitoring system18:42
kanzurethink of galaga except on steroids to destroy destractions automatically so you can continue working uninterrupted :-)18:42
fennhow do you determine what's a distraction?18:42
kanzureI don't even know what a 'distraction' is in terms of the brain or signals/noise18:43
kanzurewhatever it is, it stops me from doing really awesome work18:43
fennthe way i see it, when you're doing research, each answer opens up ten more questions.. which of those questions are distractions and which are important?18:43
kanzures/work/stuff/18:43
kanzuredon't know if that's the right way to put it18:44
kanzurethe way that distractions work for me are quite weird18:44
kanzureit's not a matter of useless research -- that's bound to happen anyway methinks unless you get more keen at selection -- it's like running through a program and not following it through to completion18:45
kanzureblah, you have me quite off track now18:47
fennits hard to finish projects when you dont have a specific end state18:48
kanzurebtw, a good solution to the Mouse problem is to just leave the instrumentation question open18:48
kanzuresure18:48
kanzure*open and extendable by other programmers18:48
kanzureactually the idea wasn't so much 'attention' as it was whatever I mentioned on humancortex.html18:53
kanzureI think the way I worded it was 'system administrator's guide to the brain" or not so much guide as framework for dealing with it18:53
kanzuredistraction-elimination would be one technique to employ, should it ever end up with an actual technique18:53
kanzureso in a sense it could be said that the Mouse/map thing could be used as a way to figure out what possible 'internal variables from that pinhole to draw through' and methods of dealing with them are possible18:54
kanzureit's easy to say in words, but not the implementation methinks18:54
kanzureEEG/MRI/neurofeedback on the sysadmin side of things18:54
kanzurebut again, finding easy variables to isolate from users? ehh ...18:55
kanzuresounds kinda useless 19:03
kanzurebut it's somewhat based off of the distinction between nearterm (me) versus longterm (something less fucked in the long run than I am) http://heybryan.org/recursion.html19:04
kanzureand buildingbrains.html19:04
* kanzure goes to eat19:04
kanzureHeh.19:20
kanzurehearing terrible tales of compiler errors from across the tables19:20
kanzureawesome.19:20
-!- Splicer [n=p@h208n2c1o261.bredband.skanova.com] has joined #hplusroadmap19:23
ybitsounds like fun19:27
ybitreferring to geek speak19:27
kanzurefenn: http://p2pfoundation.net/Category:Design19:57
kanzurehm20:09
kanzurephase-sensitive optical low coherence reflectometery? http://www.google.com/url?sa=t&ct=res&cd=12&url=http%3A%2F%2Fwww.opticsexpress.org%2Fviewmedia.cfm%3Furi%3Doe-12-11-2377%26seq%3D0&ei=QBiJSJOLDqjeigGDj7zSDA&usg=AFQjCNFizM-n14E1yRHQHVfg6fT5cJb4Xg&sig2=WeSUyJPuAdjmux-KNaTXEA20:09
kanzureaka backscattered light upon action potentials20:09
kanzurethis is another one picked up from whatever the hell it is that Superkuh is up to20:09
kanzureR. U. Maheswari, H. Takaoka, H. Kadono, R. Homma, and M. Tanifuji, “Novel functional imaging technique20:10
kanzurefrom brain surface with optical coherence tomography enabling visualization of depth resolved functional20:10
kanzurestructure in vivo,” J. Neuroscience Methods 124, 83-92 (2003).20:10
kanzureNow doesn't that sound promising?20:10
kanzureT. Akkin, “Biomedical applications of a fiber based low-coherence interferometer for quantitative differential phase measurements,” Dissertation: The University of Texas at Austin, Austin, (2003).20:11
kanzure   Rapid mechanical changes were reported in the garfish olfactory nerve using a stylus recording device fastened to a piezoceramic bender [6,7]. The recorded swelling signal and20:13
kanzureaha, that's why he's looking for the Tasaki paper20:13
kanzurefinding this one might be interesting: K. Iwasa, and I. Tasaki, “Mechanical changes in squid giant axons associated with production of action potentials,” Biochem. Biophysic. Res. Comm. 95, 1328-1331 (1980).20:13
fennreminds me of flanagan's neurophone20:17
kanzureinteresting idea20:22
kanzuredoes it .. work?20:22
* kanzure was reading http://www.neurophone.com/tech.htm20:22
-!- jm [n=jm@p57B9FD0C.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)]20:40
fennhmm they changed their story.. used to be claiming that you could send information down ordinary tactile-sense nerves and it would be decoded as audio by the brain21:23
fennhe started out using high voltage pads you stuck to your thigh or palm or whatever21:24
fennthen they figured out that the high voltage would cause the skin to deform like a piezo actuator, so they tried it with piezo's and got the same result21:25
fennthe claim used to be that even nerve-deaf people could hear with it21:25
kanzureodd :)21:52
ybitnext year, aarpa can do my taxes22:25
ybitick22:25
ybitthat´s been on the to-do list for awhile now :P22:25
ybithave any of you read neverness?22:26
ybitthat´s on the todo list22:26
ybitfenn, i´m guessing since that it´s on your sever you have read the book?22:26
ybitaarp*..22:27
kanzureybit: I've read it.22:33
kanzureI'm the one recommending it to all of you :)22:33
fennybit: just finished it22:47
fennuh, i think he covered all the bases :)22:47
kanzurewho?22:48
fennegan22:48
kanzureneverness is zindell22:48
fennah whoops a bit of selective blindness on my part, sorry22:48
kanzurebut still, that's what I was wondering22:48
kanzureyou might have meant "bryan covered the bases by saying he's the one recommending it"22:49
fennok, i just finished diaspora22:49
fennshooting people with introdus would make a great "non-violent" video game22:51
kanzure(2008-07-24 22:11:26) kanzure: Hey.23:10
kanzure(2008-07-24 22:11:32) Carissa: hello23:10
kanzure(2008-07-24 22:11:41) Carissa: it works?23:10
kanzure(2008-07-24 22:12:23) Carissa: it took me forever to figure out why this thing wouldn't work23:10
kanzure(2008-07-24 22:12:46) Carissa: oh? were not on pidgin are we???????????///23:10
kanzure(2008-07-24 22:12:51) Carissa has left the conversation.23:11
kanzure(2008-07-24 22:14:11) kanzure: I'm on pidgin, but you might not be.23:11
kanzure(2008-07-24 22:14:28) Carissa: it doesn't say i am23:11
kanzure(2008-07-24 22:14:36) kanzure: Are you in gmail?23:11
kanzure(2008-07-24 22:14:49) Carissa: no i'm on theinternet23:11
kanzure(2008-07-24 22:14:58) kanzure: You're in gmail.23:11
kanzure(2008-07-24 22:15:01) Carissa: no23:11
kanzureHow cute.23:11
fennuh, check your clock?23:15
fenn"Aromantic asexual. And atheist! Triple A rating!"23:17
kanzurewtf23:17
kanzuremy clock is wrong?23:18
kanzurealso, the p2pfoundation guy did Technocalyps, nsh23:18
kanzurewhat an odd coincidence23:18
fenneither that, or an insidious conspiracy23:19
kanzureinsidious conspiracy for 200 please23:19
fennyour answer: "The richest collection of free modes available anywhere in the net and on the planet."23:21
fennwhat was the question?23:21
kanzurewhere arez teh porns?23:24
fennhttp://thehomebased.com/?p=2823:26
fenni still can't figure out the 12th pic23:27
kanzurewhat an oddly specific answer :-)23:28
kanzureconspiracy!23:28
ybitkanzure: haha, i´m assuming the above was an excerpt from you and your little sis ^^23:41
ybitargh, i am indeed getting tired, hopefully i´ll get my provigil soon23:42
kanzureyes, Carissa is she23:42
ybitmy mum just mentioning going to dubai, india for a vacation o_O23:42
ybiti wouldn´t mind at all, but when she realizes that ´dubail´ is india, she´ll change her mind quickly23:43
kanzuremy mom's evil twin (or is that the other way around) was living in India a few years after living in Taiwan, China, Spain, England, all sorts of assbackwards places.23:43
ybitthat´s cool, but spain and england aren´t assbackwards :)23:43
kanzureon the contrary23:43
kanzureever take a dump in England?23:44
kanzurein a toilet23:44
ybitlol23:44
ybiti have not23:44
kanzureexactly23:44
ybitlol23:44
ybitbut i know people who have (besides you) :)23:44
kanzureI haven't.23:44
ybitthey don´t have backwards asses23:44
ybitheh23:45
kanzurebackwards arses23:45
kanzurefine23:45
ybitgetting english on me are you?23:45
kanzurei bloody well ought to23:45
ybiti´m tasked with choosing a vacation resort23:45
kanzureAustin, Texas.23:45
kanzureor Bloomington, Indiana and you can bug fenn23:46
* ybit looks for resorts in spain23:46
ybithttp://en.wikipedia.org/wiki/Category:Seaside_resorts23:47
ybitbah23:48
ybitspain doesn´t have any nice resorts23:48
kanzureI think I'll give up on attention23:49
kanzurebullshit idea anyway23:49

Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!