2008-06-24.log

--- Day changed Tue Jun 24 2008
-!- Phreedom [n=freedom@www.online.dn.ua] has quit [Read error: 113 (No route to host)]00:25
-!- You're now known as fenn03:17
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has quit [Read error: 110 (Connection timed out)]04:12
-!- nsh [n=nsh@eduroam-80.uta.fi] has joined #hplusroadmap04:26
-!- nsh- [n=nsh@wikipedia/nsh] has joined #hplusroadmap04:30
-!- nsh- [n=nsh@wikipedia/nsh] has quit []06:13
-!- Phreedom [n=freedom@www.online.dn.ua] has joined #hplusroadmap08:02
-!- Vedestin [n=Vedestin@d58-111-95-212.sbr3.nsw.optusnet.com.au] has joined #hplusroadmap08:09
-!- Vedestin [n=Vedestin@d58-111-95-212.sbr3.nsw.optusnet.com.au] has quit ["User pushed the X - because it's Xtra, baby"]09:39
-!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has joined #hplusroadmap10:47
kanzureHey.10:48
kanzureSo I had an idea yesterday.10:48
kanzureI want to run it by you guys.10:48
kanzurehttp://heybryan.org/self_replication.html10:48
kanzureWhat I've been mentioning is that self-replication is kind of like a single point-of-failure. It doesn't fail, but I mean that if your system component isn't integrated into everything else, it's not going to be self-replicating. 10:49
kanzureTwo howevers:10:49
kanzure(1) What if the parent unit did a secondary construction process that gave the child unit more functionality? For example, a human -> baby, but the hu gives the kid <something that the hu parent makes>.10:50
kanzure(2) And what if anything that isn't fully integrated into the self-replication process was completely "replaceable", in the sense that "I can use my arms to take them off and replace them with something else" ? Would that be a way to cheat the definition of self-replication?10:50
kanzureIt occured to my sometime yesterday, so I threw up the web page and ran off.10:51
kanzureSince the real self-replicator in the organism is the ribosome, I was wondering how to compartmentalize the ribosome. At the moment there's the nascent chain stopping techniques in ribosome display, as well as various emulsion techniques, lysosome techniques, but those don't really interest me. 11:04
kanzureI came up with a list of all possible ribosome compartmentalization techniques: have it output into a shell that is directly attached; have it attach to a surface that has deposition holes in it; have it make its own shell that wraps completely around it; 11:05
kanzureAll of these require molecular mechanisms that are, frankly, too hard to come up with. It would be great if we could have a ribosomal machine that makes its own buckyball to sleep in, but I'm not sure I want to go yakshaving on that.11:05
kanzureAlso, the fact that it's all protein folding really, really sucks, and means that you're not really going to be laying down exact wires for electronics ever, that's unacceptable. But one possible method -- if we really, really wanted the self-replicable, evolvable ribosomal manufacturing unit, is to use an STM tip to inject mRNA into a ribosome at the end of the tip, and then the output channel would go into another nanotubule as 11:07
* kanzure is flipping through some notes11:08
kanzureSince we can do electromechanical integration pretty well (brain interfaces aren't impossible, we can get a few bits out of the system, things are looking good), and when we considering physical manufacturing, where's the crossover? The best answer I can come up with is "your hydraulics system" (motors that pump water through pipes). We can work with electricity and magnetism pretty well, but not 11:09
kanzuremass/matter, which I'm still stuck on. Supposedly the answer would be the ribosome (digital (DNA) data => physical manufacturing with mass/matter (the amino acids)), but you lose the precision that is required.11:09
nshprecision is an artifact11:10
kanzuremeh?11:10
kanzurethe context that I'm considering is, say, electrical wiring11:10
nshit is part of a paradigm that is untenable in the biological domain11:10
kanzureright11:10
kanzurewell,11:10
kanzurenot without some very fancy evolutionary experimentation11:10
kanzurei.e., see spirals11:11
kanzurebut that's not going to help me here11:11
nshour understanding of engineering is distinctly idiosynchratic to the various accidents of our development (paper, metal, steam, etc.)11:11
kanzurewhile their discovery and applications are accidental,11:11
kanzureour ability to go dig a trench and make a brick wall 11:11
kanzureis not.11:11
* nsh nods11:11
nshbut the underlying metaconcepts are not as amenable on the nanoscale as they are on the mesoscale11:12
kanzuremeh, I don't care much about that11:12
kanzurethe scale doesn't matter that much to me11:12
nshunfortunately, physics does :-)11:12
kanzureThe reason why I was thinking about ribosome is because it's the interface between digital information and manufacturing11:12
kanzurebut I'm not in love with ribosomes.11:12
kanzureSo, one idea that I have had in the back of my head that might be worth mentioning :-)11:13
kanzurehave you heard of a protein array?11:13
kanzureI think it's just this giant 2D array of proteins tied on to a surface11:13
kanzureand then there's some way to do interaction studies11:13
kanzureUsually this is paired with mutagenesis to see what changes might do to the proteins and the interactions and whatever.11:13
kanzureThat sounds cool. ;-)11:13
kanzureWhat I'm considering is exploding the ribosome into its subunits (it's made up of other proteins that bind together)11:13
kanzureand then from the action sites that we have determined from the literature, split those up as well11:13
kanzureand then I want to individually test all of the action sites with a big, giant repetitive experiment11:14
kanzurei.e., I want to reverse engineer the ribosome11:14
kanzureand see what problem space that evolution was playing around with11:14
nshinteresting idea11:14
kanzureso, instead of proteins on an array, we have partially folded polypeptides, maybe some proteins sometimes, mutated proteins to see what the differences would be, etc.11:14
kanzureI'm assuming that there's -- to some extent -- a mapping between the physical interactions and the actual problem that evolution struggled with11:15
kanzurebut all of this isn't a guarantee at all -- it's kind of a last resort scenario11:15
kanzureI'm pretty sure there should be some way to do this on the macroscale, ignoring biology completely. 11:15
kanzurewhere 'this' = that interface between digital/computation and manufacturing11:16
kanzurefrankly I find it odd that the interface is "the ribosome" -- a named, single unit11:17
kanzureShouldn't the design of the ribosome be implementable on a larger scale too?11:18
kanzurehttp://www.google.com/search?q=%22design+of+the+ribosome%22 grr, first result is Kurzweil11:18
* nsh chuckles11:21
nsha potential problem is that much of the historical context to the evolutionary development may be lost11:22
kanzuretechnically the proteins fold mostly /outside/ the ribosome, so maybe my talk of ribosomes is not needed11:22
kanzureyep11:22
nshas the environment has changed along with the organism11:22
kanzurethe problem that evolution was solving probably wasn't protein folding11:22
kanzuresince protein folding is of the amino acids themselves, no?11:22
kanzureand "solving" protein folding would mean something like being able to digitally specify what protein to make11:23
kanzureand as far as we know, no biological system does this11:23
kanzure(not even us unfortunately)11:23
* nsh smiles11:23
nshit's dangerous to implicitly imagine a biological 'problems' in the same manner as problems solved by conscious agents11:24
nshs/a //11:24
kanzureof course11:24
kanzureon that note, of dangerous things with bio, http://heybryan.org/evolution.html11:24
nshthere is usually no knowledge of the challenge, nor what would consitute success, save that ones genes are around where they might not have bind11:24
nshaparatus has no perspective11:25
nshmm11:25
kanzurewhy is polymerase and ribosome two separate entities?11:26
kanzurebesides the obvious locality issues?11:26
nshbecause we define entity based on a molecular bias?11:26
kanzurenot entirely, the ribosome is made up of multiple proteins for example11:26
* nsh is always dubious about entities and discrete components11:26
kanzureif you can show me somebody who has a nondiscrete, but completely usable interpretation of molecular biology, please :) be my guest11:27
* nsh smiles11:27
* kanzure realizes that saying 'molecular biology' is against the idea of nonentitious interpretations11:27
kanzureso maybe call it nondiscrete nonentititous process-theoretic nanoscale biology11:28
kanzureor maybe information theoretic, whatever11:28
kanzureYep, I was wrong.11:29
kanzureI forgot that the coding elements themselves are really the special features here -- the code specifies the proteins, everything else is just facilitation11:30
nshhmm11:30
kanzureso, escaping the biological domain, let me pull out a little bit11:30
kanzurethe important aspect is that mass/material can be integrated because of the digital specifications (coding)11:31
kanzurethat's what I'm trying to capture here too; it's directly applicable to the design compiler and the other ideas that float around in this channel11:31
nsh"Those original emergent chemical conditions were dependent on the environment -- which we can determine from the rock records to some extent -- as well as by the fact that we figure they would have to be fairly simple, not some monstrous process more complicated than current organisms [although this alternative hasn't been extensively explored to my knowledge]."11:31
kanzure:)11:31
nshi like that you considered that.11:31
kanzureI can do electrical interfaces very, very well -- computation, discrete stuff, it's all good11:33
kanzurebut /me just can't figure out how to deal with mass/material11:33
kanzureit just kind of sits there11:33
kanzureand you need to pick it up11:33
kanzurebut to pick it up you need a motor11:33
kanzurebut to get that motor in the first place you need that mass11:33
kanzureI suppose it's an "all or nothing principle" -- the food for an organism supplies every single aspect of it, right on down to the nucleic acids. Whereas the supplies to a manufacturing process are just 'touched'.11:35
nshi'm not sure i follow11:35
nshyou're referring to consumption vs. manipulation?11:35
kanzureperhaps; ultimately what I am trying to do here is escape object-orientation in manufacturing11:36
kanzurefenn mentioned a few days ago the idea of having constraints and genetic algorithms and other evolutionary techniques so that "something like a spoon" could be manufactured for a process,11:36
nshthat'd be nice11:36
kanzurebut that's still object-orientation, no? 11:36
kanzureyes, it would be :)11:36
kanzureI thought about it later that day and I kind of just sat there for a while until I came to a lengthy conclusion11:37
* kanzure takes a deep breath11:37
kanzureSo, it kind of relates back to my two approaches to brains at this point11:37
kanzureas you might know, I like the idea of self-augmentation http://heybryan.org/recursion.html11:37
kanzurebut at the same time I am also a programmer and know when the hell to sandbox changes11:38
kanzuredon't want to run around deleting neurotransmitter receptors, do we? ;-) BAD, BAD11:38
* nsh smiles11:38
kanzureat the same time it also kind of sucks11:38
kanzuresince we can't get much information out of these brains11:38
nshwell, maybe could freeze a few...11:38
kanzureyes, we can try freezing them, we can try implants, we can try augmentation their structures and so on11:38
kanzureyes11:38
kanzurebut I'd rather have something more guaranteed11:38
kanzurewe've frozen cat brains before (7 years); and they did show electrical activity11:39
kanzureand I'd be willing to do that, but in the mean time I'd rather also try some other solutions11:39
kanzureanyway, we're basically screwed like that11:39
kanzurewhich is fine, I'm okay with that, because there are some ways that I can cope with that (freezing, implants, whatever)11:39
kanzurebut what about a brain that isn't so majorily screwed?11:39
kanzurethat's the "next generation" stuff -- the "building brains", neurofarms, neuropods, using Markram's simulations and so on11:40
kanzurewhich, coincidentially, are very useful for figuring out what changes to make per http://heybryan.org/recursion.html11:40
kanzure*cough* it's not a coincidence11:40
nshhmm11:40
nshwhat are markram's simulations?11:40
kanzurewhole brain simulations11:40
kanzurehe has accurate simulations of cortical columns11:40
kanzurei.e., he throws all of these molecular, neurophysiological, ion channel, all of these models together11:41
kanzurehttp://heybryan.org/mediawiki/index.php/Henry_Markram11:41
kanzurethere's an amazing video fi you haven't seen it yet, it's on the second link on that page, and then click the 'talk' link (it might be the first) on the resulting page11:41
nshoh, this was in the google video linked here a while back, right?11:41
kanzureyes11:41
nshright, cool11:41
nshnever got around to giving that my full attention11:41
kanzureokay, so we can build brains :) cool11:41
fennyak yak yak yak11:41
kanzureheh, that's partly my point in all of this -- attention, etc.11:41
kanzurefenn: bah, hold on11:41
fennribosome is 98% RNA11:41
fennthe protein is just there to stabilize it11:42
kanzuregrr, train of thought is crashing11:42
kanzureanyway, my idea was that those brains that we could supposedly build11:42
kanzurethe ones that won't be screwed11:42
kanzurethe ones that could be compartmentalized and digitized11:42
kanzurethat's where we'd need that computation-manufacturing interface, no?11:42
kanzurethat's where losing object-orientation would be a Good Idea.11:42
kanzureespecially if the brains are able to manufacture themselves, or if the brain farm is completely self-replicable11:43
kanzurebrains can already self-replicate to some extent, no? 11:43
nshagency problems again11:43
kanzurehm?11:43
nshis it the brain self-replicating, or the toenails11:43
nshor the genome11:43
kanzureno, hold on11:43
kanzureso adding in machinery (like a petri dish + MEA + other equipment) or infrastructure (to manage a huge farm) ... which can't make itself much? That's not a good idea.11:44
kanzureI think I have a better reason somewhere 11:44
kanzurebetter than "it's not a good idea"11:44
kanzureoh,11:44
kanzureit was the issue that (1) the brains can do electromechanical interfaces very, very well already11:44
kanzurebut (2) dealing with mass/matter on a scale other than "it's what sustains life and all of the nucleic acids and all of the damn proteins in the body, you fool" is the difficult part11:44
kanzureand where I suspect the lack of object-orientation can help out.11:45
kanzureaha. there we go.11:45
nshwhat are you defining mass/matter in opposition to?11:45
nshinformation?11:45
kanzurefenn: by 'ribosome is 98% RNA' does that mean mRNA or ribosomal RNA (rRNA) ?11:45
fennribosomal rna11:45
fennribozyme11:45
kanzurensh: well, I'm just considering it in opposition to electricity for example11:45
kanzurewhich I know is bad11:45
kanzuresince the electron does in fact have mass11:45
nshhmm11:45
kanzurebut consider E=MC^2 or something silly like that11:45
kanzureI hate to bring that up, but the mass-energy equivalencies for example11:46
kanzurewhile they might be 'equivalent', they are of different forms11:46
fenna topological defect in the fabric of spacetime11:46
nshbut is the difference 'fundamental', or merely a product of our perceptual and conceptual dichotomising?11:47
nshafaik, there is no evidence that nature necessarily conforms to a binary logic11:47
kanzurethere has to be some way to integrate manufacturing in a /useful/ way without shoving it into the 'fundamental food' of the system (by this I mean, bio's food, which fuels the whole damn system)11:47
fennoccasionally subatomic particles decay into pure energy11:47
fenni'm not sure about the reverse11:48
kanzureif the design of the system doesn't incorporate the variety that we need it to execute, does it mean that the system can't effectively deal with that variety? by this I refer to the variety from the periodic table11:48
kanzure*need it to execute/implement11:48
kanzurethat's where I came up with the original statements in here today11:48
kanzurethe idea of 'replaceable components'11:48
kanzureor the idea of 'the parent self-replicates, but then adds in some componentry as well' 11:49
kanzureand whether or not those can meet the definition of self-replication11:49
kanzurehttp://heybryan.org/self_replication.html11:49
kanzurensh: so, is the difference fundamental?11:49
kanzureI'm wondering if I'm going to randomly run into a roadblock due to a lack of understanding of gravity or something :-p11:49
kanzureI hope not :)11:49
fennyou've been treating a replicator as a particle when it's really a wavicle11:50
* nsh smiles11:50
fenni.e. some things are more strongly associated with perpetuating the system than others11:50
kanzureokay, sure11:50
fennthe wave 'passes through' the steak you had for dinner11:50
kanzurehuh?11:50
fennis the cow part of kanzure?11:51
kanzurehow could a dependency loop include processes that are not depended on?11:51
* nsh decides to walk home, thinking on these things11:51
fennyou hjust set your sights too narrow11:52
fennever heard of the gaia hypothesis?11:52
kanzureyes11:52
fennwell, even earth isnt a closed system11:52
fennwe wouldn't be here if it weren't for water comets bombarding the earth for billions of years11:52
fennso, where's the replicator?11:52
kanzurebut you said it's a wavicle, implying that it's somewhat like a particle11:53
kanzurewhat about this notion?11:53
kanzurescrew the replicator as an object, just imagine it's untouchable in that sense11:53
fennindeed, most of the functionality is in a nice little package you can send up to a space station and around the edge of infinity and back11:53
kanzureit has some functionality that we may or may not know about11:54
kanzurebut the only things that might be 'objects' are the decisions that we make in order to interface with it or hack it11:54
kanzurewould that be appropriate?11:54
kanzurefenn: no, your example of a cow disagrees with the edge of infinity journey11:54
kanzurei.e., dependency on the biosphere processes11:54
-!- nsh [n=nsh@eduroam-80.uta.fi] has quit [Client Quit]11:54
kanzureand autotrophs and whatnot to gather energy11:55
fennhm i guess11:55
kanzurebut I'm not so much concerned about "where's the replicator"11:55
kanzureoh11:55
kanzurewell, I guess you were asking that wrt my question of whether or not it's okay to have "replaceable parts" or something11:55
kanzurei.e., hu makes a kid, but kid doesn't have arms, hu makes arms, gives kid the arms later11:55
kanzureand then if the kid can do the same, then it's a replicator11:55
kanzurethis is my way of getting around the manufacturing problem11:56
fenni dont get the analogy11:56
fennwhat manufacturing problem?11:56
kanzurewe can do electromechanical interfacing stuff very, very well, but managing materials -- without having them implemented in our food and abilities to eat food and turn them into our own bodies -- is foreign so far.11:56
kanzurethe bootstrapping problem11:56
fennwhat's "electromechanical interfacing"?11:56
kanzuremany things:11:56
kanzureyour interface with your arm11:57
kanzuremy interface with the computer11:57
fennmy arm doesnt have an interface11:57
kanzureyou have neurons to it11:57
kanzureto the muscles11:57
kanzurebrain implants would also be appropriate11:57
fenni can solder wires to random points in a circuit, that doesn't make it an interface11:57
kanzuresemantics11:58
kanzuresuggest another word for it11:58
fenni use my arm to do stuff?11:58
fennlike press buttons11:58
fennbut my arm isn't well suited for manipulating molecules, is that what you're getting at?11:59
kanzurenot quite?11:59
kanzuremaybe this is a better way to put it11:59
kanzureyou can retool your arm, or various parts of your brain11:59
kanzurebut if we wanted the hu embryo to grow into a human + spaceship, we'd only have one option at the moment12:00
kanzurebasically hoping that the hu child would make it12:00
kanzurei.e., CAN make it versus IT WILL make it12:00
kanzurewhereas, if we are to take the approach of a self-replicator, isn't it supposed to be completely integrated into the design of the system? I mean, isn't it supposed to be automatic and not something left up to chance?12:00
kanzurethis is hard to explain12:00
kanzureI really thought that this was the same thing as the bootstrapping problem that we face12:01
kanzurewe have implementation ideas, but now we need the matter/energy in order to make it happen12:01
kanzure(bootstrapping of autogenix and the actual, physical manufacturing)12:02
kanzureI don't want wavicles, I want just 'waves' (I'd rather call it processes, computation, etc.). 12:05
kanzurethere's no reason manufacturing must be object-oriented12:05
fennOT i'm concerned about how i didn't notice anything weird about this guy when he was at my house last week, talked to him for a couple hours and he seemed totally normal, although according to others he was abnormal: http://www.idsnews.com/news/story.aspx?id=51308&comview=112:05
fennsupposedly he had been strung out on acid for several weeks12:07
kanzurehuh12:07
kanzurehere's the challenge:12:09
kanzurefigure out a way to consider mass/materials/matter without object-orientation12:09
kanzurefor example, a rock12:09
kanzurethe context is it's on my front lawn12:09
kanzurewith energy like electricity we can run around collecting it and charging batteries, running motors and generators and do whatever we like 12:12
kanzurewith mass, though, what does it do but sit there? Some of the materials could be used to make the electrical systems that we need,12:12
kanzurebut those systems aren't self-replicating (argh!)12:12
kanzure"Hi RepRap! Here's the rock from my front yard!."12:12
kanzure<reprap just sits there>12:13
kanzurethe typical case in the biosphere is for the hundreds of trillions of prokaryotes to run around and possibly integrate the new mass/material into their systems12:14
kanzureand then you can trace them through the ecosystems and food chains12:14
kanzurebut surely there's a way to do it without aggregating organisms like that ?12:15
fennand then hopefully the prokaryotes are easier to digest than the rocks12:15
kanzureand without integrating it into the fundamental-food-of-the-system12:15
fennplants do it all the time12:15
kanzureit's basically (1) organism eats some stuff it doesn't know about, something that could be toxic on a large scale or something12:16
kanzure(2) the organism has some evolutionary pressure to mutate, or it just does so anyway12:16
kanzureand then by the grace of #2, #1 might come into use12:16
kanzureit's all an accident12:16
kanzurewhereas our brains seem to acknowledge these materials and objects that we have around us12:17
kanzuremaybe I should just ignore it all12:17
kanzurebut then how would you ever extend your functionality beyond your hands ?12:17
fennone aspect of intelligence is modeling things outside experience12:17
kanzurenaughty12:18
kanzureyou know I don't like the I word12:18
fenntoo bad12:18
kanzureheh12:18
fenni was going to say "brain" but that was too narrow12:18
fennand inaccurate in any case12:18
fennso, it seems you want a "theory of everything" really a model of everything12:19
kanzureI hope not12:19
kanzureif that's the case then I can accept the errors of my thoughts :)12:19
fenn"with mass, though, what does it do but sit there?" is what made me think that12:20
kanzuresometimes people treat 'whole systems' as if its own 'process'12:20
kanzureas if its own 'spirit' or 'soul' or 'directed action' or something12:20
fennsure, "om" is the sound of the entire universe12:20
kanzureI don't know how to say that and I haven't considered if it's bunk12:20
kanzureso, that 'holism'/emergence works in a number of interesting cases12:21
kanzureant colonies, the brain and especially intelligence, etc.12:21
kanzureand maybe even debian12:21
fenncybernetic systems hah12:21
fennlet the buzzwords fly12:22
kanzuresure, cybernetic systems12:22
kanzureokay12:22
kanzureso given those buzzwords12:22
kanzureblah, I guess the real problem is that I don't actually have much of a problem that I am trying to solve here12:22
kanzureI'm losing my train of thought12:22
kanzureI really thought that the bootstrapping problem was a serious one12:23
fennwe're still alive mate, no worries :)12:23
fennits 99% done because of that12:23
fennits the old 99/1 problem12:23
fenni think you're butting heads with the replicator myth12:24
fennthere is no replicator12:24
fennwe're the all-singing all-dancing crap of the world12:25
kanzureit also might be my distinction between electricity/energy and mass/materials12:25
kanzureor the distinction between process and object12:25
kanzureoh12:26
kanzureI guess objects could just be aggregated skdb files in the end anyway12:26
kanzureand some of these might come with a physical ,actual attachment12:26
kanzurei.e., "download chair;" "are we still in Austin, Texas?" "yes" "supplies negotiated and will arrive by noon"12:26
kanzure(that's assuming a social backbone to that individual's system)12:27
fennyour naive conception of 'electricity' as a thing shows that you havent worked with it much12:27
kanzureor something :-/12:27
kanzurewhat do you mean though?12:27
kanzureI mena, the brain is able to interface with electrical systems, but we can't interface with a rock12:27
fennyou cant just run any old voltage through the tubes12:27
kanzurewithout busting our heads open :)12:27
kanzurethat's true12:27
fennlikewise, where does that electricity come from?12:27
kanzurewell, do we mean the flow of electrons?12:28
fennit's the result of charged or magnetic masses moving about12:28
fennthere's also the photoelectric effect, but those photons are the result of masses oscillating12:28
kanzureI guess I could make an analogy to molecular computation12:29
kanzuremost of our computation and so on is done in terms of energy, processes, and information12:29
kanzureI haven't seen an information theoretic approach to matter12:29
fennanyway, you say 'we can use electricity to run motors and stuff' but its sorta like saying 'we can use water to turn turbines to make electricity'12:29
kanzureok12:30
kanzureyes, you're right12:30
fenndo i get a cookie?12:30
kanzureso my interface distinction is null and void12:30
kanzurebut then why is it that we can program and make digital circuits12:30
fennbecause the mathematical models for electrical circuits are very simple, linear, and predictable12:31
kanzureand yet our 'circuits' with matter (I'm imagining either manufacturing facilities or giant chemistry lab setups) aren't similarly programmed?12:31
kanzureand why can't this programming be integrated into my own programming12:31
* kanzure whines12:31
fennyou cant integrate digital circuits either12:31
kanzurehm?12:31
kanzurebrain implants?12:31
fennyou can't just stick a DIP package into your skull and be done with it12:32
kanzureAND gates made out of neurons?12:32
kanzureheh :)12:32
kanzurehttp://heybryan.org/docs/neuro/12:32
fenneven if you had a brain implant, you couldnt 'download' procedural memory12:32
kanzureright12:32
fennthis is why we have dreams12:32
fennthe temporary storage buffer neurons have to 'teach' the long term memory neurons how to do stuff, and the only way is to re-live the experiences needed to use those procedures12:33
fenner, experiences that cause those procedures to be used12:33
fennso, you'd have to have a programmable dream machine i guess12:34
kanzureyou're working off of your statmeent of 'modeling things that haven't happened yet' ?12:34
kanzureI'm guessing.12:34
fenni didnt quite say that12:34
kanzureerm12:34
fennmodeling things outside of our experience12:35
kanzureyes, I realize I butchered that12:35
fennlike, i can picture you sitting there in front of your computer12:35
fenncat on the bed, too many books on the floor12:35
kanzureI do not want to have 'intelligence' as a requirement of skdb12:36
kanzureever.12:36
kanzureblargh, as a fundamental requirement12:37
kanzureas an optional requirement in some random package or whatever, meh12:37
fennwhat are the fundamental goals of skdb?12:37
kanzurewell, originally it was to bootstrap a self-replicating machine, namely its design12:37
kanzureand this was because of the requirements for recursion.html12:37
kanzureat least from my perspective12:38
kanzureplus all of the extras that it offers, like DIY manufacturing and such12:38
fenn'plus' is not a fundamental goal is it?12:38
kanzurenope12:38
fennso, you have to define what  self replicating machine is12:38
fenndon't feel bad if you can't come up with an answer, i dont think biologist can define life either12:39
kanzuremy ability to recombine my genome with a woman's12:39
* kanzure cheated12:39
fennyeah that has nothing to do with self replication12:39
kanzureI took 'self' seriously12:39
kanzureokay, so the idea was to get exponential replication 12:40
fennin any case, your child would be only halfway you (and none of your environmental influences)12:40
kanzureor at least exponential duplication or whatever12:40
kanzuremost manufacturing and building is n^2 and I want 2^n12:40
fennexponential growth12:41
kanzureyes12:41
fennin that case the only 'self' is the whole growing system12:41
* kanzure wonders if the definition of 'growth' is important here12:41
fennyep12:41
fenngrowth is based on functionality i think12:42
kanzurewhen I was thinking about exponential self-replication for making brain farms and brain simulations I was thinking of clanking replicators12:42
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has joined #hplusroadmap12:42
kanzurebut that's object-oriented12:42
kanzurehow about a nonobject-oriented definition of growth?12:42
kanzureuhm12:42
fennwell, an increase in the number of functions possible would be considered growth12:43
kanzurewell, the goal was to do lots of simulations of brains or at least physically implemented neural tissue slices for experimentation12:43
fennthis reminds me of the AI "hard takeoff" bootstrapping scenario12:43
kanzureso as long as that goal references objects, the definition of growth will be object-oriented12:43
kanzuresure12:43
kanzurelinux treats procs as objects12:44
fennthey are objects, because if linux didnt herd them around they'd die12:44
fennthat didnt come out right12:45
fenni dont like this obsession about whether something is object oriented or not12:46
fennit seems to be religiously motivated12:46
kanzureit's shorthand for a deeper issue12:46
kanzurewe don't care about the objects, we care about the functionality12:47
fennwhen you speak of 'a rock' the object is the model used to represent the rock, right?12:47
kanzurethe object is the rock12:47
kanzureit's right there12:47
kanzuresitting there12:47
kanzure(on my lawn)12:47
fennbut its just a bunch of atoms and energy12:47
kanzureright, so what the hell are we going to do with ti12:47
kanzure*it12:47
kanzureespecially with respect to our exponential growth system12:48
fennyou dont care about the absolute positions of all the electrons in the rock at a particular time12:48
kanzureright12:48
fenn(unless you do)12:48
kanzureI don't think I do12:48
kanzurewhat is your statement trying to say?12:48
kanzureis it asking if I do?12:48
fennbut what if the rock had consciousness and you wanted to do a 'rock scan' so you can upload it into rock heaven?12:49
kanzuregah?12:49
fennobject orientation is mainly a way to organize data by relevance12:49
kanzurecontext flip for a sec12:49
kanzuredo you recall my work on a matter beam?12:50
fennnot really12:50
kanzureok, well, laser beams we have12:50
kanzurewe also have matter beams12:50
kanzurehttp://heybryan.org/projects/atoms/12:50
fennok, ion laser, big deal12:50
kanzureno12:50
kanzureBose-Einstein Condensates12:50
kanzureit's where matter functions as a wave12:50
kanzureyou have to get it ultracold12:50
fennyeah i know, and then you can slap a hologram onto it and a puppy coalesces out of the mists12:50
kanzuresomething like e-9 torr12:51
fennthat's pressure12:51
kanzureignore the hologram shit12:51
fennwhy? it makes sense12:51
kanzureI don't care12:51
kanzuremy point is that the matter is a beam/wave12:51
fennuh, sorta12:51
fennit's a coherent system12:51
kanzurefor some reason the temperature difference forces us to treat it differently when it's at room temperature or something12:51
kanzurewhereas we get to treat the flow of electrons as a, well, flow12:52
kanzurebut a flow of rocks on a conveyor belt has to be a flow of rock-objects, :-/12:52
kanzureI guess we could replace flow-of-rock-objects with a differential equation12:52
kanzurethat describes the damage done to the walls of the conveyor belt system or something12:52
kanzureright?12:52
kanzureand then we could just aggregate these models together in the typical skdb sense that we've considered before 12:52
fennno, an equation is not a real thing12:53
fennyou can model the rocks with a differential equation12:53
kanzurehow the hell do you model a lump of matter with a diff-eq?12:53
kanzureit's mostly constant12:53
fenni dunno, what's a diff-eq12:53
kanzure"yep, still sitting there"12:53
kanzuredifferential equation12:53
fenndid you pay attention to what i said earlier about why we use electronics for digital logic circuits?12:54
kanzuremodelability?12:54
fennthe models correspond to observations very well, and it makes using electronics very predictable12:55
kanzureoh, and then we get to the dream machine because observations and model making requires the intelligence 12:55
fennthe dream machine is just a way to get knowledge into a brain12:55
kanzureor at least some sort of genetic algorithm for model construction (where the scoring function is of the new phenomenon that we're dealing with, yadda yadda)12:55
fenna genetic algorithm for model construction might work ok for many situations12:56
fenni'd go so far as to call it intelligence12:57
fennwhy did you bring up bose einstein condensates?12:57
kanzuretrying to get matter as a wave12:58
fenni think it's still more complex than a simple EM wave12:58
fennrubidium is still rubidium when it comes out, not gamma rays or whatever12:59
kanzureI think the fundamental issues are solved though12:59
fenni think of it as something like a crystal in the aether12:59
kanzurethe solution was social aggregation, just like before13:00
kanzurejust like we determined before13:00
kanzureplus some way to do mutagenesis on brains or something so that we can do more pre-emptive modeling13:01
fenninstructables is 'mere' social aggregation13:01
kanzureor GAs for model construction13:01
fennwe're adding a formal system on top of that to allow computer manipulation of knowledge13:01
fenna grammar if you will13:01
kanzureprogramming is like manipulating all of those experiences that resulted in nuggets of knowledge and applying the 'gist' of the situation to a new ant mound13:01
fenneh, i think programming is really the reverse of that13:02
kanzuredebuggign?13:02
kanzure*debugging?13:02
fenntaking the generalized 'law' and turning it into discrete, specific cases13:02
kanzureso which one is the skdb package13:02
kanzureis it the generalization13:02
kanzureor is the case-package ?13:02
fennthe specific case13:03
fennwith varying levels of specificity13:03
kanzureok13:03
fennspoon, wooden spoon, silver spoon, utensil13:03
kanzurebut then why is it that I seem to be able to construct simple objects like that in my mind13:04
kanzurewithout having to rely on experimentation and testing13:04
kanzuremaking a new type of eating utencil, for example13:04
fennuh, who says you dont have to rely on experimentation and testing?13:04
kanzureevery four year old has their idea of an awesome new spork with extra wings or something13:04
fenngo try it, make a new eating utensil and see if it works13:04
kanzureI can see a way to fold a food-scooping mechanism with paper and tape13:05
fennthe difference between theory and practice is muh greater in practice13:05
fennthe concept of 'data point' is useful here13:07
kanzurenow how the hell do we write code for all of this13:07
kanzuredesign compiler, inventory (data point? :p) management, blargh13:07
fennmeh13:07
kanzureI have to go back and trace the ideas here and recombine them with the arguments I was presenting earlier13:07
kanzureoh13:08
kanzurea package of tools for the modeling of new experiences or materials or things13:08
fennno13:08
kanzurewhich would then be submitted back to an skdb aggregation system13:08
kanzureno?13:08
fennthat's AI13:08
kanzurenononono13:08
kanzureconsider bacteria13:08
kanzurethey can integrate new materials into their system13:08
kanzurelike randomly coming across an awesome metal or something13:09
kanzurewithout ai13:09
fennthey do it with evolution which is a form of intelligence13:09
kanzuredon't tell me that the Cherbnoyl reactor bacteria were intelligent ;-)13:09
kanzure(it might have been fungi, and fungi does weird shit, so maybe they /were/ intelligent)13:09
fennthey dont make any models of the metal and decide it might be a good idea13:09
kanzureright, and we can do evolution with software to some extent13:09
kanzurelike modeling and equations software and so on13:09
kanzureGAs too13:09
fennso feed the software and magically it pops out a set of formulas describing the system?13:10
kanzurenah13:10
kanzurewhat I'm thinking is that there are some tools and methods that we can use to model systems to some extent13:10
kanzureand then finding the right combination or sequence is the trick13:10
fennwhat's the context here?13:10
kanzurethat's difficult to say13:11
kanzureconsider a new chemical reaction ?13:11
kanzureor maybe consider a new thing that would be socially aggregated13:11
kanzureone of those buzzword situations, where something 'new' has been found13:11
fenna new thing13:11
kanzureand it needs to be 'crystalized'13:11
fennnot much semantic content13:11
kanzureso that other skdb users can go around using it13:11
kanzureright, but the semantics could be developed over time or something13:11
fennare you just talking about the process of creating a formal grammar for describing systems?13:12
fennbecause computers are notoriously bad at that13:12
kanzureno, I was thinking of the case where you're out "on the field" and you have a computer, a something 'new', and you have some tools, like a GA and some information on similar things or context-related packages or whatever, so now you're going to plug things together and try to constrain possibility space to get models or something that can be packaged for skdb13:12
kanzurethe intelligence could be used for constraints on the possibility space that would be explored13:13
fenna robot scientist13:13
kanzurebut otherwise, the system should be able to, say, automatically build that lab 13:13
kanzurehttp://expo.sf.net/ re robot scientists, who we need to go talk with eventually13:13
fenni dont think it's part of the fundamental 'what makes skdb skdb'13:13
kanzurethey have university funding so they might be another case of reprap, but they are of the bio community and so might be more friendly (if not territorial like biobarcamp)13:13
fennand i stick by my assertion that it's AI13:13
kanzureit's definitely not a fundamental component13:13
kanzurehowever13:13
kanzureyou pointed out 'just' a while back13:14
kanzurewhen I mentioned that it was social aggregation13:14
kanzureand so I was providing a few ideas on making the social aggregation actually happen13:14
kanzuremaking some tools that allow new things to be thrown into the database would be a good idea methinsk13:14
kanzure*methinks13:14
fenni find myself clinging to the cathedral style of development13:17
fennnot wanting a bunch of machine generated junk dropped into the system13:17
fennbut humans will do that anyway13:17
kanzuredebian has quality assessment teams13:17
kanzurethat guard the gateways13:17
kanzureI thought we were going to do something like that anyway ?13:18
fennyeah but it's not all designed to work together13:18
kanzureit harms nobody since they git-clone the whole thing anyway13:18
kanzurehm?13:18
kanzureoh13:18
kanzurefrom the ground up13:18
fennit just happens to work after some massaging and cooddling13:18
kanzurerefactoring sucks13:18
* kanzure suspects it's why firefox still sucks13:18
kanzurethe code base I mean13:19
kanzureeverything else about it is pretty awesome13:19
fennfirefox is developed as a cathedral.. i'm not following13:19
kanzureoh13:19
kanzurewell, I was referring to the nasty codebase and how it would be a good idea to refactor it13:20
kanzureso that it doesn't have the issues that I've been pointing out for a while now (re: my tab addiction problem)13:20
kanzureand I mentioned that refactoring sucks13:20
kanzurenot so much the cathedral distinction13:20
fennbut nobody else has that problem so it isnt aproblem13:20
kanzuresimply that refactoring a large code base because something was wrong in the beginning, really, really sucks13:20
kanzurebah13:20
fenndammit the wireless works fine when i'm the only person awake13:20
kanzurethat's like saying just because you're sick but nobody else is, means that you're therefore not sick13:21
kanzureI think I once theoretically solved the refactoring problem13:21
kanzureI don't remember how.13:22
kanzureI wish I had some notes on that.13:22
fennif the system as a whole does what it was designed to do, then it doesnt matter if i'm upset that there's a bunch of crap floating around the skdb intarweb13:22
kanzureI was complaining about it a lot one or two months ago, and I mentioned to somebody my plans about it13:22
kanzureit might have involved my brain plans13:22
kanzuresure13:22
fennsounds like a bad plan then13:22
kanzurefenn: but the data that you /do/ receive might have to be refactored anyway13:22
kanzureespecially if the people that are submitting it do it their own way13:22
kanzuresee, it's still cathedral 13:23
kanzuresince they have to agree on the standards and whatever13:23
fennthat's a lot of work13:23
kanzureof course, automatic filtering of noncompliant package13:23
fennyou cant just turn everything away because it's not your way13:23
kanzureright :(13:23
fennsee how far that got stallman :P13:23
kanzureor can I?13:23
kanzureget lost13:23
kanzureheh13:23
kanzurewell, he's not dead yet13:23
kanzure99 percent, remember?13:24
kanzure:P I kid13:24
fennyeah maybe he'll finish hurd by the time he's dead13:24
fennor maybe he'll take some acid and realize he's not interested in software anymore13:24
fenndo you believe in self determination?13:25
kanzurewhat the hell does that mean13:25
kanzure:)13:26
fennare we just particles banging around probabilistically, all of our desires simply a result of chance13:26
fennno blame, no credit13:26
kanzurethis is a complicated issue13:26
kanzureand I thought we went over it before13:26
kanzurebrb13:26
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has quit [Read error: 110 (Connection timed out)]14:18
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has joined #hplusroadmap14:21
kanzurefenn: so, to answer your question15:37
kanzure"I am physics"15:38
kanzureperhaps not all of it15:38
kanzureand really only a quite limited subset of the possible combinations of phenomena15:38
kanzureand I don't understand probability/statistics. Or at least people's reliance on it. It's an easy way to 'get around' some problems. It's not so much getting around as ignoring ...15:42
-!- Splicer [n=p@h77n3c1o261.bredband.skanova.com] has joined #hplusroadmap15:51
kanzureHey Splicer.15:52
kanzurehm15:52
Splicerhi 15:52
kanzureSo if I go do laundry, I get to go order that new computer15:53
kanzurehm15:53
kanzureWhat a deal.15:53
Splicera man can never have too many computers15:54
kanzureAnd in the background I'm backing up all of my data. I think the total is coming out to close to 300 GB. 15:54
kanzureIt's interesting to see cp in the background there and all of the old, old folders and packages of stuffs that I've made.15:54
kanzurelike what the hell is /mnt/edmini/books/2008-06-21_server/home/bbishop/new_logdir2/cyanopicapica/15:55
kanzureoop15:55
kanzure2008-02-21_server15:55
kanzureAnyway, with any luck my cache will be web accessible soon.15:56
kanzure452 days remaining. Geesh. And I'm being kicked out, I think.16:22
nshhrm?16:23
kanzurensh: I'm not sure.16:39
kanzureBut I was told that it would be a "good idea". And that arrangements have already been made.16:40
* nsh frowns16:45
Splicer... it's alive16:48
kanzureHm.16:56
kanzureOh well. I think it might be alright. There might be some free food out of it.16:57
kanzureAnd faster pipes.16:57
nshpipes are always a bonus17:02
kanzurehttp://heybryan.org/buildingbrains.html17:13
kanzureOops, refresh.17:16
kanzureI seem to be on 11 day cycles.17:24
kanzure(when it comes to documents)17:24
nshhrm17:51
nshthis channel just went back in time by half an hour17:52
nshah, no. t'was ##neuroscience17:52
kanzurehaha :)17:52
-!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has joined #hplusroadmap18:01
kanzureHey willPow3r.18:04
kanzureLemme upload today's logs.18:04
willPow3rcool18:04
kanzurehttp://heybryan.org/chats/2008-06-24_hplusroadmap.html18:05
kanzurehttp://linuxpmi.org/trac/wiki/OpenMosixPatchCommentaries commentary on the linux kernel18:42
kanzuregot this from #openmosix18:42
kanzurean impressive woman wrote the page18:42
willPow3ri didn't think there was such thing as an "impressive woman"18:51
nshthat might be a self-fulfilling assessment18:52
willPow3rif you tend to disagree, you should come on out to san diego and check out the STI incubators that attend vernor vinge's SDSU18:54
* nsh smiles18:57
willPow3ri'm biased in my view. please forgive me for my misogyny.18:57
nshit don't confront me; long as i get my money next friday18:58
-!- Phreedom [n=freedom@www.online.dn.ua] has quit [Remote closed the connection]19:01
kanzurensh: Not self-fulfilling really, but by 'impressive' I mean to say she's not transexual and i.e., really a guy19:11
-!- nsh- [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has joined #hplusroadmap19:17
kanzurensh: Not self-fulfilling really, but by 'impressive' I mean to say she's not transexual and i.e., really a guy19:18
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has quit [Connection timed out]19:30
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has joined #hplusroadmap19:31
-!- nsh- [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has quit [Read error: 110 (Connection timed out)]19:36
kanzureUhm. I lost a train of thought.19:40
kanzureWhat am I doing?19:40
-!- fenn_ [n=pz@adsl-76-251-85-53.dsl.bltnin.sbcglobal.net] has joined #hplusroadmap19:56
-!- Topic for #hplusroadmap: Intro: http://www.youtube.com/watch?v=ZXKbzbeipmI http://diybio.org/ http://openwetware.org/ | diy bio toolkit: http://biohack.sf.net/ | Automated societal knowledge (put it to work): http://heybryan.org/exp.html | Channel wiki: http://heybryan.org/mediawiki/ | F/OSS perspectives on Kurzweil: http://heybryan.org/fernhout/19:56
-!- Topic set by kanzure [] [Sun Jun 8 13:00:15 2008]19:56
[Users #hplusroadmap]19:56
[ fenn ] [ freer ] [ percent_] [ Splicer ] [ wrldpc] 19:56
[ fenn_] [ kanzure] [ procto ] [ willPow3r] [ ybit ] 19:56
-!- Irssi: #hplusroadmap: Total of 10 nicks [0 ops, 0 halfops, 0 voices, 10 normal]19:56
-!- Channel #hplusroadmap created Sat Mar 22 15:44:12 200819:56
-!- Irssi: Join to #hplusroadmap was synced in 37 secs19:57
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has joined #hplusroadmap19:59
willPow3rtaking your adhd medication?20:08
kanzurewillPow3r: Hm?20:08
kanzureYes, today at least.20:09
willPow3rheh ;)20:09
kanzureThe reason my output has decreased is somewhat because I've been doing only half my dosage recently since I've been at the lab and not at home, but I'm fixing that starting tomorrow. It's too stupid losing so many nights just because I can't reach a cabinet fast enough.20:09
-!- fenn [n=pz@adsl-76-251-82-18.dsl.bltnin.sbcglobal.net] has quit [Read error: 110 (Connection timed out)]20:09
* kanzure wasn't at the lab today.20:09
willPow3rwhat are you taking?20:10
kanzureAdderall.20:10
kanzureAnd it /works/. Off of the medication, one minute I can be as I am now, the next minute I find myself playing in the grass outside.20:10
willPow3rinattentive type20:10
willPow3ri used to take adderall, but this new med came out20:11
willPow3rcalled vyvanse20:11
kanzureYeah, I tried that too.20:11
willPow3rits a prodrug, lisdexamphetaime20:11
willPow3ri really like it20:11
kanzurepatch form.20:11
kanzuremarketed towards kids20:11
willPow3ryes, thats usually where adhd medication is aimed when it comes to market20:12
kanzureI went in to the doc's office in late 2007 and the guy basically told me that I had Hitler syndrome, so he took me off Adderall for a week and put me on vyvanse. Blargh. Didn't go well.20:12
kanzureI didn't want to kill jews, so I don't think I had Hitler syndrome ;-)20:12
kanzureI'm actually more the H in ADHD.20:13
willPow3ryou were going the way of pinky and the brain then?20:13
kanzureNo, there's some other things going on, other than just ADHD.20:13
kanzureI'm suspecting some Aspergers.20:13
kanzurehttp://heybryan.org/intense_world_syndrome.html20:13
willPow3rclass in 15. bbl20:13
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has quit [Read error: 110 (Connection timed out)]20:25
kanzurewillPow3r: Just say when you're back.20:37
kanzurehttp://en.wikipedia.org/wiki/Intensive_Care_Unit20:48
kanzureI want one.20:48
kanzurehttp://en.wikipedia.org/wiki/DO-178B ' standard methods for producing life-critical avionics software'20:54
wrldpcwhat was that?21:22
-!- Netsplit niven.freenode.net <-> irc.freenode.net quits: percent_, freer21:25
kanzureHm?21:25
kanzurewrldpc: What?21:25
-!- Netsplit over, joins: percent_, freer21:25
wrldpcdid you just receive the system-wide message?21:26
wrldpcI glanced at it.21:26
-!- Netsplit niven.freenode.net <-> irc.freenode.net quits: percent_, freer21:26
-!- Netsplit over, joins: percent_, freer21:27
-!- Netsplit niven.freenode.net <-> irc.freenode.net quits: willPow3r21:28
-!- Netsplit over, joins: willPow3r21:28
kanzurewrldpc: No, I didn't.21:29
-!- Netsplit niven.freenode.net <-> irc.freenode.net quits: willPow3r21:30
-!- Netsplit over, joins: willPow3r21:30
-!- Netsplit niven.freenode.net <-> irc.freenode.net quits: wrldpc21:38
-!- Netsplit over, joins: wrldpc21:39
-!- Netsplit niven.freenode.net <-> irc.freenode.net quits: kanzure21:41
-!- Netsplit over, joins: kanzure21:41
-!- Vedestin [n=Vedestin@d58-111-95-212.sbr3.nsw.optusnet.com.au] has joined #hplusroadmap23:33

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