--- Log opened Sat Aug 19 00:00:53 2017
00:17 -!- poppingtonic [~brian@unaffiliated/poppingtonic] has quit [Ping timeout: 240 seconds]
00:57 -!- y0no [y0no@2001:bc8:212d:201:ff01::a] has joined ##hplusroadmap
01:14 < fltrz> finally got mysql on the external drive sub folder: completely remove mysql installation, mount -bind the folder to /var/lib/mysql, then reinstall mysql to hypnotrick mysql into believing everything is still retentive in /var/lib/mysql..
01:28 -!- Guest84243 [~HEx@hexwab.plus.com] has quit [Remote host closed the connection]
01:34 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Remote host closed the connection]
01:34 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has joined ##hplusroadmap
01:38 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Ping timeout: 240 seconds]
01:57 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-ikfshneustwqacxm] has joined ##hplusroadmap
01:58 -!- preview [~quassel@2407:7000:842d:4000::3] has quit [Read error: No route to host]
02:09 -!- y0ur1 [~y0ur1@gateway/tor-sasl/y0ur1] has quit [Remote host closed the connection]
02:10 -!- augur [~augur@198-27-215-123.static.sonic.net] has joined ##hplusroadmap
02:12 -!- y0ur1 [~y0ur1@gateway/tor-sasl/y0ur1] has joined ##hplusroadmap
02:14 -!- augur [~augur@198-27-215-123.static.sonic.net] has quit [Ping timeout: 240 seconds]
02:21 < fltrz> kanzure: that rant is very accurate imho
02:24 < fltrz> and although its evidently clear that cellular dynamics are not at all like central processing unit running single threaded software, I still think the analogy with analog electronics is high (under the condition that one recognizes a nonstraightforward analogy)
02:26 < fltrz> in analog electronics conductors are in most conditions on the same potential throughout the conductor, and the conductors connect different "components"
02:28 < fltrz> in a vacuole chemical species are in most conditions homogenously diffused throughout the vacuole, and we could pretend a specific species concentration in vacuole is like a specific species potential in a circuit)
02:29 < fltrz> instead of electrical n-terminal components we have reactions connecting the chemical species
02:30 < fltrz> so instead of an instruction set documentation and a debugger, we need something like a million channel oscilloscope (but the bandwidth doesn't need to be Ghz)
02:31 < fltrz> so functionally its like an oscilloscope, but really something that non-invasively tracks the concentration of all the species in the vacuole
02:31 < fltrz> instead of a voltage on a conductor
02:34 < fltrz> cells don't need conductors to keep signals separate, by virtue of only reacting with in the reactions (=components in circuit analogy) they play a role in, they are free to mix in the vacuole
02:35 < fltrz> and it's true analog electronics is much harder than blinking lights on arduino
02:35 < fltrz> and that's analog electronics where we have datasheets for each component
02:37 < fltrz> thats why I don't understand all the hype around finding the crystal structure of a protein, the interesting information is all in the reaction kinetics, the rate constants etc
02:38 < fltrz> thats like being obsessed about the exact path of a trace on a pcb, which is relatively arbitrary compared to which components it connects
02:39 < fltrz> if we have even moderately accurate reaction kinetics for all reactions in a vacuole, we can start modeling vacuole behaviours today
02:40 < fltrz> and make predictions about how things will react to external conditions and compare them with experiment
02:48 < fltrz> once we have the reaction kinetics of all reactions (for a specic cell say, more on that later) we can discover deterministically verifiable cell types (the stable fixed points of the reaction equations)
02:49 < fltrz> now that computation will initially be very large, and gradient descent might often result in the simulated cell suiciding itself (to protect the rest of the organism from cancer)
02:50 < fltrz> but once we have the correct cell type we can easily discover new cell types by watching what happens at cell division, consider cell types A -> B + C
02:51 < fltrz> assuming we have found the correct concentrations for cell type A, gradient descent will GIVE the concentrations for B and C, we no longer need to brute force find stable fixed point to find B and C, only follow A
02:53 < fltrz> so naturally once we have reaction equations, we just need to measure the concentrations in the fertilized egg cell type (or one of its close descendants) and we can explore the tree downwards
02:54 < fltrz> the leaves will be final cell types which eventually suicide itself or remain stable
02:55 < fltrz> now it could happen that our initial brute force found (stable fixed point) cell type was not in fact the fertilized egg cell but one of its descendants
02:56 < fltrz> so it can happen that we have to brute force it a couple of times (each time finding a new subtree in cell type lineage) but at some point we will discover the simulated sperm cell type, and also at some point the simulated egg cell cel type
02:57 < fltrz> by joining their contents (=concentrations) we will have a true fertilized egg cell and from there on its gradient descent again, and we can find ALL cell types
03:00 < fltrz> of course "once we have all the reaction kinetics" is a hard bitch, and it's not as romantic as "decrypting the secret code of life" but instead involves someone taking a couple of the millions of species (a protein/molecule/ion/...), mixing them and measuring their reaction kinetics in vitro
03:02 < fltrz> there seems little prestige in performing an important but miniscule task, and thats why it's not happening as quick as it should
03:05 < fltrz> sorry for the rant :) I identify very much with the one you posted
04:17 -!- Urchin [~urchin@unaffiliated/urchin] has joined ##hplusroadmap
04:21 -!- Urchin [~urchin@unaffiliated/urchin] has quit [Remote host closed the connection]
04:44 < kanzure> protein structure matters because it tells you what the protein is doing, like the dna + polymerase structures that have been uploaded to pdb.
04:49 < fltrz> right what it is doing (i.e. a reaction), it tells us which species interact and result, but thats only the connectivity of conductors (i.e. proteins) to a specific component (i.e. reaction), nothing about its reaction rates
04:50 < fltrz> to find a stable fixed point, for each species the production rate equals the consumption rate (in a stable way, so no runaway unstable fixed points)
04:51 < fltrz> so yes knowing that there is a reaction and what its result is A + B -> C + D is better than knowing only A, B, C, D exist, but it still doesn't give us the rate
04:53 < fltrz> i.e. we need either way more biologists or way more automation under control of these biologists
04:53 -!- Gurkenglas_ [~Gurkengla@dslb-178-000-218-087.178.000.pools.vodafone-ip.de] has joined ##hplusroadmap
05:08 < kanzure> why do you need the rate? what?
05:10 < adlai> wonderful analogy from a dude that just sat in on the first kinetics class before vanishing for the rest of the semester: "kinetics is the gas pedal, and thermodynamics is the brakes"
05:10 < kanzure> look at the polymerase structure; if we could hack the shit out of it, we could do many useful things. but the actual rate is totally irrelevant.
05:10 < adlai> ie, understanding the kinetics of various processes tells you what actually happens (as opposed to thermodynamics, which tells you where it'll theoretically end up someday)
05:11 < kanzure> http://diyhpl.us/~bryan/papers2/bio/protein-engineering/?C=M;O=D
05:20 < adlai> how a polypeptide folds into the final structure is a question of kinetics. you could easily end up in global non-minimum, thermodynamically speaking, but it's reached due to kinetic effects that lead you into a ~local~ minimum
05:20 < adlai> kinetics also leads to things like binding affinities
05:22 < fltrz> yes
05:23 < kanzure> i don't think it would matter how it folds, only that it reliably produces the folds you require.
05:24 < adlai> "reliably produces" = kinetics
05:25 < fltrz> kanzure: I may have mistakenly left the impression that I think structure is entirely irrelevant, that would be wrong, of course it is still important, what I mean is more like, to understand spontaneous cancers not due to DNA damage, the structures are not that interesting, we need to understand the cell types
05:25 < fltrz> and for each cell type the transition probability to other cell types, and the conditions that affect it
05:25 < kanzure> cell structure is function, tho.
05:26 < adlai> "cellular kinetics", i wonder whether that's an official field of study yet
05:26 < kanzure> well neurophysiology is..
05:26 < adlai> looks like it has been mentioned in 'the literature', but hasn't qualified for a wikipedia page yet
06:13 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap
06:23 -!- Gurkenglas_ [~Gurkengla@dslb-178-000-218-087.178.000.pools.vodafone-ip.de] has quit [Ping timeout: 248 seconds]
06:48 -!- jaboja [~jaboja@jaboja.pl] has quit [Ping timeout: 248 seconds]
06:54 < kanzure> hmm randal koene is now employed by kernel.co
07:11 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap
07:24 -!- ybit [~ybit@unaffiliated/ybit] has joined ##hplusroadmap
07:24 < ybit> http://ivory.idyll.org/blog/2017-pof-software-archivability.html
07:28 < kanzure> i believe i've spoken with the author, titus brown, as consequence of http://diyhpl.us/~bryan/irc/2017-02-03-beacon.pdf
07:30 -!- jaboja [~jaboja@jaboja.pl] has quit [Ping timeout: 255 seconds]
07:32 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap
07:35 -!- preview [~quassel@2407:7000:842d:4000::3] has joined ##hplusroadmap
07:53 -!- preview [~quassel@2407:7000:842d:4000::3] has quit [Remote host closed the connection]
07:59 -!- preview [~quassel@2407:7000:842d:4000::3] has joined ##hplusroadmap
08:00 -!- jaboja [~jaboja@jaboja.pl] has quit [Ping timeout: 255 seconds]
08:47 -!- yashgaroth [~yashgarot@2606:6000:cd4d:3300:f5e0:f867:a11d:8d52] has joined ##hplusroadmap
08:53 < kanzure> yashgaroth: https://twitter.com/QuantumG/status/43472344756387840
08:55 < yashgaroth> I...yeah, p much
09:26 -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-ikfshneustwqacxm] has quit [Quit: Connection closed for inactivity]
09:38 -!- Storyteller [~Storytell@unaffiliated/storyteller] has joined ##hplusroadmap
09:39 -!- Gurkenglas_ [~Gurkengla@dslb-178-000-218-087.178.000.pools.vodafone-ip.de] has joined ##hplusroadmap
09:41 < kanzure> "DeepRebirth: Accelerating deep neural network execution on mobile devices" https://arxiv.org/abs/1708.04728
09:50 -!- Burninate [~Burn@pool-96-241-130-178.washdc.fios.verizon.net] has quit [Quit: Leaving]
09:50 < kanzure> https://www.reddit.com/r/Futurology/comments/6up5qi/i_am_dr_roman_yampolskiy_author_of_artificial/
09:51 < kanzure> "When will ai exceed human performance? Evidence from ai experts" https://arxiv.org/pdf/1705.08807.pdf
09:54 < kanzure> roving epistemology gangs spotted https://streetepistemology.com/resources/
10:01 < kanzure> "SMASH: One-shot model architecture search through hypernetworks" https://arxiv.org/abs/1708.05344 https://twitter.com/hardmaru/status/898589592873336832 https://www.reddit.com/r/reinforcementlearning/comments/6uioms/smash_oneshot_model_architecture_search_through/
10:02 < Storyteller> https://petitions.whitehouse.gov/petition/buy-baja-peninsula-instead-building-wall
10:02 -!- mode/##hplusroadmap [+o kanzure] by ChanServ
10:02 -!- Storyteller was kicked from ##hplusroadmap by kanzure [Storyteller]
10:02 -!- mode/##hplusroadmap [-o kanzure] by kanzure
10:14 < kanzure> http://approximatelycorrect.com/2017/08/14/death-note-finally-an-anime-about-deep-machine-learning/
10:56 -!- Gurkenglas_ [~Gurkengla@dslb-178-000-218-087.178.000.pools.vodafone-ip.de] has quit [Ping timeout: 255 seconds]
11:02 < kanzure> "Computational principles of synaptic memory consolidation" http://www.nature.com/neuro/journal/v19/n12/full/nn.4401.html (2016)
11:08 -!- Burninate [~Burn@pool-96-241-130-178.washdc.fios.verizon.net] has joined ##hplusroadmap
12:10 < fenn> 'I’d recommend Psycho Pass to anyone who’s into anime & deep learning. The whole theme of Psycho Pass is basically about a huge classifier (“Sibyl System”) which spots criminals from the ordinaries.'
12:10 < fenn> the style of psycho pass reminded me of ghost in the shell SAC
12:23 -!- y0ur1 [~y0ur1@gateway/tor-sasl/y0ur1] has quit [Quit: y0ur1]
12:44 < fenn> "data is going to become useless because the software needed to interpret it is vanishing. (We're already seeing this with 454 sequence data, which is less than 10 years old; very few modern bioinformatics tools will ingest it, but we have an awful lot of it in the archives.)"
12:44 < fenn> wut
12:46 < fenn> "454 data does not provide individual base measurements from which basecalls can be derived. Instead, it provides measurements that estimate the length of the next homopolymer stretch in the sequence (i.e., in "AAATGG", "AAA" is a 3-mer stretch of A's, "T" is a 1-mer stretch of T's and "GG" is a 2-mer stretch of G's). A basecalled sequence is then derived by converting each estimate into a
12:47 < fenn> homopolymer stretch of that length and concatenating the homopolymers."
12:47 < fenn> this does not sound that hard to do
13:40 < kanzure> the STORM paper has 4013 citing articles.
13:40 < fenn> it's pretty sweet
13:41 < kanzure> fig1 "STORM imaging reveals distinct organization of actin filaments in the axons and dendrites of neurons" https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3815867/
13:42 < kanzure> "The 2015 super-resolution microscopy roadmap" http://www.rug.nl/research/portal/files/39740478/Hell_2015_J._Phys._D_Appl._Phys._48_443001.pdf
13:45 < fenn> really wants an in-browser 3d volumetric flyaround viewer
13:45 < fenn> storm data is just a bunch of points
13:46 < fenn> you'd have to use completely different compression algorithms than images use
14:03 < adlai> what is this vanishing software? have these people not heard of massively distributed backups aka the bitcoin utxo set?
14:05 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has joined ##hplusroadmap
14:10 < kanzure> that's not how utxos work
14:22 -!- Malvolio [~Malvolio@unaffiliated/malvolio] has quit [Ping timeout: 240 seconds]
15:03 -!- HEx1 [~HEx@hexwab.plus.com] has joined ##hplusroadmap
15:08 -!- jaboja [~jaboja@jaboja.pl] has joined ##hplusroadmap
15:36 < kanzure> "A buddy of mine was working in the lab of the scientist who "invented" unnatural amino acid technology (rumblings about how it was taken from the postdoc in the lab across the hall). That was the sexy project de jour of the lab; but my buddy was working on a less sexy project - catalytic antibodies. This is where you take the concept that antibodies can be adapted to fit any molecular shape ...
15:36 < kanzure> ...and try to make it fit something that looks like the transition state of a chemical reaction, resulting in acceleration of the chemical reaction. (designer enzymes!) .... finally his boss modified his project, instructing him to graft a metalloprotease domain, onto the antibody in an attempt to get it working. A breath of fresh air! Suddenly he was invigorated with a new approach to the ...
15:36 < kanzure> ...project. But not long after that, he was back to the old routine of being burned out ... In the end, it never worked. It turned out that the metalloprotease domain was designed by Homme Hellinga. Years after this, the scientific community discovered that Homme Hellinga was faking his enzyme design work." https://news.ycombinator.com/item?id=15055058
16:18 < kanzure> jrayhawk: here are some comments about migrating debian to nix packages https://news.ycombinator.com/item?id=15011743
16:19 -!- jrayhawk [~jrayhawk@unaffiliated/jrayhawk] has joined ##hplusroadmap
16:26 < kanzure> 16:18 < kanzure> jrayhawk: here are some comments about migrating debian to nix packages https://news.ycombinator.com/item?id=15011743
16:30 < jrayhawk> 'Nix doesn't integrate with Debian well - you can also use yum on Debian and nobody does that either. If you're using Debian, it's because you want to manage a Debian system with Debian packages.' wow, one lone warrior is still fighting the deb/rpm war of 20 years ago
16:32 < jrayhawk> in reality people use debian because it's got more packages across more architectures with stronger policy guarantees and more manpower than any other distribution; i don't think anybody's particularly enamoured with the packaging tools these days except insofar as lintian forces policy compliance
16:34 < jrayhawk> seeing joey hess go nuts and mash nix support into debhelper 8 would not be particularly surprising
16:36 -!- preview [~quassel@2407:7000:842d:4000::3] has quit [Read error: Connection reset by peer]
16:42 -!- aeiousomething [~aeiousome@183.82.170.54] has joined ##hplusroadmap
16:46 -!- preview [~quassel@2407:7000:842d:4000::3] has joined ##hplusroadmap
17:09 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Remote host closed the connection]
17:09 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has joined ##hplusroadmap
17:11 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Remote host closed the connection]
17:11 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has joined ##hplusroadmap
17:13 -!- preview [~quassel@2407:7000:842d:4000::3] has quit [Read error: Connection reset by peer]
17:14 -!- preview [~quassel@2407:7000:842d:4000::3] has joined ##hplusroadmap
17:40 < kanzure> "Magneto-thermal genetic deep brain stimulation of motor behaviors in awake, freely moving mice" https://elifesciences.org/articles/27069
17:40 < kanzure> "The approach uses alternating magnetic fields to heat superparamagnetic nanoparticles on the neuronal membrane. Neurons heat-sensitized by expressing TRPV1 are activated with magnetic field application. Magnetothermal genetic stimulation in the motor cortex evoked ambulation, deep brain stimulation in the striatum caused rotation around the body-axis, and stimulation near the ridge between ...
17:40 < kanzure> ...ventral and dorsal striatum caused freezing-of-gait. The duration of the behavior correlated tightly with field application. This approach provides genetically and spatially targetable, repeatable and temporarily precise activation of deep-brain circuits without need for surgical implantation of any device."
17:50 < kanzure> ".. remote neural excitation through the activation of the heat-sensitive capsaicin receptor TRPV1 by magnetic nanoparticles. When exposed to alternating magnetic fields, the nanoparticles dissipate heat generated by hysteresis, triggering widespread and reversible firing of TRPV1+ neurons"
17:50 -!- Malvolio [~Malvolio@unaffiliated/malvolio] has joined ##hplusroadmap
17:53 < kanzure> "Engineering genetically-encoded mineralization and magnetism via directed evolution" https://www.nature.com/articles/srep38019 (216)
17:57 < kanzure> "MagR alone is insufficient to confer cellular calcium responses to magnetic stimulation" https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5352684/
17:58 < kanzure> "A proposal for an electron-transfer mechanism of avian magnetoreception" https://arxiv.org/pdf/1705.00735.pdf
17:58 < kanzure> cc nmz787
18:07 < kanzure> "CRISPRcon .. the University of California, Berkeley campus, kicked off yesterday, August 16, 2017"
18:07 < kanzure> http://crisprcon.org/
18:08 < kanzure> low signal quality. nevermind.
18:22 -!- maaku [~mark@173.234.25.100] has joined ##hplusroadmap
18:22 < maaku> pasky: you done any work with word2vec?
18:37 < aeiousomething> hi maaku
18:43 < maaku> aeiousomething: hello
18:44 < aeiousomething> maaku i've managed to get my hands on some modafinil for my ADD finally in India
18:44 < aeiousomething> the comedown is horrible.
18:45 < kanzure> the what now?
18:45 < kanzure> "comedown" sounds made up.
18:56 < fltrz> are those input consoles (and some display) on a sattelite exterior? https://upload.wikimedia.org/wikipedia/commons/d/d4/LDEF_over_payload_bay.jpg
19:00 < fenn> materials being exposed to space environment
19:01 < aeiousomething> kanzure: ... when the drug wears off towards the end of the day
19:01 < aeiousomething> I have massive headaches
19:02 < aeiousomething> and severe miscoordination
19:09 < fltrz> huh weird, the caption for the image on wikipedia states this is STS-41 which according to wikipedia launched Ulysses probe, but if I look at the part of the Ulysses probe and its propulsion systems, none look like the object with a terminal in the picture above...
19:11 -!- aeiousomething [~aeiousome@183.82.170.54] has quit [Read error: Connection reset by peer]
19:11 -!- aeiousomething [~aeiousome@183.82.170.54] has joined ##hplusroadmap
19:12 < fltrz> ah its the retrieval of LDEF and its STS-32
19:12 < fltrz> so the caption was wrong
19:14 < fltrz> ah no I am wrong too, its the deployment of LDEF but STS-41C
19:28 -!- aeiousom1thing [~aeiousome@117.195.129.174] has joined ##hplusroadmap
19:29 -!- aeiousomething [~aeiousome@183.82.170.54] has quit [Ping timeout: 240 seconds]
19:29 -!- aeiousom1thing [~aeiousome@117.195.129.174] has left ##hplusroadmap []
19:33 -!- Storyteller [~Storytell@unaffiliated/storyteller] has joined ##hplusroadmap
20:32 < maaku> .tell aeiousomething does it at least work though?
20:32 < yoleaux> maaku: I'll pass your message to aeiousomething.
20:33 < maaku> .tell aeiousomething context: the modafinil
20:33 < yoleaux> maaku: I'll pass your message to aeiousomething.
20:34 < maaku> I hope I'm not the only one who gets massively nostalgic everytime I see STS images
20:34 < maaku> amazing course lecture on designing the space shuttle : https://ocw.mit.edu/courses/aeronautics-and-astronautics/16-885j-aircraft-systems-engineering-fall-2005/index.htm
20:35 < maaku> Aircraft Systems Engineering taught by Jeffrey Hoffman (probably the most famous STS astronaut)
21:01 -!- Storyteller [~Storytell@unaffiliated/storyteller] has quit [Ping timeout: 255 seconds]
21:06 -!- berndj [~berndj@mail.azna.co.za] has quit [Quit: ZNC - http://znc.in]
21:06 -!- berndj [~berndj@mail.azna.co.za] has joined ##hplusroadmap
21:22 -!- jaboja [~jaboja@jaboja.pl] has quit [Remote host closed the connection]
21:26 < fltrz> lol "I don't understand how people manage to be in an office with their boring-ass jobs from 9-5 Monday through Friday till eternity (excluding their measly 2-3 weeks of vacation). I honestly think most people lack consciousness."
22:23 -!- TC [~talinck@cpe-174-97-113-184.cinci.res.rr.com] has joined ##hplusroadmap
22:23 -!- TC is now known as Guest31807
22:27 -!- hehelleshin [~talinck@cpe-174-97-113-184.cinci.res.rr.com] has quit [Ping timeout: 240 seconds]
22:58 -!- yashgaroth [~yashgarot@2606:6000:cd4d:3300:f5e0:f867:a11d:8d52] has quit [Quit: Leaving]
23:27 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Remote host closed the connection]
23:27 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has joined ##hplusroadmap
23:31 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Ping timeout: 240 seconds]
--- Log closed Sun Aug 20 00:00:54 2017