--- Log opened Tue Jul 26 00:00:07 2016
00:05 -!- fleshtheworld [~fleshthew@2602:306:cf0f:4c20:14f:b32d:cdbc:7ce1] has quit [Read error: Connection reset by peer]
00:08 -!- augur [~augur@c-73-202-137-143.hsd1.ca.comcast.net] has joined ##hplusroadmap
00:50 -!- nildicit [~nildicit@unaffiliated/nildicit] has quit [Ping timeout: 260 seconds]
00:54 -!- esmerelda [~andares@unaffiliated/jacco] has joined ##hplusroadmap
00:56 -!- augur [~augur@c-73-202-137-143.hsd1.ca.comcast.net] has quit [Read error: Connection reset by peer]
00:56 -!- augur [~augur@c-73-202-137-143.hsd1.ca.comcast.net] has joined ##hplusroadmap
01:20 -!- nildicit [~nildicit@unaffiliated/nildicit] has joined ##hplusroadmap
01:36 -!- Aurelius_Home [~cpopell@c-73-200-185-48.hsd1.va.comcast.net] has quit [Read error: Connection reset by peer]
01:43 -!- jtimon [~quassel@55.31.134.37.dynamic.jazztel.es] has joined ##hplusroadmap
01:47 -!- Aurelius_Home [~cpopell@c-73-200-185-48.hsd1.va.comcast.net] has joined ##hplusroadmap
01:58 -!- zeroach [~CharlieNo@24-241-64-111.dhcp.stls.mo.charter.com] has quit [Read error: Connection reset by peer]
02:29 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
03:05 < archels> "A Bayesian is one who, vaguely expecting a horse, and catching a glimpse of a donkey, strongly believes he has seen a mule."
03:12 < ebowden_> lol
03:27 -!- jtimon [~quassel@55.31.134.37.dynamic.jazztel.es] has quit [Ping timeout: 252 seconds]
03:30 -!- SloanOnLinux [~sloan@162.72.132.128] has quit [Read error: Connection reset by peer]
03:31 -!- SloanOnLinux [~sloan@162.72.132.128] has joined ##hplusroadmap
03:32 -!- drewbot [~cinch@ec2-54-158-9-16.compute-1.amazonaws.com] has quit [Remote host closed the connection]
03:32 -!- drewbot [~cinch@ec2-174-129-73-19.compute-1.amazonaws.com] has joined ##hplusroadmap
03:43 -!- Orpheon [~Orpheon@213.200.193.129] has joined ##hplusroadmap
04:05 -!- Orpheon [~Orpheon@213.200.193.129] has quit [Read error: Connection reset by peer]
04:07 -!- Orpheon [~Orpheon@213.200.193.129] has joined ##hplusroadmap
04:13 -!- augur [~augur@c-73-202-137-143.hsd1.ca.comcast.net] has quit [Quit: Leaving...]
04:26 -!- Gurkenglas_ [Gurkenglas@dslb-178-005-223-183.178.005.pools.vodafone-ip.de] has joined ##hplusroadmap
04:48 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
04:51 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
04:53 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
04:53 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
04:57 -!- jtimon [~quassel@55.31.134.37.dynamic.jazztel.es] has joined ##hplusroadmap
05:03 -!- Gurkenglas_ [Gurkenglas@dslb-178-005-223-183.178.005.pools.vodafone-ip.de] has quit [Ping timeout: 244 seconds]
05:07 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
05:09 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
05:11 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
05:12 -!- ebowden_ [~ebowden@2001:8003:100e:c500:1cfd:f1aa:440c:383c] has joined ##hplusroadmap
05:12 -!- ebowden_ [~ebowden@2001:8003:100e:c500:1cfd:f1aa:440c:383c] has quit [Changing host]
05:12 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
05:12 -!- c0rw1n [~c0rw1n@193.47-244-81.adsl-dyn.isp.belgacom.be] has quit [Read error: Connection reset by peer]
05:12 -!- c0rw1n [~c0rw1n@193.47-244-81.adsl-dyn.isp.belgacom.be] has joined ##hplusroadmap
05:13 -!- c0rw1n [~c0rw1n@193.47-244-81.adsl-dyn.isp.belgacom.be] has quit [Client Quit]
05:15 -!- c0rw1n [~c0rw1n@193.47-244-81.adsl-dyn.isp.belgacom.be] has joined ##hplusroadmap
05:50 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
05:52 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
06:05 -!- eudoxia [~eudoxia@r186-55-145-241.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap
06:33 -!- Gurkenglas_ [Gurkenglas@dslb-178-005-223-183.178.005.pools.vodafone-ip.de] has joined ##hplusroadmap
06:37 < kanzure> nmz787: yes tdt has low efficiency, it falls off after adding one or two nucleotides, and only adds up to ~25 nt but i think this can be improved with selection experiments.
06:42 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-zhfofmegyzqleakl] has quit [Quit: Connection closed for inactivity]
06:44 -!- Jenda` [~bablbam@dekatron.hrach.eu] has quit [Ping timeout: 276 seconds]
06:49 -!- winsen [~winsen@unaffiliated/winsen] has joined ##hplusroadmap
07:10 -!- winsen_ [~winsen@unaffiliated/winsen] has joined ##hplusroadmap
07:14 -!- winsen [~winsen@unaffiliated/winsen] has quit [Ping timeout: 265 seconds]
07:15 -!- winsen [~winsen@unaffiliated/winsen] has joined ##hplusroadmap
07:15 -!- winsen_ [~winsen@unaffiliated/winsen] has quit [Ping timeout: 244 seconds]
07:26 < JayDugger> Good morning, everyone.
07:27 -!- Malvolio [~Malvolio@unaffiliated/malvolio] has quit [Quit: hacked by feminists]
08:01 -!- Gurkenglas_ [Gurkenglas@dslb-178-005-223-183.178.005.pools.vodafone-ip.de] has quit [Ping timeout: 252 seconds]
08:09 < kanzure> http://www.the-odin.com/diy-bacterial-crispr-kit/
08:12 < ebowden_> Oh cool.
08:18 -!- eudoxia [~eudoxia@r186-55-145-241.dialup.adsl.anteldata.net.uy] has quit [Quit: Leaving]
08:22 -!- SloanOnLinux [~sloan@162.72.132.128] has quit [Quit: KVIrc 4.9.2 Aria http://www.kvirc.net/]
08:24 < kanzure> .wik galantamine
08:24 < yoleaux> "Galantamine (Nivalin, Razadyne, Razadyne ER, Reminyl, Lycoremine) is used for the treatment of mild to moderate Alzheimer's disease and various other memory impairments, in particular those of vascular origin." — https://en.wikipedia.org/wiki/Galantamine
08:25 -!- SloanOnLinux [~sloan@162.72.132.128] has joined ##hplusroadmap
08:50 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
08:53 < kanzure> i wonder what ralph merkle thinks about the ethereum hard-fork. since he has switched to ethereum and "smart contracts". heh.
09:01 < JayDugger> Umm...did he mention it in that LTB podcast?
09:02 < JayDugger> Sorry, https://www.youtube.com/watch?v=dvzpvXLbpv4, Epicenter Bitcoin.
09:25 < kanzure> no, he pubished a paper
09:25 < kanzure> instead of working on molecular nanotech :\
09:32 -!- nmz787_i [ntmccork@nat/intel/x-jnhbqqhvcnqqetpq] has joined ##hplusroadmap
09:34 -!- Gurkenglas_ [Gurkenglas@dslb-178-005-223-183.178.005.pools.vodafone-ip.de] has joined ##hplusroadmap
09:44 < CaptHindsight> how many people would buy a $3k POSaM + PCR + CRISPR Kit?
09:46 < kanzure> well, i don't really have a use for dna hybridization, so... not me. if it did conjugation and ligation, yeah sure. i'd buy a $15k version, a $20k version, etc.
09:47 < CaptHindsight> that version will cost a bit more at first
09:47 < CaptHindsight> probably add a zero
09:48 < kanzure> why $200k?
09:48 < CaptHindsight> someones gotta pay for our time
09:51 < gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=ca9cc72a Bryan Bishop: render petertodd things >> http://diyhpl.us/diyhpluswiki/transcripts/mit-bitcoin-expo-2016/fraud-proofs-petertodd/
09:57 < maaku> Merkle had a looong oped in the latest issue of Cryonics about DAOs
09:57 < CaptHindsight> in medicine when a gene is considered a defect, how many base pairs have to be different to be considered a defect?
09:58 < kanzure> maaku: hopefully this ethereum blunder will convince him to work on other more interesting problems
09:58 < kanzure> CaptHindsight: so far biologists have studied "single nucleotide polymorphisms" the most, but there are other kinds of changes of course.
09:59 < kanzure> tweetosphere blowing up https://twitter.com/kanzure/status/757962716103311360
10:00 < CaptHindsight> referring only to single gene disorders
10:01 < kanzure> synthetic viruses are a good delivery tool for crispr gene editing stuff
10:01 < kanzure> eventually we will deploy an app store and such
10:02 < CaptHindsight> synthetic gene store
10:02 < CaptHindsight> order before 5 and click same day shipping
10:03 < CaptHindsight> order 3 or more genes and get one free virus for delivery
10:05 < CaptHindsight> for example with CF or sickle cell how much of the gene is actually wrong/defective?
10:06 < CaptHindsight> they have IDed 1000 variations in the CF gene
10:06 < CaptHindsight> The most common mutation (in 70% of cystic fibrosis patients) is a three-base deletion in the DNA sequence, causing an absence of a single amino acid in the protein product.
10:07 < CaptHindsight> looks like an easy fix
10:07 < CaptHindsight> so why no gene therapy yet for this?
10:08 < CaptHindsight> oh it's in clinical trials
10:19 < nmz787_i> CaptHindsight: CF has done gene therapy trials before... as I recall, first patient seemed to be succes, second's immune system reacted MUCH MUCH more than the first's and killed the person.... so dose-response varies MUCH more widely with that tech ( pretty sure that was viral packaging the fix )
10:19 < nmz787_i> that memory is from about 6 years ago
10:20 < nmz787_i> CaptHindsight: that is precisely why CRISPR trials are getting so much attention
10:20 < nmz787_i> because there is great precendent, but also great promise
10:20 < nmz787_i> (greatly horrible precendent?)
10:23 -!- winsen [~winsen@unaffiliated/winsen] has quit [Quit: Leaving]
10:24 < CaptHindsight> last year the Chinese did some tests on non-viable human embryos using CRISPR
10:25 < CaptHindsight> wasn't it something like 4 out of the 30 that worked and the others had odd mutations
10:26 < CaptHindsight> only four had the desired CCR5 mutation out of 45
10:26 < kanzure> better to stick with viral gene therapy
10:27 < CaptHindsight> On the plus side, the team did not find any unintended mutations in the embryos,
10:27 < CaptHindsight> I'd like to see the actual results
10:28 < CaptHindsight> what are the reasons for immune system response in the gene therapy with viral delivery?
10:28 < kanzure> immunological attack
10:28 < CaptHindsight> poor choice of virus? virus not modified enough?
10:29 -!- ArturSha1 [~ArturShai@195.114.248.3] has quit [Ping timeout: 258 seconds]
10:29 < CaptHindsight> what are the exact triggers?
10:30 < CaptHindsight> from my outsiders perspective I'm surprised by how much info is still missing
10:30 < CaptHindsight> looks like a wild west for research in all of this
10:31 < CaptHindsight> and the ethical arguments are like listening to right wing radio
10:33 < nmz787_i> CaptHindsight: definitely wild west
10:33 < CaptHindsight> what was behind the lack of stem cell research during Dubya term? Drug co roadblock or fear of offending the invisible angry man that lives in the sky?
10:34 < nmz787_i> Dubay...Dubai? THE CONNECTIONS!!!!
10:34 < nmz787_i> CaptHindsight: I thought it was religion
10:35 < CaptHindsight> Dubya/Dubya's
10:35 < nmz787_i> www.www.www
10:38 < CaptHindsight> In 2005, the State of California took out $3 billion in bond loans to fund embryonic stem cell research in that state.
10:39 < CaptHindsight> Jesus, how do you blow 3B on this and have so little to show for it?
10:39 < CaptHindsight> very carefully
10:40 < CaptHindsight> and purchase everything through Sigma
10:42 < nmz787_i> CaptHindsight: QC is a major issue with this kind of research
10:42 < nmz787_i> and clinicial trials in general
10:43 -!- cynsia [~cyn@pool-71-125-214-39.nycmny.east.verizon.net] has joined ##hplusroadmap
10:43 < kanzure> CaptHindsight: here are some things about gene therapy,
10:43 < kanzure> http://diyhpl.us/~bryan/papers2/gene-therapy/Engineering%20adeno-associated%20viruses%20for%20clinical%20gene%20therapy%20-%202015.pdf
10:43 < nmz787_i> as far as we know, sigma's purity/quality certificates aren't fake
10:43 < kanzure> http://diyhpl.us/~bryan/papers2/gene-therapy/Genome-editing%20technologies%20for%20gene%20and%20cell%20therapy%20-%202016.pdf
10:43 < kanzure> http://diyhpl.us/~bryan/papers2/gene-therapy/Therapeutic%20genome%20editing:%20prospects%20and%20challenges%20-%202015.pdf
10:44 < kanzure> http://diyhpl.us/~bryan/papers2/gene-therapy/Non-viral%20vectors%20for%20gene-based%20therapy%20-%20review%20-%202014.pdf
10:44 < kanzure> http://diyhpl.us/~bryan/papers2/gene-therapy/Two%20decades%20of%20clinical%20gene%20therapy%20-%20success%20is%20finally%20mounting%20-%202010.pdf
10:44 < nmz787_i> CaptHindsight: I wonder if any of that money went to lobbying
10:44 < CaptHindsight> I want to read over that CF case that ended badly
10:45 < CaptHindsight> you modify 3 base pairs in one gene, deliver by virus and the immune system went nuts
10:48 < CaptHindsight> selling the tools for simple gene mods like this could fund all the crazy full genome synthesis
10:49 < kanzure> autodesk has a synthetic virus production line that they have funded
10:49 < nmz787_i> maybe this was the person, but I thought it was specifically about CF that I am remembering https://en.wikipedia.org/wiki/Jesse_Gelsinger
10:49 < nmz787_i> CaptHindsight: looks like a CF trial had issues where virus genes turned up in the liver
10:49 < nmz787_i> http://www.jyi.org/issue/evaluation-of-the-clinical-success-of-ex-vivo-and-in-vivo-gene-therapy/
10:49 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Leaving]
10:50 < nmz787_i> last interactive thing here is good, but no citations http://learn.genetics.utah.edu/content/genetherapy/casestudy/
10:50 < nmz787_i> maybe this http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2570152/
10:50 < nmz787_i> .title
10:50 < yoleaux> Gene Therapy Using Adeno-Associated Virus Vectors
10:50 < kanzure> 2008?
10:50 < CaptHindsight> those idiots should have known better
10:51 < nmz787_i> kanzure: I'm sure there is a more recent version
10:51 < superkuh> Appropriate username.
10:52 < CaptHindsight> Inclusion of Gelsinger as a substitute for another volunteer who dropped out, despite Gelsinger's having high ammonia levels that should have led to his exclusion from the trial;
10:52 < kanzure> "Photosensitivity of neurons enabled by cell-targeted gold nanoparticles" http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4393361/
10:52 < CaptHindsight> bad choice of virus
10:53 < nmz787_i> 'bad' == 'only' quite possibly back then
10:54 < CaptHindsight> the cancer trials using the cold virus haven't ended in any deaths
10:54 < CaptHindsight> I guess they were hoping things would go well and for the associated fame
10:54 < nmz787_i> I am unfamiliar, but viruses can have specific entry-points, or organs they target
10:55 < CaptHindsight> what makes them so great for delivery
10:55 < CaptHindsight> how well is that understood?
10:56 < CaptHindsight> the chemistry of their membranes
10:56 < CaptHindsight> and target cells
10:57 < CaptHindsight> seems once this gets fully mapped out you can deliver whatever fits into a given virus that matches the target cells
10:58 < CaptHindsight> then it's on to synthesizing customized viruses vs modifying found
10:58 < kanzure> autodesk's synthetic virus stuff is re: dog cancer, it's a personalized synthetic virus
10:58 < kanzure> specifically targeting only that one dog patient's cancer
10:58 < CaptHindsight> for cancer they have use Cold, Polio, Measles and HIV
10:59 < CaptHindsight> autodesk also funded a SLA printer
11:00 < CaptHindsight> >100m for a oxygen permeable membrane that you don't even need to print as fast or faster that it can
11:00 < CaptHindsight> >$100m
11:02 < CaptHindsight> http://carbon3d.com/
11:02 -!- nmz787_i [ntmccork@nat/intel/x-jnhbqqhvcnqqetpq] has quit [Quit: Leaving.]
11:04 -!- Gurkenglas_ [Gurkenglas@dslb-178-005-223-183.178.005.pools.vodafone-ip.de] has quit [Ping timeout: 240 seconds]
11:04 < CaptHindsight> https://www.autodeskresearch.com/blog/hacking-viruses
11:04 < CaptHindsight> http://bionano.autodesk.com/
11:05 < kanzure> yep that's them
11:05 < kanzure> andrew is the one from the human genome project
11:05 < CaptHindsight> Import Cadnano format design files  oh god not here to
11:06 < CaptHindsight> all we need are DNA files in proprietary autodesk format
11:07 < kanzure> cadnano is open-source https://github.com/cadnano
11:07 < CaptHindsight> for how long?
11:07 < CaptHindsight> its Autodesk
11:08 < kanzure> CaptHindsight: they asked me to quit my job and work for HGP instead, http://diyhpl.us/~bryan/papers2/bio/The%20human%20genome%20project%20-%20write,%20hgp-write%20-%202016.pdf
11:08 < CaptHindsight> and write software for them?
11:09 < kanzure> run the project... organize teams, make up software roadmaps, etc.
11:09 < kanzure> release synthetic viruses unto the ocean.... and such.
11:09 < CaptHindsight> are they actually trying to make anything?
11:09 < kanzure> human genome synthesis, synthetic viruses, dna synthesis tech
11:09 < CaptHindsight> or is this just a big fund raiser?
11:10 < kanzure> they are raising $100-250 million, yes
11:11 < CaptHindsight> that's all I expected
11:11 < kanzure> well they will need to actually do something. which is where i come in. if they don't know what to do then i have a long list of things they should be doing.....
11:11 < CaptHindsight> would be nice if they build something that works
11:11 < Aurelius_Work2> kanzure : you accepted?
11:11 < kanzure> yes building things should be highest priority
11:12 < kanzure> Aurelius_Work2: sort of? it's still being defined.
11:12 < CaptHindsight> if they want sub-contract out the "build something" part I know people that do that  :)
11:13 < kanzure> i want the mto be committed to this. i still need to bring them an actual proposal. this is partly the reason for the polymerase draft doc that we have been editing last few days..
11:14 < CaptHindsight> I'm really good at developing acronyms as well
11:15 < kanzure> ah sounds like you'd fit in with the military
11:16 < kanzure> for some reason they estimate it should take 10 years to synthesize only one human genome molecule. but i think we can do better than that.
11:16 < CaptHindsight> like always, it depends on who is doing it
11:16 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap
11:17 < kanzure> they are modeling this off of the synthetic yeast project and such (see the refs in their hgp-write paper)
11:17 < CaptHindsight> number conveniently pulled from thin air
11:17 < CaptHindsight> also safe, no rush
11:18 < kanzure> well that's only $10M/year if it's 10 years and only raising $100M/year. $10M wont do much, unless you have the right team.
11:18 < CaptHindsight> Brain Research through Advancing Innovative Neurotechnologies (BRAIN) initiative centers]
11:19 < CaptHindsight> how many people do you need?
11:19 < CaptHindsight> I'll be done way before then
11:20 < CaptHindsight> kind of pointless since not all gets used anyway
11:28 < CaptHindsight> the tools are easy, what they should focusing on is all the cross referencing, proteins, immune system chemistry etc
11:36 < kanzure> tools are easy, but only when people are working on them
11:38 < CaptHindsight> in 10 years you'll be able to order them on alibaba
11:41 -!- Burnin8 is now known as Burninate
12:16 -!- jtimon [~quassel@55.31.134.37.dynamic.jazztel.es] has quit [Ping timeout: 258 seconds]
12:31 -!- duper [~super@ca4.blackcatz.org] has joined ##hplusroadmap
12:40 < kanzure> CaptHindsight: i think it's clear that most of biologists are not tool builders. so we have to outline the plan for them.
12:57 -!- nmz787_i [~ntmccork@134.134.139.70] has joined ##hplusroadmap
12:58 -!- jtimon [~quassel@55.31.134.37.dynamic.jazztel.es] has joined ##hplusroadmap
13:00 -!- nmz787_i1 [~ntmccork@134.134.139.70] has joined ##hplusroadmap
13:00 -!- nmz787_i [~ntmccork@134.134.139.70] has quit [Remote host closed the connection]
13:00 -!- nmz787_i1 [~ntmccork@134.134.139.70] has left ##hplusroadmap []
13:12 -!- hazirafel [~hazirafel@bzq-79-180-185-151.red.bezeqint.net] has joined ##hplusroadmap
13:20 -!- augur [~augur@2601:645:c100:63f1:c841:c5e7:97ac:44b3] has joined ##hplusroadmap
13:25 < kanzure> hazirafel: duper: hi.
14:05 < duper> hi
14:14 < CaptHindsight> kanzure: I'm not sure how much biology they do either since there is so much missing info
14:15 < kanzure> they haven't started yet. afaik.
14:33 -!- nmz787_i [ntmccork@nat/intel/x-ibsamndidaxtnigd] has joined ##hplusroadmap
14:35 < nmz787_i> kanzure: "falling off" explains nothing to me about 'efficiency' in terms of overall polymerization though, it can just hop back on in a moment... 25nt limit seems like it could be related to oligo interfering somehow, but a hard limit like that also seems to be unrelated to 'efficiency'
14:36 < nmz787_i> I mean, in synChem efficiency is just statistical chance extension will happen each round, most things I've read have been a simple scalar, not an equation
14:36 < kanzure> yes i think the mechanism limiting tdt to 25 to 50 bp is not known at the moment
14:36 < nmz787_i> i.e. 98%
14:36 < nmz787_i> nothing about 25nt the efficiency changes, etc
14:36 < nmz787_i> (i mean, it seems we aren't comparing apples to apples even though we are using the same word 'efficiency')
14:36 < kanzure> it's a little strange that the de novo dna synthesis capabilities of vent polymerase are not more seriously characterized
14:36 < kanzure> i guess nobody has a strong use case for random dna
14:37 < kanzure> or its friend, "deep vent polymerase"
14:38 < kanzure> there was something in one of these papers that mentioned tdt purification was a really annoying procedure, or some other set of problems
14:41 < pasky> We now have an MVP for our custom image classification tool - http://images.ailao.eu:7777/ - thoughts, ideas, usecase candidates, or pointers to potential customers welcome :)
14:44 < kanzure> pasky: btw did you see this one? "Synthesizing the preferred inputs for neurons in neural networks via deep generator networks" http://arxiv.org/pdf/1605.09304v1.pdf
14:48 -!- Orpheon [~Orpheon@213.200.193.129] has quit [Remote host closed the connection]
14:48 < kanzure> pasky: one of the projects that maaku wants to work on (and for me to fund) is live real-time audio transcription of the type that i do at events ( http://diyhpl.us/wiki/transcripts/ ). any chance that you could put some effort into that?
14:49 < pasky> kanzure: oooh that's awesome! thanks
14:50 < kanzure> and http://diyhpl.us/~bryan/papers2/ai/Towards%20an%20integration%20of%20deep%20learning%20and%20neuroscience%20-%202016.pdf
14:50 < pasky> kanzure: sorry that doesn't fit that well into what I'm doing now
14:51 -!- Orpheon [~Orpheon@213.200.193.129] has joined ##hplusroadmap
14:55 -!- augur [~augur@2601:645:c100:63f1:c841:c5e7:97ac:44b3] has quit [Remote host closed the connection]
14:59 < kanzure> https://www.src.org/calendar/e006096/#tab-agenda
15:03 < kanzure> "plant electronics" is interesting. i guess someone got tired of stringing up the lights on xmas trees.
15:03 -!- eudoxia [~eudoxia@r186-50-217-194.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap
15:05 < kanzure> "The Harvard Stem Cell Institute recently said they would start GDF11 human trials in 2026.  2026?  Well some of us may not make it that long.  At least now, the GDF11 door has been kicked wide open and my small, courageous GDF11 cohort has helped determine a safe and effective dose."
15:17 -!- c0rw1n [~c0rw1n@193.47-244-81.adsl-dyn.isp.belgacom.be] has quit [Quit: Konversation terminated!]
15:25 < maaku> pasky: yes I owe kanzure a project writeup with costs
15:26 < maaku> I see it as more than just solving a specific near-term problem though, and would like this to be part of a larger NLP / semantic processing project
15:26 < maaku> near term goal is definately just transcription related though
15:33 -!- ebowden_ [~ebowden@war2000509.lnk.telstra.net] has joined ##hplusroadmap
15:33 -!- ebowden_ [~ebowden@war2000509.lnk.telstra.net] has quit [Changing host]
15:33 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
15:34 < pasky> it's very interesting for me mid-term+long-term but not short-term
15:34 < pasky> I'm looking at bootstraping a revenue-making business here so that these efforts can become self-funding and progress grows fast
15:35 < ebowden_> Oh, what's the revenue maker?
15:35 < kanzure> pasky: i think transcripts are more likely to make money than image classification in the short-term :)
15:36 < kanzure> even semantic markup of audio streams (if transcription is too difficult)
15:38 < pasky> kanzure: who are the customers?
15:39 < pasky> ebowden_: for me, people who want machine learning; the trouble is, people usually want solutions, not machine learning, and you get busy developing the end-to-end solution and never get around to plugging in substantial machine learning, which is annoying and we're still trying to figure out ways around that
15:40 < pasky> so image classification is one of the best well-formed tasks that's actually possible to explain briefly, which is why we started with that
15:40 < ebowden_> Wanting solutions and not what you want? Those monsters!
15:40 < pasky> even though I'm much more interested in NLP
15:40 < pasky> ebowden_: yes, really!
15:40 < kanzure> pasky: for transcription? any business that has a meeting, so essentially all of them. N employees sitting around a table * avg hourly salary per employee, with an 8 person meeting you're spending $800/hour so you might as well throw a few bucks towards having a log.
15:42 < pasky> hmm, so why isn't anyone doing it yet?
15:42 < pasky> it makes sense
15:43 < kanzure> in the past, it was annoying to setup because nobody had microphones laying around, or you would have to dial in an extra service or something
15:43 < kanzure> but now since most employee meetings involve many smartphones, the interface can be much easier to setup
15:43 -!- jaboja [~jaboja@vps.jaboja.pl] has joined ##hplusroadmap
15:45 < pasky> I'll be thinking about that; you have a lot of transcripts, do you also have the audio recordings for some of these?
15:45 < maaku> pasky: don't get trapped in the consulting rabbit hole. make products
15:45 < pasky> for audio labelled training data is always big part of the pain
15:45 < maaku> pasky: audiobooks
15:45 < kanzure> there are dial-in services for delayed transcription i think, like some person on the other end of the line listening to the whole conversation, but i think people prefer near real-time output on a display or something. instead of just a black hole where you get a document back a few days later.
15:46 < kanzure> i don't have audio for many of these. there are some youtube videos, sometimes...
15:46 < maaku> kanzure: there are those two. we used one at the virtual institutes at NASA
15:46 < pasky> audiobooks vs. smartphone input of room full of random people seated differently is huge difference
15:46 < kanzure> audiobooks is a better idea.
15:46 < maaku> they would do "real-time" transcriptions of scientific talks
15:46 < pasky> and audiobooks are basically entirely unaligned
15:46 < maaku> basically the same as closed captioning on 24/hr news (another target market)
15:46 < kanzure> pasky: you could play the audio and then apply noise, make it play in a 3d environment and add noise, e.g. like all those stupid second life projects
15:46 < maaku> pasky: walk before you run
15:47 < kanzure> maaku: who does real-time transcriptions of scientific talks ? nasa itself?
15:47 < pasky> i think the different speakers (plus different accents) are the most important missing ingredient
15:47 < kanzure> or are they contracting that out to some company?
15:47 < maaku> kanzure: some contractor in colorado, don't know details
15:47 < maaku> was a federal contract
15:47 < kanzure> weird.
15:47 < pasky> one idea i seriously entertained was learning speech synthesis using different voices from audiobooks
15:47 < kanzure> think of all the science they are losing by not recording everything :)
15:47 < pasky> the advantage is that it's easier to align the data
15:48 < pasky> but i suppose you could do it in reverse too
15:48 < maaku> pasky: mid-term for this project I'd like to wear a pebble watch (or something) that does a nice weblog of my daily interactions
15:48 < kanzure> also it's possible that voice recognition is a hard problem and maybe google/amazon is the only one with enough data to get audio recognition right
15:48 < pasky> maaku: battery
15:48 < kanzure> battery is not a concern for audio recording
15:49 < pasky> i thought pebble has a small battery
15:49 < pasky> maybe i'm wrong
15:49 < maaku> it has a 7-day battery
15:49 < pasky> audio recording is probably cheaper than transmitting the waveform somewhere else
15:49 < pasky> maaku: 7-day for typical pebble power draw
15:50 < kanzure> this is a dumb argument. we can trivially build long-life audio recording equipment.
15:50 < maaku> recording is only going to reduce that by a small (<10x) factor
15:50 < maaku> but yes what kanzure said
15:50 < kanzure> if we can't build audio recording equipment then we might as well give up on life
15:50 < pasky> yes, i didn't want to get into that argument, sorry
15:50 < maaku> I have a cheap chinese made wristband recorder that lasts 20hrs on record
15:51 < pasky> after all for the near-term usecases you don't need battery
15:51 < pasky> (or it's a damn smartphone)
15:51 < kanzure> "Phone recognition for mixed speech signals: Comparison of human auditory cortex and machine performance" https://www.semanticscholar.org/paper/Phone-Recognition-for-Mixed-Speech-Signals-Chang-Edwards/7d9881141b42da2e63d363732406647dd5df838d/pdf
15:51 < maaku> pasky: there's obvious DARPA uses for this technology (or whatever your local european equivalent is)
15:52 < pasky> that's what concerns me :)
15:53 < maaku> yes well it's not the tool but who uses it...
15:53 < kanzure> even without actual transcription, having topic highlights would still be useful
15:53 < kanzure> such as for jumping around in audio streams
15:53 < ebowden_> Let's not make medicines either. Did you know that the military uses medicines!?
15:54 < maaku> ebowden_: in all seriousness I think pasky is right to be concerned ;)
15:54 < kanzure> you could have something that spits out the 3 locations where a certain word or cluster of words were said
15:54 -!- esmerelda [~andares@unaffiliated/jacco] has quit [Ping timeout: 276 seconds]
15:54 < kanzure> and then use that instead of randomly hopping around a long video feed or whatever
15:54 < ebowden_> Because it could drastically facilitate mass surveillance?
15:55 < maaku> I'm talking about technology that if subvertly installed on people's phones or baseband processors could record and process every conversation everywhere
15:55 < pasky> i really think the main problem is data, i think audiobooks won't do much good - but what i'd be curious about is how well do the stock speech to text engines do on audiobooks
15:56 < kanzure> if your concern is that audiobooks don't have enough noise, it is trivial to add noise to any audio stream
15:56 < kanzure> you can even look at common types of noise in speech recognition tasks, and emulate those specific types of noise
15:56 < pasky> 00:47 < pasky> i think the different speakers (plus different accents) are the most important missing ingredient
15:57 < kanzure> audiobooks can give you accents actually
15:57 < kanzure> but unfortunately i don't think we have a large freely available audiobook corpus anyway
15:57 < pasky> you could stitch together different audiobooks
15:57 < pasky> why does it have to be free?
15:58 < kanzure> because who wants to spend $20,000 on books at amazon.com?
15:58 < kanzure> have to get the data somehow
15:58 < kanzure> maybe there's a giant torrent
15:59 < pasky> i don't think you need that much data
15:59 < ebowden_> pasky, have you considered that intelligence bodies might already possess this capability?
16:00 < kanzure> it's not like they will give you the tech if you ask them nicely, ebowden
16:00 < pasky> you can easily get thousands of hours e.g. from song of ice and fire, harry potter, discworld audiobooks, and at some near point your available gpus become the limiting factor
16:01 < kanzure> er how many hours do you need? i was expecting... a lot.
16:01 < ebowden_> kanzure, I know, but what it could mean is that developing such tech might not give such actors any new capabilities.
16:01 < pasky> I'd expect you could train a decent domain-limited speech to text engine from a single harry potter audiobook, the much bigger problem is data
16:01 < kanzure> ebowden_: imho i think deploying it as a service would be the place of concern. they will intercept all the audio and transcripts.
16:02 < pasky> just having an audio stream and text stream is definitely not enough, you need alignment and the more alignment the better
16:02 < pasky> but you could abuse an existing speech reco engine to get that
16:02 < kanzure> pasky: oh that's a really small amount of audio. interesting.
16:02 < pasky> so all in all, this sounds like a really fun experiment
16:02 < kanzure> well, actually, alignment isn't too bad
16:02 < pasky> i wish i had the time :(
16:02 -!- esmerelda [~andares@unaffiliated/jacco] has joined ##hplusroadmap
16:03 < ebowden_> pasky, you could make your own spy agency! :D
16:03 < kanzure> you spit out some text, and then you run some text distance similarity between the generated text, and then similarity against all N word phrases from the book
16:03 < kanzure> in fact, you might not want alignment
16:03 < kanzure> for example: if the model *misses* an entire word, that's fine, in comparison to getting the rest of the sentence wrong.
16:04 < kanzure> also i think a prank version of this could be made, where instead of acutal speech recognition, it just spits out anything that sounds vaguely similar to the words.
16:05 < pasky> but in fact, this gets really complicated because the first question is, what's actually the task here, formally? on the input you have a time series of cepstral coefficients or something, on the output...?
16:05 < kanzure> task is to replace my typing activities
16:05 < pasky> that's not the formal version ;-)
16:06 < pasky> there's all kinds of time warping you may want to do, and then decide to maybe label the time series with phonemes, or do something else
16:06 < kanzure> speaker recognition would also be an interesting task to start with.
16:06 < pasky> but for that you need some target variable value for each time step, i.e. phoneme-level alignment
16:07 < ebowden_> Sometimes I wonder if building little wetware brains and training them would be easier than this.
16:07 < pasky> kanzure: if the company, for these 8 people, have hours of voice recordings where only they are speaking, that's totally doable!
16:07 < pasky> *they = each one
16:07 < kanzure> well, they probably have hours of them speaking, but it's probably a disjoint hour, heh
16:07 < pasky> that's no problem, but it's unlikely it'll be *only* them speaking
16:08 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has joined ##hplusroadmap
16:08 < maaku> kanzure: I have the harry potter audiobooks
16:08 < maaku> that's more than enough to get started.. that guy has incredible voice range too
16:09 < pasky> maaku: i think running a stock speech to text engine (cmu sphinx, kaldi, ?julius?) on the audiobooks would be a great experiment and baseline
16:10 < maaku> pasky: agreed. and thanks for the starting list of engines
16:11 < kanzure> maybe the solution is to just donate my brain to science and then we can see what neural connectivity is appropriate for this problem
16:11 < kanzure> there's a lot of real-time adjustment that i do when i'm typing like that
16:11 < kanzure> based on the speaker
16:13 < kanzure> ibm was using 2000 hours of audio for training http://arxiv.org/pdf/1505.05899.pdf
16:13 < kanzure> "The most relevant data are the transcripts of the 1975 hour audio data used for training the acoustic model, consisting of about 24M words."
16:16 < pasky> yep that's the ballpark i expected
16:16 < pasky> maaku: there are two versions of the audiobooks btw - https://www.youtube.com/watch?v=zbd2mzpYRrE
16:17 < pasky> i listened to the UK one, so I prefer it ;) but voldemort is better by jim dale
16:17 < kanzure> NIST speech data http://www.itl.nist.gov/iad/mig/tests/spk/
16:19 < kanzure> http://www.lscp.net/persons/dupoux/bootphon/zerospeech2014/website/
16:27 < kanzure> so....
16:27 < kanzure> why are all these speech recognition papers about words? when i'm typing, i am typing characters, not words.
16:29 < pasky> how do you figure out what characters to type?
16:29 < pasky> and how often are the characters not parts of words?
16:29 < pasky> in, say, japanese, character-level transcription would be probably a lot more tractable
16:29 < kanzure> i type what sounds right
16:30 < pasky> what sounds right is determined by the words they form
16:30 < pasky> furthermore, even the word context is important, i.e. the language model
16:32 < maaku> pasky: in Japanese I would think you would need a semantic component to figure out what characters
16:32 < maaku> actually maybe a markov model is sufficient
16:39 < maaku> pasky: regarding your company have you done any work towards selling products directly?
16:40 < nmz787_i> kanzure: can't you just abuse google translate (english audio in, english out)... or google voice (pass your recorded audio through google voicemail, which gets transcribed and emailed to you)? I mean, I get wanted 'your own' solution.
16:41 < pasky> maaku: we have one paying customer for the vize.it product already (not the frontend, but using the backend) and getting in touch with a few others
16:41 -!- eudoxia [~eudoxia@r186-50-217-194.dialup.adsl.anteldata.net.uy] has quit [Quit: Leaving]
16:41 < nmz787_i> i've used google translate in the opposite way, text in, MP3 out
16:43 < nmz787_i> https://github.com/pndurette/gTTS
16:43 -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Leaving]
16:47 < kanzure> nmz787_i: google translate will block you after a certain number of queries
16:48 < kanzure> maaku: there's also some interesting ways to "cheat" with speech recognition. you don't have to type what was said. all that matters is that you type something that is vaguely similar and holds most of the same semantic meaning.
16:53 < streety> that seems more like a "cheat" for transcription, you still need speech recognition to be good enough to get the meaning
16:55 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Remote host closed the connection]
16:56 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-hfntczzkxbsiewrt] has joined ##hplusroadmap
16:57 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
17:01 -!- esmerelda [~andares@unaffiliated/jacco] has quit [Ping timeout: 276 seconds]
17:02 < kanzure> well there other things you can cheat at; you don't need the exact words, you can often use other words.
17:02 < kanzure> or different spellings
17:05 < kanzure> "An example of such a system is the MIT Spoken Lecture Processing project [8], where the developers had collected ove 500 hours of recordings, of which more than 200 hours had been transcribed. Each lecturer had between 1–30 hours of speech .. language models were trained on more than 6 million English words. This system achieved a Word Error Rate (WER) of 17% [9]."
17:05 -!- jaboja [~jaboja@vps.jaboja.pl] has quit [Ping timeout: 244 seconds]
17:06 < kanzure> " Kemp et. al [11] found unsupervised training to decrease WERs from 32.1% to 20.6%, using as little as 30 minutes of transcribed and 50 hours of untranscribed data"
17:06 < kanzure> from http://www.nwu.ac.za/sites/www.nwu.ac.za/files/files/v-must/Publications%202014/devilliers-2014-lecture-transcription.pdf
17:15 < nmz787_i> well it is past EOD and no news on Lightning Terminators...
17:16 < nmz787_i> maybe I'll have a friend with a business (and lab address with lease papers) email try
17:16 -!- ebowden_ [~ebowden@war2000509.lnk.telstra.net] has joined ##hplusroadmap
17:16 -!- ebowden_ [~ebowden@war2000509.lnk.telstra.net] has quit [Changing host]
17:16 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
17:17 -!- esmerelda [~andares@172.58.47.9] has joined ##hplusroadmap
17:17 -!- esmerelda [~andares@172.58.47.9] has quit [Changing host]
17:17 -!- esmerelda [~andares@unaffiliated/jacco] has joined ##hplusroadmap
17:17 < kanzure> hmm, the japanese parliament has been running speech recognition since 2010 https://www.semanticscholar.org/paper/Transcription-System-Using-Automatic-Speech-Kawahara/6e033da6944b39d50b5b785bd786250c502b01d4/pdf
17:17 < nmz787_i> http://hackaday.com/2016/07/26/pressure-formed-parabolic-mirror-from-a-mylar-blanket/
17:27 < nmz787_i> .title http://www.techbriefs.com/Briefs/May98/GSC13783.html
17:27 < yoleaux> NASA Tech Briefs Archive :: NASA Tech Briefs
17:27 < nmz787_i> "Deflection of Circular Membrane Under Differential Pressure"
17:28 < kanzure> maaku: http://diyhpl.us/~bryan/papers2/ai/speech-recognition/?C=M;O=A
17:28 -!- c0rw1n [~c0rw1n@193.47-244-81.adsl-dyn.isp.belgacom.be] has joined ##hplusroadmap
17:29 -!- nmz787_i1 [ntmccork@nat/intel/x-uvqcbzlhkvavpplq] has joined ##hplusroadmap
17:29 -!- nmz787_i1 [ntmccork@nat/intel/x-uvqcbzlhkvavpplq] has quit [Client Quit]
17:31 -!- nmz787_i [ntmccork@nat/intel/x-ibsamndidaxtnigd] has quit [Ping timeout: 244 seconds]
17:51 -!- HEx1 [~HEx@hexwab.plus.com] has quit [Ping timeout: 260 seconds]
17:54 -!- HEx1 [~HEx@hexwab.plus.com] has joined ##hplusroadmap
18:11 -!- maaku [~quassel@173-228-107-141.dsl.static.fusionbroadband.com] has quit [Quit: No Ping reply in 180 seconds.]
18:12 -!- maaku [~quassel@173-228-107-141.dsl.static.fusionbroadband.com] has joined ##hplusroadmap
18:41 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has joined ##hplusroadmap
18:45 -!- yashgaroth [~yashgarot@2602:306:35fa:d500:f5e0:f867:a11d:8d52] has joined ##hplusroadmap
18:52 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Remote host closed the connection]
19:00 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has quit [Ping timeout: 276 seconds]
19:00 < kanzure> yashgaroth: hi
19:00 < yashgaroth> yo
19:05 -!- justanotheruser [~Justan@unaffiliated/justanotheruser] has joined ##hplusroadmap
19:09 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
19:12 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-hfntczzkxbsiewrt] has quit [Quit: Connection closed for inactivity]
19:23 -!- ebowden_ [~ebowden@war2000509.lnk.telstra.net] has joined ##hplusroadmap
19:23 -!- ebowden_ [~ebowden@war2000509.lnk.telstra.net] has quit [Changing host]
19:23 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
19:44 -!- strangewarp [~strangewa@c-76-25-206-3.hsd1.co.comcast.net] has quit [Read error: Connection reset by peer]
19:45 -!- strangewarp [~strangewa@c-76-25-206-3.hsd1.co.comcast.net] has joined ##hplusroadmap
19:54 -!- strangewarp [~strangewa@c-76-25-206-3.hsd1.co.comcast.net] has quit [Read error: Connection reset by peer]
19:55 -!- ArturSha1 [~ArturShai@195.114.248.3] has joined ##hplusroadmap
19:55 -!- strangewarp [~strangewa@c-76-25-206-3.hsd1.co.comcast.net] has joined ##hplusroadmap
20:00 < kanzure> hm.
20:05 < kanzure> "Successful serial recloning in the mouse over multiple generations" http://www.cell.com/cell-stem-cell/fulltext/S1934-5909(13)00008-8
20:05 < kanzure> ".. Previously, we proposed that repeated rounds of genomic reprogramming via serial cloning might lead to an increase in efficiency over successive generations because of the selection of easily reprogrammable cells. Disappointingly, however, it has been found that the success rate in fact decreased with each iteration. In one study, only one cloned mouse was produced in the sixth generation from more than 1,000 nuclear transfer ...
20:05 < kanzure> ... attempts—but it was cannibalized by its foster mother (Wakayama et al., 2000)."
20:10 < yashgaroth> lotta problems with somatic cell cloning b/c of mutations that accumulate over time; methylated CpGs regulating genes for a different tissue type that'll get mutated
20:11 < kanzure> yashgaroth, if i decided to reign viral chaos unto the lands would you join me in support
20:11 < yashgaroth> ya ok
20:12 < kanzure> kthx
20:12 < yashgaroth> np
20:23 -!- jtimon [~quassel@55.31.134.37.dynamic.jazztel.es] has quit [Ping timeout: 252 seconds]
20:34 -!- withanr [~notauser@185.65.134.75] has joined ##hplusroadmap
20:35 -!- HEx1 [~HEx@hexwab.plus.com] has quit [Ping timeout: 260 seconds]
20:37 -!- HEx1 [~HEx@hexwab.plus.com] has joined ##hplusroadmap
20:40 < kanzure> maaku: okay another interesting cheat we should propose using is to simplify the target language and insist that everyone speaks the simplified sound system
20:41 < kanzure> maaku: we could also do some other tricks like, having the speaker's smartphone relay data through ultrasound regarding their speech style and other gotchas
21:38 < fenn> beep boop
21:46 -!- Burnin8 [~Burn@pool-108-31-204-167.washdc.fios.verizon.net] has joined ##hplusroadmap
21:47 -!- Orpheon [~Orpheon@213.200.193.129] has quit [Ping timeout: 240 seconds]
21:49 -!- Burninate [~Burn@pool-108-31-204-167.washdc.fios.verizon.net] has quit [Ping timeout: 240 seconds]
22:04 -!- withanr [~notauser@185.65.134.75] has quit [Quit: Leaving]
22:30 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has joined ##hplusroadmap
22:50 -!- augur [~augur@noisebridge130.static.monkeybrains.net] has quit [Remote host closed the connection]
22:56 -!- yashgaroth [~yashgarot@2602:306:35fa:d500:f5e0:f867:a11d:8d52] has quit [Quit: Leaving]
23:12 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-hbewnqxaxfwdleex] has joined ##hplusroadmap
23:13 -!- Albatross [~Albatross@unaffiliated/albatross] has joined ##hplusroadmap
23:24 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has quit [Remote host closed the connection]
23:50 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-hbewnqxaxfwdleex] has quit [Ping timeout: 250 seconds]
23:51 -!- PatrickRobotham [uid18270@gateway/web/irccloud.com/x-dpmflowjlvrcrllq] has joined ##hplusroadmap
23:51 -!- augur [~augur@2601:645:c100:63f1:8906:11fa:cb00:449a] has joined ##hplusroadmap
23:53 -!- ebowden_ [~ebowden@CPE-101-180-249-107.lnse3.cha.bigpond.net.au] has joined ##hplusroadmap
23:53 -!- ebowden_ [~ebowden@CPE-101-180-249-107.lnse3.cha.bigpond.net.au] has quit [Changing host]
23:53 -!- ebowden_ [~ebowden@unaffiliated/ebowden] has joined ##hplusroadmap
--- Log closed Wed Jul 27 00:00:08 2016