| --- Day changed Sun Jul 27 2008 | ||
| ybit | is kanzure_ a bot? | 00:04 |
|---|---|---|
| kanzure | No. | 00:37 |
| kanzure | it's my server | 00:37 |
| -!- Phreedom [n=freedom@195.216.211.159] has quit [Read error: 110 (Connection timed out)] | 01:29 | |
| kanzure | fenn: ruby on rails or django looks like it autogenerates codebases from classes and models of the saem sort that we want to use YAML for serialization | 01:32 |
| kanzure | ignoring the fact that it's used for web tech, eh | 01:32 |
| kanzure | just imagine it not being web and for something else related to skdb | 01:32 |
| -!- procto [n=boo@logarchy.org] has quit [Read error: 60 (Operation timed out)] | 02:47 | |
| -!- procto [n=boo@logarchy.org] has joined #hplusroadmap | 02:49 | |
| -!- Phreedom [n=freedom@195.216.211.159] has joined #hplusroadmap | 04:40 | |
| -!- Netsplit over, joins: procto | 04:51 | |
| -!- Phreedom [n=freedom@195.216.211.159] has quit [Read error: 110 (Connection timed out)] | 06:09 | |
| -!- jm [n=jm@p57B9DB1D.dip.t-dialin.net] has joined #hplusroadmap | 07:55 | |
| -!- Phreedom [n=freedom@195.216.211.159] has joined #hplusroadmap | 09:08 | |
| -!- kanzure [n=bryan@adsl-76-229-131-47.dsl.pltn13.sbcglobal.net] has quit [Connection timed out] | 12:10 | |
| -!- kanzure [n=bryan@76-220-201-251.lightspeed.oshkwi.sbcglobal.net] has joined #hplusroadmap | 12:11 | |
| kanzure | heh, I like these people | 14:13 |
| kanzure | I had something of a 20-depth argument tree going with Steve for a while, and he actively addressed it as the argument tree | 14:14 |
| kanzure | hm | 14:14 |
| fenn | paul stamets mushroom antiviral extract is interesting | 14:32 |
| kanzure | hm, django has no flat file object relational mapper, no flat file database support whatsoever | 14:38 |
| kanzure | otherwise that might have been a plug-and-play solution right there | 14:38 |
| kanzure | http://showmedo.com/videos/video?name=1100000& "create a wiki in 20 minutes with django" | 14:38 |
| kanzure | django is only using things like mysql, postgresql, sqlite etc. | 14:38 |
| kanzure | just need to write it in flat files instead | 14:39 |
| kanzure | there's a python-git module that could be used, I think, but I haven't investigated the architecture yet | 14:39 |
| kanzure | okay, guess django isn't right for it | 14:51 |
| kanzure | it was close, or I'm just terribly confused | 14:51 |
| fenn | i'm still trying to figure out what it is exactly | 14:52 |
| kanzure | ruby on rails for python | 14:53 |
| kanzure | I think what it does is use your models to generate a mysql db and then the input/output pages automatically for whatever you're working on | 14:53 |
| kanzure | back when I used to do web app development nonsense I was always writing pages to interface back with the database | 14:54 |
| fenn | yeah well i cant figure out what the big deal about ruby is either :\ | 14:54 |
| kanzure | i.e., view.php, admin.php, edit.php, all sorts of stuff | 14:54 |
| kanzure | so this just takes your model that you write up, and spits out those pages automatically | 14:54 |
| kanzure | that's about it. | 14:54 |
| -!- wrldpc [n=wrldpc@pool-71-174-89-247.bstnma.fios.verizon.net] has joined #hplusroadmap | 15:38 | |
| wrldpc | terry grossman is being interviewed here tonight: http://www.ustream.tv/channel/sunday-evening-update | 15:40 |
| fenn | kanzure: i just watched 'django wiki in 20 minutes' and i think i like it | 15:43 |
| fenn | you can really get your hands into the guts, but it's still clean and simple | 15:43 |
| fenn | reminds me of Mason without the suck factor | 15:55 |
| -!- Splicer [n=p@h93n1c1o261.bredband.skanova.com] has joined #hplusroadmap | 16:15 | |
| kanzure | okay, good | 16:32 |
| Phreedom | fenn: should also take a look at RoR ;) | 16:42 |
| -!- wrldpc_ [n=wrldpc@pool-71-174-89-247.bstnma.fios.verizon.net] has joined #hplusroadmap | 16:51 | |
| -!- wrldpc [n=wrldpc@pool-71-174-89-247.bstnma.fios.verizon.net] has quit [Read error: 60 (Operation timed out)] | 16:52 | |
| kanzure | stem cells into gametes? | 16:55 |
| kanzure | http://heybryan.org/docs/uncertain_future.doc <-- is what we're doing heh' | 17:02 |
| fenn | yay doc files | 17:12 |
| fenn | here's a summary for those of you that hate java and OOO | 17:12 |
| fenn | The Uncertain Future is an innovative probabilistic modeling system where you insert your guesses about variables related to the future of Artificial Intelligence and humanity, and are presented with colorful graphs and statistics regarding how the future might turn out if your beliefs are roughly correct. You can then email these results to your friends, post them on your blog or Myspace. | 17:12 |
| fenn | axe grinding through bayesian probability calculations | 17:14 |
| Phreedom | ah this is for prophets | 17:24 |
| Splicer | It reminds me a bit of a course I took called "Financial evaluation of entrepreneural companies" | 17:41 |
| Splicer | It was about 15 parameters that had to be subjectively evaluated | 17:42 |
| wrldpc_ | grossman canceled :\ | 17:51 |
| kanzure | fenn: yep | 18:58 |
| kanzure | http://www.singinst.org/blog/2008/06/23/siai-2008-summer-research-project/ | 19:08 |
| kanzure | "They also hope to create an early version of an eventual, publicly viewable wiki, mapping out the total space of scenarios involving AGI and existential risk." | 19:08 |
| kanzure | publicly viewable wiki | 19:08 |
| kanzure | so the group isn't SIAI, but rather "we're researchers and would like to do something interesting related to what you work on please give us a grant" "okay" | 19:12 |
| kanzure | hm | 19:27 |
| kanzure | Steve is published | 19:28 |
| kanzure | Fierer, N, M Breitbart, J Nulton, P Salamon, C Lozupone, R Jones, M Robeson, R Edwards, B Felts, S Rayhawk, R Knight, F Rohwer, RB Jackson (2007; in press) Metagenomic and small-subunit rRNA analyses of the genetic diversity of bacteria, archaea, fungi, and viruses in soil. Applied and Environmental Microbiology. | 19:28 |
| kanzure | Angly F, B Felts, M Breitbart, P Salamon, RA Edwards, C Carlson, AM Chan, M Haynes, S Kelley, H Liu, J Mahaffy, JE Mueller, J Nulton, R Olson, R Parsons, S Rayhawk, CA Suttle, F Rohwer (2006) The marine viromes of four oceanic regions. PLoS Biology. 4 (11). 2121-2131. | 19:28 |
| kanzure | (Steve Rayhawk) | 19:28 |
| kanzure | that last one: | 19:28 |
| kanzure | http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1634881 | 19:28 |
| fenn | this reminds me of 'limits to growth' - perhaps you could contact the club of rome :) | 19:33 |
| kanzure | huh? | 19:36 |
| kanzure | I don't remember 'limits to growth' | 19:36 |
| fenn | it was before your time (and mine) http://en.wikipedia.org/wiki/Limits_to_growth | 19:41 |
| fenn | apparently it got eric drexler annoyed enough to prove it wrong | 19:41 |
| fenn | so that's why he started thinking about nanotech | 19:41 |
| fenn | something like that anyway | 19:42 |
| kanzure | oh, it's Meadows | 19:42 |
| kanzure | "Dancing with systems" and "12 leverage points" | 19:42 |
| kanzure | she's with the Whole Earth magazine crowd | 19:42 |
| kanzure | or she was, at least | 19:42 |
| fenn | right, that's how i learned about it | 19:43 |
| kanzure | *at least until she died | 19:46 |
| fenn | death == limit to growth | 19:48 |
| * kanzure reminds everyone to stop dying | 19:52 | |
| fenn | fools! death is for the weak! | 19:52 |
| kanzure | and the old! | 19:53 |
| kanzure | oops | 19:53 |
| fenn | something seems wrong about them lumping the largest number of deaths into "85 and over" http://www.disastercenter.com/cdc/Leading Cause of Death 1999-2005.html | 19:55 |
| fenn | wow alzheimers killed a lot | 19:59 |
| kanzure | bah, they want to do the calculations in the browser | 20:01 |
| kanzure | in javascript | 20:01 |
| kanzure | methinks this is not a good idea | 20:01 |
| fenn | um, no they don't? | 20:02 |
| fenn | oh you're talking about the siai survey | 20:03 |
| kanzure | yes | 20:03 |
| fenn | if it's all browser-side then how do you get any survey results? | 20:03 |
| kanzure | xmlhttprequest them back to the server | 20:03 |
| fenn | do they want javascript for zoomable graphs and sorting n stuff? | 20:04 |
| * fenn would rather just get a nice optional svg | 20:04 | |
| kanzure | no, not just that | 20:04 |
| kanzure | I'm not entirely certain how this is supposed to work, but | 20:04 |
| kanzure | here's what I know so far: the user configures up to 40 different variables via the graphs, these are then sent back to the server into some black box that then runs through 100,000 permutations | 20:04 |
| kanzure | of a logic tree for statistical reasoning, and then this data is saved as a session perhaps, and when it's done the user gets to play around with different graphs for the scenario. | 20:04 |
| fenn | it seems a lot easier to write the code for server-side calcs at least | 20:05 |
| kanzure | basically it's "we don't trust people to do probabilistic reasoning" | 20:05 |
| kanzure | yeah | 20:05 |
| kanzure | exactly, but nobody is telling me what these calculations are | 20:05 |
| kanzure | so. :) | 20:05 |
| fenn | i imagine it looking something like a lisp expression | 20:05 |
| fenn | i think writing those calcs will be the hard part | 20:06 |
| fenn | or at least getting it right | 20:06 |
| fenn | anyone can throw some bullshit together and call it statistics | 20:06 |
| kanzure | there's a few whiteboards surrounding me with lots of fancy calculus for differential equations that I don't exactly know how to solve | 20:06 |
| fenn | have you brought up open source considerations? | 20:06 |
| kanzure | licensing of the source code? | 20:07 |
| fenn | there's not really much point to just putting some black box up on the web that you just have to believe | 20:07 |
| kanzure | correct | 20:07 |
| fenn | it's scientifically irresponsible | 20:07 |
| fenn | you can't have a meaningful discussion | 20:08 |
| fenn | this was one of the major problems with limits to growth | 20:08 |
| fenn | they didnt publish their calculations until like 4 years later | 20:08 |
| kanzure | lame. | 20:08 |
| fenn | (the other problem was they got the calcs wrong :P) | 20:08 |
| fenn | just try to make sure they dont use some awful proprietary simulator software ok :) | 20:10 |
| kanzure | fenn: so in general, what Anna, the one who flew me up here and the one who is leading this whole thing, she simply wants ways to reduce existential risk of human extinction, and thinks this is a good step forward | 20:10 |
| kanzure | verbalizing the alternative epistemology is not easy | 20:10 |
| fenn | any kind of world game is a good idea | 20:10 |
| fenn | better if it's meaningful (accurately represents reality) | 20:11 |
| fenn | maybe you could start out with sliders at default values that siai picks | 20:11 |
| kanzure | right, right | 20:11 |
| fenn | or have various futurists' opinions | 20:11 |
| kanzure | for some reason there's going to be a hardcoded limit to the input factors that the users have to variablize or slide or whatever | 20:12 |
| fenn | a sort of meyers-briggs test for futurists :) | 20:12 |
| kanzure | for instance, one of the 40 factors might be "hardware costs into the future" | 20:12 |
| kanzure | so a 40% probability of $5/flop in 2010 or something | 20:12 |
| kanzure | but I don't know why it's hard coded | 20:12 |
| fenn | describing that as a probability doesn't seem right | 20:12 |
| kanzure | probably because the backend model is not automatically extendable | 20:12 |
| fenn | do they actually have code already? | 20:13 |
| wrldpc_ | to nublabs! | 20:13 |
| -!- wrldpc_ [n=wrldpc@pool-71-174-89-247.bstnma.fios.verizon.net] has quit [] | 20:13 | |
| kanzure | nope | 20:14 |
| kanzure | and it's kind of weird trying to write this web app without actually knowing what that model is doing | 20:15 |
| kanzure | or the input/output for that matter | 20:15 |
| fenn | if they dont have code then talking about hard coded parameters seems silly dont you think? | 20:15 |
| kanzure | just inputting a median and bound via point-and-click, and then -- there is no "and then", there's no where to send this data to. No testing server, no blackbox, ... | 20:15 |
| kanzure | yeah :( | 20:16 |
| fenn | well, let me put in a good word for functional programming | 20:16 |
| kanzure | as opposed to OOP? | 20:17 |
| kanzure | :) | 20:17 |
| fenn | right | 20:17 |
| fenn | this is a functional sort of problem | 20:17 |
| fenn | numbers in, numbers out, and all sorts of hell in between that may or may not get calculated | 20:17 |
| fenn | are you familiar with the "model, view, controller" school of software design? | 20:28 |
| fenn | basically it's a way to ensure that you can use multiple different interfaces without lots of hacking | 20:30 |
| kanzure | haven't heard of it phrased like that, but I'm sure most of my work is like that | 20:30 |
| fenn | doing the model calcs in the viewer app would totally ignore all the goodness you get from MVC | 20:32 |
| fenn | anyway it's a buzzword you can point to, so its not just 'wah, i'm the new guy and i dont like it' | 20:32 |
| kanzure | yeah .. I recommended a cgi gateway or something, but it looks like they really want to do it in the viewer app because they don't want to have any server hardware requirements | 20:37 |
| kanzure | ahem. :( | 20:37 |
| fenn | why not.. its not like anyone is going to look at it :) | 20:37 |
| fenn | just write it so you have something to play with | 20:39 |
| fenn | then they can use it or not | 20:39 |
| kanzure | heh | 20:40 |
| kanzure | for now they want a 'slide show of questions for svg" for parts 1-1 through 1-4 in that doc file | 20:40 |
| kanzure | and then this would be some sort of input to some magical calculation from the mathgnome that I haven't heard of yet, and then an output graph will be generated | 20:41 |
| kanzure | so I need some way to separate the interface, the graphing, the svg, user input, and so on, without screwing everything up | 20:41 |
| kanzure | the fact that there's no SVG graphing toolkit kinda sucks | 20:41 |
| kanzure | guess I'll have to start one | 20:41 |
| fenn | there must be one | 20:41 |
| fenn | in javascript? | 20:42 |
| kanzure | http://www.germane-software.com/software/SVG/SVG::Graph/screenshots.html chart::line | 20:42 |
| kanzure | right | 20:42 |
| kanzure | because I want the user interface aspects at the same time | 20:42 |
| fenn | google sez: http://www.liquidx.net/plotkit/ | 20:42 |
| kanzure | like moving the curve via optimization | 20:42 |
| kanzure | yeah, I was looking around google | 20:42 |
| kanzure | a few days ago, don't know if you saw my linkdump | 20:43 |
| fenn | graphing is part of the UI | 20:43 |
| fenn | computers don't read graphs :P | 20:43 |
| kanzure | uhm, crap | 20:44 |
| * kanzure still needs to remember to do a googling support tool | 20:45 | |
| kanzure | even a simple frame with a list of the queries you've done or something | 20:46 |
| kanzure | 'session tracker' was what I called it in 2006 (yeah, I'm lazy) | 20:46 |
| kanzure | hm, I can't seem to figure out whether or not the point of on a graph can be retrieved in terms of points on the screen, from plotkit | 20:56 |
| kanzure | otherwise converting the drag-and-drop dot thing is going to be a pain | 20:56 |
| kanzure | user interactivity too | 20:56 |
| kanzure | heh, macromedia flash mx for svg would be nice, except one that doesn't cost you a leg | 20:56 |
| fenn | what drag-n-drop dot thing? | 20:56 |
| kanzure | fenn: http://heybryan.org/graph.svg notice the giant dot | 20:57 |
| fenn | the blue/white dot? | 20:57 |
| kanzure | the bigger of them, yes | 20:57 |
| fenn | hey i can move it | 20:57 |
| fenn | sorta | 20:57 |
| kanzure | hurray | 20:57 |
| kanzure | yeah, sorta :) | 20:57 |
| kanzure | it's not supposed to go below y=0 | 20:57 |
| fenn | that's neat but it's not exactly robust | 20:58 |
| kanzure | don't read the code, I'm embarrased because of the existing bug that I haven't written | 20:58 |
| kanzure | oh, by no means is it robust | 20:58 |
| kanzure | haven't *fixed | 20:59 |
| -!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has joined #hplusroadmap | 21:03 | |
| -!- jm [n=jm@p57B9DB1D.dip.t-dialin.net] has quit [Read error: 60 (Operation timed out)] | 21:24 | |
| -!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has quit ["Leaving"] | 21:58 | |
| -!- PanGoat [n=pan@ovid.sensoryresearch.net] has joined #hplusroadmap | 23:09 | |
Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!
