--- Day changed Sat Jun 28 2008 | ||
ybit | even the snythesis of it looks laborious | 00:02 |
---|---|---|
ybit | http://www.erowid.org/archive/rhodium/chemistry/adrafinil.modafinil.html | 00:02 |
-!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has quit ["Leaving"] | 00:34 | |
kanzure | fenn: I like that quote from Ed's blog, thanks | 00:44 |
kanzure | re: quantified chemical analysis, http://heybryan.org/instrumentation/instru.html some with explanations and links to diy projects, many not | 00:44 |
kanzure | also, it doesn't include biological instrumentation even though it should :) | 00:44 |
kanzure | hm, now I just need to figure out what to present tomorrow | 00:49 |
kanzure | Eli singled me out on the sl4 mailing list today. :-/ | 00:52 |
fenn | heresy | 00:53 |
kanzure | hm, it's not on the archives yet | 00:53 |
kanzure | basically, Lee Corbin sent an email to the list as if he was an ai and was saying "this is why you should give me all yer keys to the thingies!" | 00:53 |
kanzure | and then I replied with a lengthy analysis pointing out all the bullshit and said | 00:54 |
kanzure | "You're freaking me out if /this/ is what all of your wonderful ideas amount to." | 00:54 |
kanzure | I even included the phrasee lee-i in the message, so it's obvious that I knew it was Lee. | 00:54 |
kanzure | but then Eli comes up and says "This isn't me; I have no idea who. Come on, isn't that obvious from the writing style?" | 00:54 |
kanzure | so he can't read headers | 00:55 |
kanzure | O.o | 00:55 |
fenn | meh | 00:56 |
fenn | is he still doing the 'ai in a box' game? | 00:56 |
kanzure | nah, but the guys on the mailing list are | 00:57 |
kanzure | I really don't know what he's up to | 00:57 |
kanzure | he just gets funding for I don't-know-what | 00:57 |
fenn | is anissimov still in his group? | 00:58 |
kanzure | yes | 00:58 |
fenn | he had some kind of deduction/inference engine | 00:58 |
kanzure | anissimov and I haven't been getting along recently | 00:58 |
kanzure | anissimov and maybe even eli take this "no, only intelligence => singularity" stance | 00:58 |
kanzure | whereas I point out the upper limits on hardware manufacturing as a limit to 'supposed' exponential growth etc. | 00:58 |
kanzure | i.e., you can copy all you want, but you'll need more hdd space :) | 00:59 |
fenn | nah, intelligence can get around so-called upper limits | 00:59 |
kanzure | maximum compression balconies? | 00:59 |
fenn | balconies? what a silly term | 00:59 |
kanzure | basically you're talking about data compression | 00:59 |
fenn | no, but there's plenty of room for that too | 01:00 |
fenn | i mean, we havent come close to exhausting lossy compression | 01:00 |
kanzure | at some point a bit is going to have to correlate to something | 01:00 |
kanzure | so, what, 1 bit per 'intelligence point', ok great | 01:00 |
kanzure | well you still are producing bits at a linear rate | 01:00 |
kanzure | heh' | 01:00 |
fenn | wtf are you on about | 01:00 |
fenn | data and intelligence are orthogonal | 01:01 |
kanzure | let's suppose we have /bin/ai | 01:01 |
fenn | ok and your unstated assumption is it runs on a computer and has no connection to the rest of the world | 01:01 |
kanzure | and let's also, for the sake of simplification, say that it's a scripted language so it's just a script realyl (not both a binary and code) | 01:01 |
kanzure | your unstated alternative assumption is that it will just bitforce its way into the interfaces or something :) | 01:02 |
kanzure | i.e., random combination #334914 now trying to activate hardware component 37 on floor 5 of ip address .... | 01:02 |
fenn | or persuade a human to do rl stuff until it has the means to make its own hardware | 01:02 |
fenn | robots come with documentation, why cant it just read the manual | 01:02 |
fenn | i dont want to just grind through all the yudkowsky examples for you | 01:03 |
fenn | if the AI is worth a damn it can figure out how to run a robot | 01:04 |
kanzure | anyway, back to my example | 01:04 |
kanzure | at some point, you're going to maximize the number of beneficial changes to the code that you can make without adding an extra bit | 01:05 |
kanzure | there's a maximum on the number of bits. | 01:05 |
kanzure | which is constrained by linear manufacturing of those bits (hdd fabber). | 01:05 |
fenn | why linear manufacturing? | 01:05 |
kanzure | I'm referring to what we currently have going for us | 01:05 |
kanzure | now, | 01:06 |
fenn | sure but that's a poor upper limit | 01:06 |
fenn | and even so it's probably more than enough | 01:06 |
kanzure | theoretically, the enhancements to the manufacturing proc can be discovered within those bits of additional uhh usefulness, before it reaches that leveling off at the max number of bits on a hdd | 01:06 |
fenn | what's the total data storage connected to the internet? anyone know? | 01:06 |
kanzure | in 2003 it was est. 161 exabytes | 01:06 |
kanzure | estimated | 01:07 |
fenn | ok i rest my case | 01:07 |
kanzure | not established | 01:07 |
kanzure | unless that's now 161^2 exabytes now ... ;-) | 01:07 |
fenn | enough code to blow up earth many times over | 01:07 |
kanzure | I feel a great disturbance in the market for pizza and coke. | 01:07 |
kanzure | I'm not sure why I was complaining about ai a few moments ago like that | 01:10 |
kanzure | technically there's that upper limit issue, sure, but in reality that's not my big beef with the whole thing of course | 01:10 |
kanzure | there's a few issues that concern me, like the implied "the ai can do this and you can't" and of course their weird Bayes approach to it all ;-) | 01:11 |
kanzure | so | 01:12 |
kanzure | what the hell am I going to be presenting tomorrow? :-p | 01:12 |
kanzure | I know what to talk about and I know how I want to say it, | 01:12 |
kanzure | but just standing up or sitting down and talking isn't a good way to go about things. | 01:12 |
kanzure | should I do the typical stupid power point? | 01:12 |
kanzure | There's a projector available, so I could just go around surfing the net. | 01:12 |
kanzure | and pointing out why something is awesome and how it can relate to the overall projects | 01:13 |
kanzure | I titled it 'automating innovation' so that I can tie in the incubation theory ideas with the artificial attention augmentation system plus some other fun stuff, like, you know, an innovation aggregation project (erm, skdb), etc. | 01:13 |
fenn | sounds like you're making a super duper AI yourself mister :) | 01:15 |
kanzure | and in which case I just need to figure out which links among the many that I'll want to select | 01:16 |
kanzure | not so, according to them the ai means all sorts of statistical inference sillyness and Bayes and Goal Hierarchical Systems and voodoo magic made for dominating the world etc. | 01:16 |
fenn | you will run into the same things while creating an automated innovation machine | 01:16 |
kanzure | why would we need a statistical inference engine | 01:17 |
kanzure | har har har | 01:19 |
kanzure | innovation automation | 01:19 |
kanzure | ia | 01:19 |
kanzure | it's like ai but all backwards. | 01:20 |
kanzure | http://heybryan.org/ia.html | 01:32 |
kanzure | hrm | 01:32 |
fenn | re: societal aspects of innovation, normally i think of 'standing on shoulders of giants' and isaac newton etc | 01:43 |
kanzure | sure, I could throw in some references to that | 01:43 |
kanzure | hm, there's a lot to throw in | 01:43 |
kanzure | I present at noon, but it starts at 10 | 01:43 |
-!- percent [n=percent@66.158.193.25] has joined #hplusroadmap | 01:53 | |
-!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has joined #hplusroadmap | 02:11 | |
-!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has quit ["Leaving."] | 02:11 | |
-!- percent_ [n=percent@66.158.193.25] has joined #hplusroadmap | 03:01 | |
-!- percent [n=percent@66.158.193.25] has quit [Read error: 110 (Connection timed out)] | 03:17 | |
-!- percent [n=percent@66.158.193.25] has joined #hplusroadmap | 03:30 | |
-!- percent_ [n=percent@66.158.193.25] has quit [Read error: 110 (Connection timed out)] | 03:46 | |
-!- percent_ [n=percent@66.158.193.25] has joined #hplusroadmap | 04:33 | |
-!- percent [n=percent@66.158.193.25] has quit [Read error: 110 (Connection timed out)] | 04:49 | |
-!- percent_ [n=percent@66.158.193.25] has quit [" "] | 05:17 | |
-!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has quit [Read error: 110 (Connection timed out)] | 06:23 | |
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has joined #hplusroadmap | 07:17 | |
-!- nsh [n=nsh@wikipedia/nsh] has quit [Read error: 110 (Connection timed out)] | 07:54 | |
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has joined #hplusroadmap | 08:00 | |
-!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has joined #hplusroadmap | 09:49 | |
-!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has quit ["Leaving."] | 10:09 | |
--- Log opened Sat Jun 28 10:40:46 2008 | ||
-!- fenn [n=pz@adsl-76-248-68-62.dsl.bltnin.sbcglobal.net] has joined #hplusroadmap | 10:41 | |
-!- 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/ | 10:41 | |
-!- Topic set by kanzure [] [Sun Jun 8 13:00:15 2008] | 10:41 | |
[Users #hplusroadmap] | 10:41 | |
[ fenn] [ freer] [ nsh] [ Overand] [ procto] [ ybit] | 10:41 | |
-!- Irssi: #hplusroadmap: Total of 6 nicks [0 ops, 0 halfops, 0 voices, 6 normal] | 10:41 | |
-!- Channel #hplusroadmap created Sat Mar 22 15:44:12 2008 | 10:41 | |
-!- Irssi: Join to #hplusroadmap was synced in 0 secs | 10:41 | |
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has quit [Read error: 110 (Connection timed out)] | 13:15 | |
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has joined #hplusroadmap | 13:21 | |
-!- nsh- [n=nsh@d85-194-245-82.cust.wlannet.com] has joined #hplusroadmap | 13:57 | |
-!- nsh [n=nsh@87-94-146-186.tampere.customers.dnainternet.fi] has quit [Read error: 110 (Connection timed out)] | 14:05 | |
-!- willPow3r [n=will@cpe-66-75-6-181.san.res.rr.com] has joined #hplusroadmap | 14:29 | |
-!- nsh- [n=nsh@d85-194-245-82.cust.wlannet.com] has quit [Read error: 110 (Connection timed out)] | 15:06 | |
-!- kanzure [n=bryan@cpe-70-113-54-112.austin.res.rr.com] has joined #hplusroadmap | 16:00 | |
kanzure | I met freenet guy. | 16:03 |
-!- ybit [n=h@unaffiliated/ybit] has quit [Read error: 104 (Connection reset by peer)] | 17:21 | |
-!- ybit [n=h@unaffiliated/ybit] has joined #hplusroadmap | 17:24 | |
kanzure | http://kevinkoym.com/ | 17:26 |
kanzure | Kevin either wants to hire me, work with me, or do something related to brain augmentation. He used to do neurofeedback apparently. | 17:26 |
kanzure | http://www.bootstrapaustin.org/wiki/index.php/Welcome_to_Bootstrap | 18:11 |
kanzure | argh | 18:13 |
kanzure | I'ms till beating my head over about the whole brain visualization stuff | 18:13 |
kanzure | I mean, if I do a 3D model of the brain and just do regions, what good does that do me | 18:13 |
kanzure | don't I need an exact neuronal model of the brain | 18:14 |
kanzure | but at the same time I need to do something that I should have started yesterday | 18:14 |
kanzure | so I'm trying to find that "balance" | 18:14 |
kanzure | the real balance might be in something totally different that I am neglecting somehow | 18:14 |
kanzure | I suspect I made a nerror in reasoning somewhere, and tracing this error is proving to be a pain. | 18:14 |
kanzure | http://www.kosmophone.com/ "The Kosmophone is a gamma-ray spectrometer operating in the range of about 3 to 7 million electron-volts (MeV) controlling a MIDI music synthesizer." in Austin. | 18:21 |
kanzure | $18 125 PSI air tank w/ guage. On the other side of the city. Worth it? | 18:37 |
fenn | how big? | 18:59 |
fenn | if <10gal no, even then you should have a use for it or it will just be in the way | 18:59 |
kanzure | k | 19:35 |
kanzure | it was 10 gal ish | 19:35 |
kanzure | why don't I have a map of the whole brain anyway, even a 'fuzzy map' | 19:35 |
kanzure | this is ridiculous | 19:35 |
kanzure | anyway, on another front, for some reason I ended up writing this 'long' email that I'm about to send out to a number of mailing lists | 19:36 |
kanzure | it's asking others how they go about acquiring useful equipment for cheap | 19:36 |
kanzure | so, you know my autoscholar script | 19:36 |
kanzure | it would be ridiculously easy to write a crawler that goes through a predefined routine always searching the local websites for deals one quipment | 19:36 |
kanzure | *deals on equipment | 19:36 |
kanzure | for instance, craigslist and ebay definitely | 19:37 |
kanzure | and also include some behavioral routines, such as driving through some predefined course and checking a number of stores / locations | 19:37 |
kanzure | the driving part can't be computationally automated obviously | 19:37 |
kanzure | but no reason that you can't be automatically keeping tabs on all of the indices for specific words and so on | 19:37 |
kanzure | right? | 19:38 |
kanzure | How weird. | 19:48 |
kanzure | http://humancortex.alleninstitute.org/has/human/advancedsearch.html | 19:48 |
kanzure | The only two possible locations: temporal cortex or visual cortex. | 19:48 |
kanzure | Uhh. Isn't there more specificity than that in the brain. | 19:48 |
kanzure | Just uhh saying :) | 19:48 |
kanzure | So obviously this is going to end up leading me back to ontologies of the brain, or internal coordinate referencing systems, or something. | 19:50 |
kanzure | Crap. | 19:51 |
kanzure | Hahah. | 21:22 |
kanzure | So I was just sitting in the garage. I found this really neat maroon chair sitting there. So I was there for a few minutes, whatever. | 21:22 |
kanzure | Clicked back to craigslist (I'm outlining some garage sales to hunt for tomorrow, and then dumpsters to try diving in) | 21:23 |
kanzure | http://austin.craigslist.org/gms/735848255.html | 21:23 |
kanzure | That's the chair. | 21:23 |
Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!