--- Log opened Fri Dec 27 00:00:31 2013
01:16 < jcorgan>  /clear
13:13 < Emcy> https://twitter.com/zestyping/status/416570841720438785/photo/1 theyre mental.
13:13 < Emcy> i wonder if its actually a statement about how fuckd up the internet is now
13:13 < Emcy> "welp everyone back to pneumtics"
13:31 < nsh> greenwald keynote starts: http://streaming.media.ccc.de/saal1/native/lq/
13:33 < TD> thanks
13:37 < BlueMatt> nsh: thanks
13:37 < nsh> np
15:31 < goedgoed> nsh: Shit. CCC totally slipped my mind. Thanks for reminding me!
15:33 < nsh> np
16:46 < Emcy> greenwald isnt do ing anything particularly amazing or insightful
16:46 < Emcy> hes just telling it like it is, according to occams razor mainly
16:47 < Emcy> unfortunately the weave of lies and obfusication of whats really going on has run so deep for so long, when he opens his mouth it feels like a lungful of air after a freedive
16:48  * nsh nods
16:57  * daira2 nods too
17:33 < gmaxwell> Went out to lunch a bit ago, heard random people talking about dogecoin.  The person promoting it was proposing it as a fun way to try out cryptocurrency.
17:34 < nsh> whatever happened to experimenting with psychedelic drugs and rock music :/
17:35 < gmaxwell> I wonder if we should rebrand testnet as Courage Wolf coin.  "Bitcoin too stable and secure.  Use testcoins!"
17:36 < nsh> go full hog and gamify the whole system of cryptocurrency experimentation
17:37 < nsh> it'll be like spore but with a lot more hashing
17:37 < gmaxwell> well, my point there is that if you really want coins for futzing around with, testnet is better than some total alt, since at least testnet is guarenteed to track development.
17:38 < Emcy> serioualy isnt dogecoin only about 3 weeks old
17:38 < Emcy> either that was very improbable or cryptocoins in general are getting some wicked mindshare
17:39 < nsh> well, doge represents the intersection of cryptocurrency and inane internet sensations
17:39 < nsh> which have a much wider and more rapid mindshare proliferation function
17:40 < Emcy> we havent had a really good meme since the great meme collapse of 2011
17:40  * nsh smiles
17:40 < Emcy> when facebook kids started unironically posting the meme templates everywhere and then arrow knee blew the lid off the whole thing
17:40 < gmaxwell> Emcy: I mean, I am in silicon valley, hearing people talk about bitcoin in public is basically a daily event.
17:41 < Emcy> oh cool didnt knew you went there
17:41 < Emcy> that skews the probabilities somewhat
18:00 < pigeons> i know people who have never used bitcoin and aren't interested in it but are trying dogecoin for fun and to buy steam games. i don't get it, but people like it for some reason
18:01  * andytoshi-logbot is logging
18:02 < Luke-Jr> gmaxwell: update topic?
18:02 < Luke-Jr> andytoshi-logbot: got a link for topic? :P
18:02 < andytoshi-logbot> I'm logging. I don't understand 'got a link for topic? :P', Luke-Jr.  Try /msg andytoshi-logbot help
18:02 < andytoshi> lol
18:03 < andytoshi> Luke-Jr: i need an ack from gmaxwell .. and i'm not sure i'll get one as long as people are talking about bitcointroll users candidly here :P
18:05 < andytoshi> the last thing we need is altoz posting "see, andy and luke really are out to get me!"
18:05 < Luke-Jr> haha
18:08 < nsh> i will happily volunteer to be a few order of magnitude more offensive than anyone else to deflect log-flac
18:08 < nsh> (just one of my many services)
18:08 < Luke-Jr> lol
18:10 < gmaxwell> logging is fine, but something more reliable than andytoshi-logbot should be doing it if we're going to have public logs. :P
18:10 < gmaxwell> also, I think the topic is public editable in here.
18:10 < andytoshi> aww, you don't trust the perl script i downloaded and ran without reading?
18:11 < gmaxwell> andytoshi: I mean I see your bot bouncing in and out a lot.
18:11 < andytoshi> i know, it is supposed to detect disconnects and come back, but it doesn't
18:12 < andytoshi> i'll spend some time working on it over in #andytoshi..
18:17 < sipa> dogecoin even still uses irc seeding...
18:17 < sipa> or rather, again
18:22 < gmaxwell> sipa: I assume it's a fork of the pre 0.8 litecoin codebase?
18:24 < Luke-Jr> gmaxwell: it's not (topic public edit)
18:25 < warren> gmaxwell: it's 0.6 based
18:27 < andytoshi> ok, i see, the logger's "reconnect" code just deletes the PID file and assumes somebody else will respawn it
18:27 < andytoshi> i'll give it its own user, write some systemd unit files and do this properly..
18:52  * andytoshi-logbot is logging
18:52 < andytoshi> ok, it shouldn't disconnect for more than a few secs from now on
18:53 < gmaxwell> andytoshi: Is your CJ thing available as a tor hidden service?
18:54 < andytoshi> oh! gimme five minutes..
18:55 < gmaxwell> yea, part of why I asked was because its so easy to setup. :P
18:58 < andytoshi> http://xnpjsvp7crbzlj3w.onion/
18:59 < andytoshi> let me just confirm that none of the links assume it is under domain.net/coinjoin/ ...
19:01 < andytoshi> cool, all good
19:06 < gmaxwell> andytoshi: cool. it works, you might want to put both links on the bottom of the page.
19:10 < andytoshi> done
19:11 < andytoshi> now i'll go spam anontalk with it..
19:12 < gmaxwell> Whats anontalk?
19:12 < andytoshi> i don't think that's what it's called .. there used to be an anonymous board at http://ci3hn2uzjw2wby3z.onion/
19:12 < andytoshi> maybe it has gone down
19:13 < gmaxwell> I guess the next missing piece for your tool is an auto-participater. e.g. something that polls periodically and if there is a open CJ of the right size, it participates for you.
19:15 < sipa> TD: what timezone are you in? :)
19:17 < TD> currently
19:17 < TD> GMT
19:17 < TD> sipa: why?
19:18 < sipa> not used to seeing you join at this time :)
19:18 < andytoshi> gmaxwell: yeah, that'd be cool
19:19 < andytoshi> right now i'm working on having any coinjoins going on, when i'm not 50% of the participants
19:20 < gmaxwell> andytoshi: yea well part of the challenge, of course, is that when someone uses it there may be no one else, or /worse/ just someone with a non-match that doesn't really increase their privacy.
19:29 < andytoshi> yeah, that happened to my this morning..
19:29 < andytoshi> ;;cjs c11fc9bd5b462946
19:29 < gribble> Coinjoin Status: session ``AMEMB Lanceros PFS Sex Tess IDB 15kg'' is completed. The submitted transaction ID was 80819213bb25df35f890fab55f8d3b71c8f5bed3b823bb949ce26e1471686e61.
19:29 < andytoshi> https://blockchain.info/tx/80819213bb25df35f890fab55f8d3b71c8f5bed3b823bb949ce26e1471686e61 i was the 0.2's
19:29 < andytoshi> it clearly said, "most popular output is 0.2. use that output size"
19:32 < gmaxwell> might be better if the person starting a join could mandate a size?
19:34 < andytoshi> maybe, i dunno, i don't want to make it too irritating .. i am already disappointed by the ~0 uptake
19:35 < gmaxwell> andytoshi: yea, sorry about that. I could have warned you.  See also the ~0 uptake in PT's dust-be-gone.
19:37 < andytoshi> lol, s'fine, i learned a lot from this
19:37 < andytoshi> and i found a bug in the rust compiler, so my github account claims i have "contributed" to rust, which is cool
19:40 < gmaxwell> My expirence with dust-b-gone suggests that even with an automated participater the usage will be low. I'm not sure what it takes to get it used.
19:41 < andytoshi> calling it doge <.<
19:41 < gmaxwell> I ... worry... that there are basically few actual cryptocurrency people in Bitcoin.
19:42 < sipa> is there anything coinjoin-related already usable/released?
19:42  * sipa has hardly followed up recently...
19:43 < andytoshi> i worry about this too, but then i remember that there are like 30-40 serious people here, and we are able to connect on #bitcoin-wizards and exchange research in a way that would've been impossible even 10 years ago, even if we'd had bitcoin back then
19:43 < andytoshi> well, irc was around 10 years ago, but i don't think the preprint archives were, nor do i think academics spent a lot of time on public forums
19:44 < andytoshi> sipa: i have a joiner at http://xnpjsvp7crbzlj3w.onion/ which is "usable" if you can deal with rawtx's
19:45 < sipa> ok
19:46 < andytoshi> one of maaku's 17 million projects is an automatic p2p joiner
19:46 < andytoshi> but idk if that's usable right now
19:48 < sipa> well, i don't think you can't expect any measurable uptake without any serious wallet application having integration or even automatic using it
19:49 < pigeons> are the sessions submitted at https://www.wpsoftware.net/coinjoin/ shared with http://xnpjsvp7crbzlj3w.onion/ ?
19:51 < pigeons> since there is such low usage i would want to submit to the one that gets the higher chance of someone else submitting a transaction if not
19:51 < andytoshi> pigeons: yes, they are the same site
19:51 < pigeons> thanks
19:51 < andytoshi> the .onion gets routed to the wpsoftware.net one at my tor node
19:52 < andytoshi> (and the webserver is on the same hardware as my tor node)
19:53 < maaku> andytoshi: haha yeah i got way too much going on right now
19:54 < maaku> i need to focus on finishing just one of them
19:54 < maaku> i'm reworking the protocol messages to include multiple bucket sizes and explicit fees
19:55 < maaku> but that stalled while i was working on the utxo validation index bips
19:55 < maaku> i'm going to work the python proof-of-concept to the point where you can do joins from the command line
19:56 < maaku> and I got an offer from someone else to handle the messaging via bitmessage + tor
19:56 < maaku> but after that, I'd rather see it reworked into C++ and integrated into (a fork of) the reference client directly
19:56 < andytoshi> oh, nice
19:59 < maaku> i'm hoping that the work i'm doing will be the foundation of a future protocol extension everyone uses
20:00 < maaku> but i have no illusions of it happening quickly :)
20:00 < nOgAn0o> SEND BTC FOR 50% GUARANTEED PROFIT.. EMAIL ME AT NOGANOO@LIVE.COM WITH YOUR ORDER!  I HAVE AN EXCHANGE EXPLOIT!   1CL67LZ94WUExLe9ZpKZfFMFJKwVEZqyDM
20:07 < nsh> really?
20:08 < nsh> (gmaxwell)
20:09 < pigeons> if i send you btc isn't that 100% profit for you rather than 50%?
20:09 < sipa> who has op here?
20:13 < nsh> gmaxwell has done all the +b that i've seen
20:14 < nsh> ty
20:19 <@gmaxwell> I was the only +o in here, but I've now added +o to petertodd amiller adam3us sipa warren maaku jgarzik Luke-Jr  (top talkers in here)
20:19  * maaku fantasy power trips
20:23 < nsh> ULTIMATE POWERRRRR
20:25 < gmaxwell> andytoshi: you might want to make the there is no current messages include a "Go to https:// to start one."
20:27 < nsh> +1
20:28 < andytoshi> oh, hey, that's a great idea
20:29 < andytoshi> there we go
20:29 < andytoshi> (i'm going to do all my testing on #bitcoin from now on for advertising purposes)
20:29 < andytoshi> so far i've netted -0.0006btc on this joiner, by participating in pretty-much every join :P
20:31 < nsh> ballads will be sung for generations to come of your entrepreneurial acumen :)
21:08 < Luke-Jr> andytoshi: what testing? #bitcoin is explicitly non-logged FYI
21:10 < andytoshi> Luke-Jr: coinjoin
21:10 < andytoshi> log testing i do on #andytoshi :P
21:10 < andytoshi> but thx for the heads up
22:34 < gmaxwell> T-25 minutes for andytoshi's coinjoin. Time to prep your transactions if you're joining.
22:34 < andytoshi> 0.5 btc outputs
22:35 < andytoshi> i guess i should prep mine..
23:01 < gmaxwell> ;;balance 1ForFeesAndDonationsSpendHerdtWbWy
23:01 < gribble> 5.46e-05
23:02 < andytoshi> lol
23:03 < andytoshi> does anyone know who did that?
23:03  * nsh is confused
23:04 < gmaxwell> told you that you should have made it a spendable address!
23:04 < nsh> oh
23:04 < gmaxwell> ;;cjs
23:04 < gribble> Coinjoin Status: current session is open for 16 more minutes. There are currently 3 transactions in the pot. The most popular output value is 0.5.
23:06 < nsh> what's the current 'accessibility' of coinjoin?
23:06 < gmaxwell> andytoshi: ^ that should probably say max(count(most_popular),ntransactions) to avoid disclosing the number of players when it wouldn't be obvious from the inputs.
23:06 < gmaxwell> nsh: you can use andy's thing if you can spend via a raw transaction.
23:07 < gmaxwell> It's actually slightly safer than a normal raw transaction, since it prevents to common all coins to fees failure mode.
23:07 < andytoshi> hmm
23:07 < nsh> how hard would it be to let any-random-noob perform the rawtx spend safely?
23:08 < nsh> (i assume it would be better for privacy is the barrier-to-entry for coinjoin was as low as possible)
23:08 < nsh> or bc.i users can do it now?
23:09 < gmaxwell> nsh: yea, ultimately there needs to be dumb wallet integrated tools. But getting there requires more expirence with the technology, so tools like andy's n00b unfriendly one are a stepping stone.
23:09 < nsh> right
23:09 < gmaxwell> nsh: bc.i has something they're calling "coinjoin" which is really only kinda coinjoin. As it depends on you trusting bc.i to do the right thing.
23:09 < nsh> (wasn't being critical in any way, just wondering how to increase the utility)
23:09 < nsh> but i guess not trusting bc.i much more than people already do?
23:10 < gmaxwell> (really part of the premise I had in promoting this style of txn is that we can't really get wide adoption if its predicated on additional trust, because the trust is a cost too)
23:10  * nsh nods
23:10 < gmaxwell> yea, if you're already using Bc.i you're already exposed, it's not really too much worse.
23:10 < nsh> right
23:11 < andytoshi> gmaxwell: what should the display say if i'm actually publishing max(count(most_popular),ntransactions) ?
23:11 < andytoshi> "there are something like 3 transactions in the pot"
23:12 < gmaxwell> andytoshi: "there are ~N transactions in the pot"
23:13 < nsh> the awesome liability-reducing power of the tilde
23:13 < nsh> :)
23:28 < andytoshi> holy shit, these cj's get confirmed fast
23:28 < andytoshi> less than a minute this time
23:29 < nsh> there's good marketing for you :) "want faster confirmations *AND* increased privacy? use coinjoin!"
23:29 < gmaxwell> andytoshi: I guess you should share a link to the txn in #bitcoin — worse the slight loss in privacy to show people that its real.
23:30 < nsh> (just add a tilde if it's not actually faster on average)
23:31 < andytoshi> i've got to jack up the fees required fee tho, that time the fee was 0.00035786
23:31 < andytoshi> and i only demanded 0.00024 from people, so it's possible that i wound up paying most of that myself
23:32 < andytoshi> :S
23:33 < nsh> charge slightly over the odds and disburse the difference as a faucet or something
23:33 < nsh> maybe
23:33 < gmaxwell> andytoshi: set fees to whatever sane value you think they need to bet to get people to use it, I'll pay you out of the CJ bounty fund (or, if the other signers don't agree, out of my own pocket) later.
23:34  * nsh nods
23:35 < andytoshi> well, i'm not too concerned about the personal loss, but rather what happens when i'm not involved with a join
23:35 < andytoshi> i suppose i could attach a faucet..
23:36 < andytoshi> right now, the only person outside of this channel i've heard comment on the fees said they were "practically nothing"
23:37 < gmaxwell> really the problem with the fees is that they dork up going from round valued outputs to round valued outputs.
23:37 < andytoshi> yeah, that's really irritating
23:37 < nsh> could you have a dummy input and output in every join that soaks up the fee, ehm, jaggedness?
23:38 < nsh> (from some holding wallet run as part of the service)
23:38 < nsh> no, that doesn't make sense
23:38 < gmaxwell> right. :P
23:38 < nsh> shh, it's late
23:39 < gmaxwell> andytoshi: could do that if he basically gave people fee tokens. e.g. send in some coin to andy and he gives you a fee token, and then you can use that in multiple txn to pay your fees (meaning andy just pays them). But it's a lot of complexity— too much for a simple manual process.
23:39 < andytoshi> well, if it was always us, i could do something like that, since i trust that people here would pay up if i asked
23:39  * nsh nods
23:39 < andytoshi> i definitely don't want to add complexit
23:39 < nsh> you even simplified the word!
23:40 < nsh> :)
23:40 < andytoshi> for now i bumped the fee up from 8000 to 10000 satoshi, since that's rounder ;)
23:40 < gmaxwell> yea, but I suspect we'll not learn more if it's just us.  What I think we should try doing is these daily ones for a few days and see if we get any more players.
23:40 < andytoshi> and in this case, we would have paid the minimum network fee even if everyone had only given 10k sat
23:40 < nsh> i wonder if you could make it a wee bit "gamier" to entice people
23:41 < nsh> (not quite gambling, but some chance element that adds 'fun')
23:41 < andytoshi> it's tough without increasing complexity .. i could do something like give all the donations to a random participant
23:41 < andytoshi> but then they'd have to provide an additional address alongside the rawtx
23:41 < gmaxwell> well and then you'll un round one of my pretty round coins. you bastard. :P
23:41 < nsh> but can that be done without trusting that you aren't getting backhanders to pick certain people to win?
23:41 < andytoshi> :P
23:42 < andytoshi> nsh: lol nope
23:42 < gmaxwell> nsh: not without making it more complex.
23:42 < nsh> is there a way to add some randomness to dispersal in script
23:42 < nsh> i thought there was an OP that gave a random bit...
23:42 < nsh> as an artefact of something or other
23:42 < gmaxwell> there isn't but there are ways to do that but not without making it more complex.
23:42 < nsh> ok
23:43 < gmaxwell> part of the point is that these txn should be generally indistinguishable from ordinary ones— except perhaps that they have many equally sized outputs.
23:43  * nsh nods
23:43 < gmaxwell> so that they're hard to exclude from tracing tools, and if a tracing tool starts excluding them, it'll be easy to make 'fake' CJ transactions.
23:43 < nsh> hmm, how would that help?
23:44 < andytoshi> (if NSA excluding these from analysis, then you can get all your ordinary transactions excluded just by spending to several outputs -- then you win)
23:44 < gmaxwell> nsh: because even without participating in an actual CJ you can form a txn all by yourself that looks like one (a bunch of equal sized outputs) and then various automatic deanonymization methods would hit it and fire of their CJ huresitic and give up.
23:45 < gmaxwell> yea, as andytoshi says.
23:45 < nsh> right
23:45 < nsh> though in practice it might rather go as "okay all of these guys are definite terrorists. *dronestrike*"
23:45 < gmaxwell> well different threat model.
23:45  * nsh nods
23:46 < gmaxwell> Either dumb web tools that automatically trace coins frequently gives BS results from CJs in which case they're easily debunked and few trust them... or they ignore CJs and you just make some fake ones and basically opt out of their tracing. win win.
23:46 < gmaxwell> NSA .. I can't help you with. You're probably screwed. :P
23:46 < andytoshi> if SR were still up i'd be sending all the donations there :P
23:46 < gmaxwell> andytoshi: you could send them to the FBI. :P
23:46 < andytoshi> hahaha
23:46 < nsh> people are still "sending" money to that addresss...
23:46 < nsh> but that's another subject
23:47 < gmaxwell> whatever jackass hacked john dillon sent a bunch of btc there that he'd sent me in a private key for the CJ fund. :(
23:47 < nsh> sucks :(
23:54 < andytoshi> ok, i tried doing the max(unsigned tx count, mpo count) thing
23:54 < andytoshi> we'll know in a minute or so if it worked..
23:54 < gmaxwell> andytoshi: perhaps the count should actually be min(distinct_in_addresses,max(ntransactions,n_most_pop_outputs))  .. otherwise on this one it would have displayed 10 which was clearly impossible vs 9 which is at least more credible.
23:55 < gmaxwell> I suppose there really should be some maximum_credible_amount  which does some value analysis.
23:57 < andytoshi> hmm
23:57 < andytoshi> the point of this display is to give people a swag of their anonymity
23:58 < gmaxwell> right, but e.g. if I submit to you a txn that itself looks like a coinjoin, e.g. two addresses each with enough to form a uniform output, the display shouldn't leak that.
23:58 < andytoshi> ah, i see
23:59 < gmaxwell> maybe it should work from a pure analysis of the transaction submitted so far, just some metric some attacker might use to guess the participants.
23:59 < andytoshi> i think i'll modify coinjoin to calculate how many participants it thinks its merging
--- Log closed Sat Dec 28 00:00:01 2013