--- Log opened Sat Dec 21 00:00:15 2013 01:12 < eristisk> /msg NickServ IDENTIFY iojfdys!df9876ds7%% 01:13 < eristisk> oops, better change that. 01:15 < _ingsoc> Lmao. 01:16 < eristisk> *^_^* 01:52 < Emcy> wow RSA took $10 measly MM of NSA cash to deliberately gimp thier crypto 01:53 < Emcy> Had a long chat with my friend who worked on RSA's BSafe. Two comments: "i saw that this morning and was filled with a sense of shame" 01:53 < BlueMatt> yea, see the thing I find the most surprising in that story is it only took $10 million... 01:53 < Emcy> But no, he had no idea NSA had done some payoff and he was working on code apparently deliberately gimped. ;( 01:53 < Emcy> hmm those are quote tweets, i jave no RSA friend 01:54 < Emcy> BlueMatt yeah its not a lot of money for flushing your company rep down the toilet is it. NSA total budget for this sort of aubterfuge is 250MM apparently 01:56 < Emcy> I SINCERELY hope the word gets out about this far and wide and the market sorts this one out 02:36 < Dylan_> has anyone heard news about truecrypt's independent audit? 02:47 < Emcy> theyve engaged an audit firm 02:47 < Emcy> what i dont know is whats stopping that firm getting a nice fat NSL about it 03:44 <@gmaxwell> Personally I'm waiting to hear about the lawsuits from ex-NSA people who are now unemployable. 04:01 < Emcy> they picked thier side 05:30 < maaku> Dylan_: http://blog.cryptographyengineering.com/2013/12/an-update-on-truecrypt.html 09:03 < adam3us> some scroll-back comment: petertodd or maaku were talking about how the minimal function needed from the network is tx ordering (if you ignore SPV functionality). i agree with this. the fact that committed tx are respendable in committed form to full nodes is in fact an illustration of this fact. 09:07 < adam3us> in some way you can see that the distributed function offered by bitcoin network if you remove tx validation (as respendable committed tx do) is that it is a secure namespace. (first come first served, first to announce owns; and can opt to transfer ownership. transferring ownership decommits because it involves a signature. anyway a distributed namespace is a slightly higher level function built on a distributed timestamp. 09:10 < adam3us> and in fact again, other than for optimization, full nodes could survive fine with a distributed timestamping service only (no name uniqueness guarantee) as the timestamping defines ordering. they can therefore build first to publish just by ignoring later republications (and validating themselves either via committed tx key knowledge, or by validating clear text but unvalidated tx) 11:36 < petertodd> adam3us: just ignoring later publications isn't good enough though, you need to be able to be sure that a prior publication *doesn't exist* at all 11:37 < petertodd> adam3us: if you can't, then you're not sure if your coins are valid 11:37 < petertodd> adam3us: assuming bitcoin-style fixed inflation that is... 11:38 < petertodd> adam3us: you could have a system where double-spends were valid if accompanied by more pow of cource 11:38 < petertodd> *course 11:39 < petertodd> adam3us: basically the scheme I came up with back in highschool for a decentralized crypto-currency after reading a certain paper about hashcash :P 11:39 < petertodd> adam3us: dunno if you've heard about it, you do a partial-preimage against... :P 12:02 < adam3us> petertodd: yes. i meant to imply the full node scans from genesis and is thereby convinced that a given string is the first copy 13:55 < maaku> adam3us: "if you ignore SPV functionality" <--- that's a big thing to ignore 13:56 < maaku> it's the difference between academic wankery and a system that is actually deployable and workable 13:57 < petertodd> maaku: this is -wizards, we're doing research and development 14:04 < andytoshi> maaku: i think if you're aware of the simplifications (and ofc petertodd is), spherical blockchain reasoning is useful 14:05 < andytoshi> eg bitcoin solves problems independently enough that you can think of timestamping apart from everything else 14:07 < petertodd> anyway, with additional technology you probably can make such systems usable on low-resources too, for instance via SCIP to compactly validate coin histories, or via economic tricks to limit the scope of fraud 14:12 < CodeShark> I wish more of petertodd's ideas were being tried in practice :) 14:13 < petertodd> CodeShark: same :P 14:13 < petertodd> CodeShark: it'll be nice finally getting some free time to work on them properly soon 14:18 < CodeShark> need any help? 14:18 < CodeShark> or rather, would you like some help? :) 14:18 < petertodd> yes! 14:18 < petertodd> although frankly, I think right now coinjoin is what needs dev effort on the most 14:19 < petertodd> other stuff is cool, but it really needs to actually see implementations 14:19 < CodeShark> not sure whether stronger privacy or blockchain/utxo prunability are a higher priority 14:20 < adam3us> maaku: yes SPV is the current scaling model. i like to re-examine assumptions. sometimes i find ways to re-write them along the way. so thinking back to the minimal function for the global part is good. more secure even. and then try other ways to scale maybe there are better ways. 14:20 < petertodd> it's easier to throw hardware at making bitcoin scale than it is to throw hardware at making bitcoin private 14:20 < adam3us> maaku: eg commitd tx have strong policy advantages over clear/validated tx 14:21 < adam3us> maaku: more resistant to centralization for example 14:21 < CodeShark> throwing more hardware at making bitcoin scale seems to encourage greater centralization, though 14:22 < adam3us> maaku: anyway the comment was part of some wide-ranging what-ifs i tried to isolate the dependency bitcoin puts on mining, and it turns out there are multiple entangled reasons 14:22 < petertodd> yes, but bitcoin will easily survive having transactions gradually become more expensive 14:23 < nsh> "There are levels of survival we are prepared to accept." -The Architect. 14:23 < petertodd> nsh: transactions already actually cost like $50 each; fees can go up a hell of a lot 14:24 < CodeShark> $50 each?!?!? 14:24 < nsh> hmm 14:24 < CodeShark> are you talking about international wire transfers, petertodd? :) 14:24 < petertodd> CodeShark: yup. total new bitcoins created out of thin air * $/BTC / # of transactions = $50 14:24 < adam3us> there are multiple paths to policy neutrality: actual decentralization, moderately central nodes having insufficient info to do policy (committed tx) 14:25 < adam3us> petertodd: yeah but thats a point in time description allocating all of reward to tx fees; as number of tx increases and reward decreases the cost/tx falls 14:25 < CodeShark> petertodd: I don't follow - block rewards don't cost the parties transacting bitcoins 14:26 < CodeShark> it has a small inflationary effect, perhaps 14:26 < CodeShark> but that affects everyone 14:26 < petertodd> adam3us: sure, but the point is the *economics* are such that bitcoin works at a real cost of $50/tx, which implies that the core usage of bitcoin is as a store-of-value/speculation 14:27 < petertodd> adam3us: maybe it'd start to get ugly at $10/tx, but we can certainely survive $1/tx 14:27 < CodeShark> petertodd: so you're saying that each transaction spreads a cost of $50 amongst all holders of bitcoins? 14:28 < adam3us> petertodd: seems like the reward is the reward, its just a distribution mechanism/bootstrap mechanism. i dont see a reason to equate it to tx cost at current tx rates 14:28 < petertodd> CodeShark: no, I'm saying the cost to run the whole bitcoin system is $50/transaction 14:28 < maaku> petertodd: that's saying subsidy is $50/transaction 14:28 < CodeShark> petertodd: who foots the bill? 14:28 < petertodd> CodeShark: that's not to say the *marginal* cost of a transaction is $50, but it strongly suggests that much higher fees are economically feasible 14:28 < adam3us> petertodd: the supposition is that % of income from mining crosses over as tx # increase so that fees take over as reward tapers 14:29 < petertodd> adam3us: exactly, and given the system functions just fine with a huge fixed cost, making that into a marginal cost is likely fine to a first approximation - my main worry is actually off-chain systems being *too good* and not supporting miners enough 14:30 < petertodd> adam3us: but we probably have ~10 years before that's a big deal... 14:30 < adam3us> so far no one made a remotely plausible off chain anything other than TDs micropaymens channel but thats point to point so its just a way to avoid aborts on a tab 14:31 < petertodd> adam3us: micropayment channels are *not* off-chain, don't call them that 14:31 < maaku> adam3us: freimarkets 14:31 < petertodd> adam3us: and I'd say fidelity bonded banks, especially w/ trusted hardware, are perfectly plausible, they just won't happen unless fees make them happen 14:31 < maaku> but i think you have some confusion over what off-chain is 14:31 < adam3us> maaku: isnt freimarkets on chain (on freicoin or other coin) 14:32 < adam3us> petertodd: there you go that is off-chain 14:32 < petertodd> adam3us: I think you a word 14:32 < maaku> adam3us: private accounting servers (with atomic transfers with the public chain, including bitcoin) are part of the spec 14:32 < adam3us> maaku: off-chain is like not on-chain ;;) 14:33 < maaku> and i'd count open transactions too 14:33 < adam3us> maaku: yeah you could say chris odom open transactions is focussing on off-chain 14:33 < adam3us> maaku: the problem is all the off-chain stuff i've seen loses fundamentaly 1 or 2 important and useful bitcoin functions 14:34 < maaku> well the key part is how value is moved on and off chain 14:34 < maaku> chris only figured that out with his "holy grail" voting pools 14:34 < maaku> which still aren't implemented, i think 14:34 < maaku> adam3us: if it didn't lose bitcoin functionality, it'd replace bitcoin entirely 14:35 < adam3us> maaku: or what properties are left once you have the coin in some offchain situation. eg what OT tokens backed in bitcoin? thts not going to be as secure, nor distributed etc 14:35 < petertodd> adam3us: and to that, so what? losing my $100 morning coffee slush funds every once in a while isn't a big deal 14:35 <@gmaxwell> :( https://github.com/spesmilo/electrum/issues/512 14:36 < maaku> let me rephase ... you can't expect an off-chain solution to be better or equal to bitcoin in every way, or else it will be strictly speaking better (off-chain scales better), so what are we doing? 14:36 < petertodd> gmaxwell: huh? every time you make a paymen tto an address it goes into the "used" bin and gets hidden 14:37 < adam3us> maaku: well itd be nice to minimize the feature loss offchain. maybe its possible to not give up anything even. we can at least try with that objective 14:38 < petertodd> adam3us: don't let perfect be the enemy of good enough 14:38 < maaku> adam3us: well i'm on board with that. recognizing that the goal is something we weill probably never achieve (and if we did we'd replace bitcoin entirely) 14:38 < adam3us> maaku: eg say btc gets to minimum amount of $10k on chain, perhaps a solution is multiple side-chains and atomic swaps into the main chain for example 14:38 < maaku> but shoot for the moon and you'll at least land among the stars 14:39 < adam3us> maaku: i just like to understand clearly the requirements (rather than think in terms of the artefacts of the current system) not all of the artefacts may be actual fundamental limitations 14:39 < maaku> who knows what the minimal, least impactful features are that we'd have to give up, so might as well try to keep them all 14:41 < CodeShark> atomic swaps would also permit a fully decentralized cryptocoin exchange :) 14:41 < adam3us> maaku: yeah well so far all my design rejigging attempts ended up making something worse, its definitely hard; seems like bitcoin only-just-works, and its multiple features so inter-dependent on mining its hard to modify anything 14:42 < maaku> CodeShark: what's the value of having more than 1-2 decentralized currencies? :P 14:42 < adam3us> CodeShark: this is true; somewhat. you also need script extension to have non-stalling (otherwise people will stuff the order book to manipulate price with cryptocoins they have no actual intention of selling). ie so you can take the ask, by definition by satisfying its price 14:43 < CodeShark> maaku: there are different use cases where different features might be more/less desirable - economic parameters, confirmation times, etc... 14:43 < adam3us> maaku: i think one digital scarcity definition (bitcoin) is the limit, 14:43 < adam3us> CodeShark: they are mostly excuses for me-too-coins aka pump & dumps with no transactions and so no intrinsic value 14:44 < maaku> CodeShark: I challenge you to come up with one real example that isn't better served by some other off-chain solution 14:44 < petertodd> gmaxwell: oh hang on, just tried that myself... weird, recv addr list not getting repopulated, yeah, that's a WTF 14:44 < petertodd> gmaxwell: recent bug I think 14:44 < adam3us> bitcoin-staging with 1:1 peg (as discussed a few days ago by BlueMatt & gmaxwell) is the answer IMO 14:45 < CodeShark> adam3us: I'm familiar with that argument - and while true that all the alt coins are essentially bitcoin ripoffs, I see it differently - I think the parameters Satoshi chose for bitcoin are completely arbitrary - what's not arbitrary is the block chain concept as a decentralized timestamping mechanism. why should we get stuck on a specific set of arbitrary parameters? 14:45 < adam3us> anyone not doing that needs their pump & dump sabotaging financially of via mining difficulty attacks 14:45 < maaku> adam3us: in my view changing the nature of the decentralized money is the only valid reason to try a different coin (like we've done with freicoin, and I'm sure there are other possible variations) 14:45 < maaku> but changing interblock time, proof of work algorithm, subsidy algorithm, etc. has ~zero real world benefit 14:45 <@gmaxwell> CodeShark: they aren't completely arbritary, as alts have been created which only changed the "arbritary" ones and turned into fireballs as a result. 14:45 < adam3us> maaku: yes freicoin actually and namecoin are not param tweaks 14:46 < adam3us> CodeShark: yeah satoshi clearly but extensive modeling into the params; pretty much all alts are outright worse 14:47 < CodeShark> adam3us: Satoshi is the wright brothers and bitcoin is the first powered airplane. 14:47 < petertodd> CodeShark: and our job as bitcoin developers is to upgrade that airplane to a modern Boeing 787, without landing 14:48 < andytoshi> gmaxwell: i saw that comment on #bitcoin too, i left 14:48 < andytoshi> idk how you can tolerate so much of that channel at once 14:48 < nsh> my secret is copious consumption of crack cocaine 14:48 < petertodd> andytoshi: my contacts at the vatican say greg's getting canonized when he kicks the bucket 14:49 < maaku> CodeShark: you know the wright brothers spent years in their private wind tunnel perfecting their airplane before it ever flew ;) 14:49 < adam3us> CodeShark: like i said i think param tweak alts that try to start a new race are pump & dumps; and if one did come along that got real transactions, it would rise to instead be dangerous to the confidence in digital scarcity which is too valuable a new concept jeopardize with toy pump & dumps 14:49 < CodeShark> maaku: the wright brothers didn't even understand swept wing designs or the fundamental subsonic limitations of propellers 14:50 < CodeShark> point is, Satoshi surely missed many things, too 14:50 < adam3us> CodeShark: thought experiment: if ltc got real transactions, overtook btc in market cap, and then perhaps btc users dump btc to buy ltc causing a btc price crash; then ltc users notice ftc catchng up with its market cap - next thing you know people lose confidence in the asset class of digital scarcity and turn the whole thing into a digital tulip 14:51 < adam3us> CodeShark: I dont want that outcome. alts must die. use bitcoin-staging to try useful new params or features. 14:51 < CodeShark> anyhow, I'm not going to get sucked into a religious debate 14:52 < CodeShark> alts are inevitable - we must learn to cope with them. they won't die 14:53 < adam3us> CodeShark: i am just saying alts are stupid and maybe even dangerous. 14:53 < CodeShark> the same decentralized nature of the technology which makes bitcoin so hard to kill makes alts hard to kill 14:53 < adam3us> CodeShark: I think most of the pump & dumps will die soon enough. they have no intrinsic value because there are no transactions. eventually they die 14:54 < adam3us> CodeShark: bitcoin has first mover advantage - big intrinsic value, infrastructure and stored-value; the alts are abuses trying to make-money-fast with param-tweaks, 99% of them. 14:54 < CodeShark> the core technology is agnostic to these parameters - the core technology consists of a decentralized timestamping mechanism using proof-of-work 14:55 < CodeShark> you can think of alts as simply param tweaks on bitcoin - I see it a different way - I would like to see the decentralized timestamping mechanism Satoshi invented applied to many problems 14:55 < adam3us> CodeShark: if you like analogies its like the wrong brothers came along and cloned the write brothers plane and painted it blue and then tried to claim they invented or profit from the wright brothers work 14:56 < CodeShark> I think the original bitcoin client isn't sufficiently modular and flexible 14:56 < adam3us> CodeShark: "look at my blue plane".... its "BLUE" so its like cool and stuff, please mine it and make me money fast :) 14:56 < adam3us> CodeShark: so work on making it modular and flexible 14:56 < CodeShark> I have been :) 15:01 < CodeShark> adam3us: my motivation here is not making money fast - my motivation is seeing this technology evolve 15:02 < adam3us> CodeShark: ok, me too. i think the best for tht to happen is bitcoin-staging with 1:1 peg. BlueMatt & gmaxwell had a plausible argument that a 1:1 peg maybe possible 15:02 < adam3us> CodeShark: the one change to rule them all as greg put it 15:03 < adam3us> CodeShark: it allows btc denominated (21 million coin cap preserving) alts or beta-coins 15:03 < jtimon> adam3us 1:1 is possible but very unconvinient 15:04 < adam3us> how so? 15:04 < jtimon> what would you make bitcoin's security depend on another chain? 15:04 < adam3us> jtimon: i think its fantastic; thats the clever part - onl the coins moved are at risk 15:05 < jtimon> well, the whole altchain is at risk if the validity of its transactions depend on bitcoin's chain 15:05 < jtimon> a reorg in one can cause a reorg in the other 15:05 < nsh> depends on the nature of the dependency :) 15:06 < jtimon> sure 15:06 < adam3us> jtimon: yes. but it seems unlikely for the mid-term that an alt will be more secure than bitcoin; and most alts are also uninteresting - no tx, and no intrinsic value 15:06 < adam3us> jtimon: i think a reorg in the alt is designed not to do anyting to btc; g like mining a long conf time 15:06 < jtimon> that doesn't say anything in favor of a bitcoin-pegged currency 15:07 < adam3us> jtimon: most of the thinly veiled excuses for alts are "oh the innovation" (like param tweak or hash function swap) 15:07 < adam3us> jtimon: so if they can peg to btc, then they have no excuse, they can do the innovation or go away 15:07 < CodeShark> alts are just tinkering with parameters that a suffiiciently modularized technology would allow you to freely tweak anyhow 15:08 < jtimon> I don't see how a bitcoin-pegged currency will prevent stupid people from doing and saying stupid things 15:08 < adam3us> jtimon: yes but they wont be starting a digital scarcity race and no one will risk btc in them, so its less wasteful and less dangerous 15:09 < jtimon> I disagree with your "digital scarcity" argumentation 15:09 < jtimon> you can innovate in your own testnet, you don't need a btc-pegged currency nor an altcoin for that 15:11 < CodeShark> the genie is out of the bottle 15:11 < nsh> hell, these days you can simulate an entire operating economy with some EC2 instances and historical transaction data 15:11 < CodeShark> regardless of whether or not alts are dangerous to bitcoin, they are inevitable 15:12 < jtimon> agreed 15:12 < nsh> network protocols are inevitable too, yet here we are on tcp/ip(v6) :) 15:12 < nsh> (unfortunately) 15:13 < jtimon> and a btc-pegged altcoin changes nothing with respect to the rest of altcoins 15:14 < jtimon> I think we just need more time and people losing ridiculous amounts of money by speculating in altcoins for the fever to pass 15:14 < adam3us> jtimon: it shows them for what they are - pump & dumps or they would use btc-peg (unless they are actually experimenting with the distribution model itself, like freicoin) 15:15 < jtimon> I think the btc-pegged altcoin is a bad idea and I'm still missing how it is supposed to change in any way the perception on altcoins 15:15 < jtimon> why would they use btc-peg? 15:16 < jtimon> that's not better, is worse 15:16 < jtimon> technically 15:16 < jtimon> an unecessary burden 15:16 < adam3us> jtimon: ok say someone wanted to implement freimarket extensions to make them available to bitcoin scripting. 15:17 < jtimon> cool 15:17 < adam3us> jtimon: they could do that using dogecoin or btc... you choose 15:17 < adam3us> jtimon: (and i like freimarket script extensions a lot... a couple of them i thought about before i saw it, very elegant and minimal!) 15:17 < jtimon> the hardfork on btc is much more difficult 15:18 < jtimon> so I think altcoins will have it first 15:18 < adam3us> jtimon: thats the point of bitcoin-staging and btc-peg. make that hard fork, then other hard-forks can happen in pegged-alts 15:18 < jtimon> probably freicoin first, then a freicoin-without-demurrage fork or several of them... 15:19 < adam3us> jtimon: thats why i said gmaxwell called it to the one change to rule them all - its literal, other forks dont need forks after that 15:19 < jtimon> just like can happen in non-pegged alts much more easily I still don't see the point 15:20 < jtimon> the whole pegging stuff is a burden in your design I don't see what it adds other than calm some of your "digital scarcity" fears 15:20 < adam3us> jtimon: the point is where woud you rather have freimarket extensions available in bbqcoin or to btc users 15:21 < jtimon> to all users 15:21 < jtimon> but are bbq devs going to rebase their code? 15:21 < jtimon> maintain it? 15:21 < adam3us> jtimon: no really teh original motivation is the unfortunate conflict between the need to be careful with btc changes, to preserve value, and the desire to implement known useful improvements 15:22 < adam3us> jtimon: bbq are a joke, thats my point; it'll probably flame out at some point when the dev gets bored and it breaks 15:22 < jtimon> I understand the motiviation, I just disagree that your proposed solution helps in any way or that altcoins are really a problem 15:23 < jtimon> many people involved in altcoins haven't seen altcoins die yet 15:23 < adam3us> jtimon: give it time. quite a few have died already in flame outs and peter outs 15:24 < jtimon> after a couple of them die in their hands they will think twice before speculating on the next proprietary altcoin 15:24 < adam3us> jtimon: btw i mean param-tweak alts... should distinguish - i am using alt as short hand for things like bbq coin an doge coin 15:24 < jtimon> they're all alts 15:24 < pigeons> probably the first ones WEEDS, 100% premined only tx fee currency, and beertokens, backed by 1 bottle of beer, have died 15:25 < jtimon> I think "backing" is a bad idea in general 15:25 < jtimon> for money 15:26 < pigeons> yes, but fun thing is you can prove something is a good or bad idea eventually 15:26 < jtimon> a year from now, there will be articles with a long list of death altcoins 15:26 < warren> it's hard for them to die 15:27 < jtimon> well, most of them are zombies as currencies anyway 15:27 < pigeons> anyway the client multicoin that sacarlson used by forking bitcoin and pulling out the parameters into config files was used by later altcoins like tenebrix and fairbrix which eventually brought litecoin even though at that point litecoin decided to fork from bitcoin again 15:28 < jtimon> yeah, multicoin was an interesting project 15:28 < pigeons> he was starting to work on plugin modules for things like difficulty adjustment filters as that started to get more complex than just changing static numbers, but then he got a real job 15:29 < pigeons> and bitcoin is kind of less interesting now that its big bucks for some 15:29 < pigeons> but for some people that makes it more interesting 15:32 < pigeons> i think a focus on decentralizing mining would be a good niche for an altcoin. i guess the tools are all there with GBT, and there are example models such as p2pool with real data to look at 15:53 < jtimon> kind of off-topic but...does anyone know if the rumors that say coinbase uses mongoDB as their primary store for financial transactions are true or not? 16:07 < nsh> jtimon, i am relatively confident they are true or not, yes 16:08 < jtimon> yeah true or not, that's what I thought 16:08 < nsh> it's always the way... 16:08 < nsh> damn you aristotle. damn you to hell... 16:11 < andytoshi> nsh: you can always study set theory if this dichotomy bothers you ;) 16:13 * nsh smiles 16:14 < nsh> (set-theoretical approches, e.g. fuzzy logic, are still fundamentally predicated upon bivalent membership identity, and can do more than concealing the dichotomy at a lower level of analysis) 16:16 < nsh> (a really non-binary system of logic has values that are qualitatively different to truth and falsity, rather than shades of the two) 16:18 < andytoshi> i meant, you can reject the law of excluded middle and do logic that way.. 16:19 < andytoshi> i don't know what field those people claim to be part of 16:19 < andytoshi> without making any claims as to what's in the middle 16:20 < andytoshi> hmm, you're still right, it's either true or not --- and false or not 16:20 < andytoshi> perhaps you should study zen then 16:24 < nsh> it's not just the law of the excluded middle. that's one pillar of bivalence. the other is the law of non-contradiction. every A is A, no A is not-A 16:25 < nsh> it's difficult to imagine a system without the law of non-contradiction. whether this is a reflection of a universal truth [sic] or a result of our historical mathematical/logical/linguistic enculturation is an open question :) 16:25 < jtimon> A + 0 = A, A + 1 = 1 16:27 < andytoshi> you don't need much in the way of axioms for a single contradiction to imply every statement is true.. 16:28 < andytoshi> so it's definitely baked pretty hard into historical logic 16:28 < andytoshi> i'm not familiar with the attempts to fix this non-robustness 16:29 * nsh nods 16:31 < nsh> there is a body of work due to Lukasiewicz but it's accompanied by an unfortunate tendency of later thinkers to reduce it back to bimodality 16:33 < nsh> More recently A. S. Karpenko. it's my occasional hobby to casually read up on it, but more slips past me than sticks, as with most matters 16:35 < nsh> discussed in some detail here: http://www.oocities.org/m_valuedlets/tranche4.html but unless you have predilection to wading through schizoform word salad it might not be much use to you :) 16:36 < andytoshi> 'fraid not :) 16:36 < nsh> fair enough 16:52 <@gmaxwell> pigeons: I thought beertoken was backed by some promise to deliver beer (not just one bottle, but some larger quantity as set by some kind of board or something) 16:59 <@gmaxwell> jtimon: I've seen a number of pretty concerning technical behaviors from coinbase, so I'd believe any random thing. 17:58 < pigeons> gmaxwell: there wasn't a beertokens comittee, it was just steve, and yes like all these things from silver certificates to mtgox usd it ultimately comes down to a promise 17:58 < pigeons> the promise was to redeem each beertoken for one bottle of a specific type of beer that steve liked and was common in thailand where he lived 17:59 < pigeons> but he didnt buy the beer and refrigeration and storage, he backed it with bitcoins, which brought up a problem as bitcoins decreased in value a lot from when he set them aside 17:59 < pigeons> he ended up buying more coins to make up the difference 18:01 < pigeons> and guys, BigDataBorat says "My contact at Coinbase say use of MongoDB strictly for reason of give client plausible deniability." 18:01 < pigeons> https://twitter.com/BigDataBorat 18:01 <@gmaxwell> what the heck does that mean? 18:02 < pigeons> "Estimate of MongoDB's value vary, one replica say $700m, one replica say $1.2 billion, one replica say 1.5 billion." 18:02 <@gmaxwell> "when we stole all the coins we could plausably deny it being theft?" 18:02 < pigeons> yes that's what it means 18:03 < Luke-Jr> lol 18:04 < nsh> hi orperelman. i liked your work on the poincare conjecture 18:04 < nsh> thanks for inventing bitcoin :) 18:05 < nsh> (i'm not personally sure ricci flow with surgery is a valid technique, but i'm not a topologist) 18:48 < maaku> andytoshi: isn't that constructivist math? (removing the law of excluded middle) 19:06 < andytoshi> maaku: yeah, that's the name for doing mathematics that way (e.g. rejecting proofs by contradiction) 19:07 < andytoshi> there are subsets of logic (which i have ~0 knowledge of) which do things like fuzzy logic and try to make this concrete 19:08 < andytoshi> my girlfriend was into constructivist math for a short while, not believing in anything that wasn't computable 19:08 < andytoshi> but it's nearly impossible to do a lot of classical mathematics that way 19:22 < nsh> depends how you define "doing mathematics" 19:22 < nsh> :) 19:23 < nsh> as a pursuit of noble platonic truths, or as a means towards solving practical problems... 19:24 < nsh> i'm not sure there are many engineering artifacts that are based predominently on an existential proof 19:24 < nsh> hmm, not so sure, now i think about it a bit more 19:53 <@gmaxwell> they're not unrelated. 19:53 <@gmaxwell> if you find some totally abstract "noble platonic truth" it can be a bridge that solves a pratical problem. 19:53 * nsh nods 19:55 <@gmaxwell> e.g. there is a bunch of NP proof stuff where you can show a proof system is sound by reducing it to a 2d graph coloring problem, and then show that if the system is unsound it would contradict four coloring, which otherwise is kinda useless trivia. 19:56 < nsh> right, came across that recently in a talk, funnily enough 21:44 < maaku> but ultimately you have to reduce it to be constructivist to enter the realm of engineering 21:45 < maaku> e.g. if you look at real numbers constructively, you get this funny think called numerical analysis ... 21:47 < nsh> analysis was pretty practical when it came to aiming canon :) 21:48 < nsh> computers are still named after the art of ordinance in french... 21:55 < nsh> (philosophically, it fascinates me that the assumption of the continuum, even though actual algorithmic infinities are avoided, yields such powerful results in anaylsis. we can calculate things in continuuous sets that would suffer combinatoric explosion over discrete structures...) 21:56 <@gmaxwell> Stirling's approximation <3 21:57 < phantomcircuit> wtf 21:57 <@gmaxwell> being able to answer questions like from an infinite distribution of 50/50 true/false how likely is it you'd draw 30 and get 5 true... answering it combinitorically is impossible. 21:57 < phantomcircuit> i just noticed google wallet is still using checkout.google.com 22:00 * nsh nods --- Log closed Sun Dec 22 00:00:17 2013