| 00:00:06 | Taek42: | ========== |
| 00:00:13 | Taek42: | There would be a large advantage to 100ms block times |
| 00:00:16 | Taek42: | or 1ms block times |
| 00:00:26 | Taek42: | or the ability to release blocks at any second |
| 00:00:28 | Taek42: | *any moment |
| 00:00:38 | Taek42: | The advantage is that you don't need mining pools anymore |
| 00:01:29 | Taek42: | If the difficulty is 1, and the network has some magical way of propagating blocks instantly, pretty much no ASIC owner would need a pool. |
| 00:59:18 | pigeons: | yes lots of things become possible when you have magic |
| 01:18:42 | gmaxwell: | Taek42: Yes other people have previously pointed out that useful work complicates incentives. I believe the ASIC hard faq on bitcoin.ninja makes this observation. |
| 01:20:23 | gmaxwell: | I think it's somewhat less bad than you think (and that I originally thought) but its certantly a complicating factor. (I guess the important thing to keep in mind is that the incentives are always over marginal returns and oppturnity costs; not over the absolutes) |
| 01:29:43 | andytoshi: | the asic faq also paints an uglier picture than is probably realistic |
| 01:30:09 | andytoshi: | like, people sabotaging the chain to avoid miners infringing on their scientific work |
| 01:44:20 | andytoshi: | realistically all that needs to be said is "optimization-free things are by definition not scientifically interesting" so they're incompatible ;) |
| 01:45:01 | Taek42: | I don't understand what that means |
| 01:45:43 | andytoshi: | Taek42: it means that if i'm investigating a problem and i necessarily have to brute-force it, idgaf about it |
| 01:45:51 | andytoshi: | this is not true for data processing kinda problems, but it is for math or CS |
| 01:46:10 | Taek42: | got it |
| 01:46:50 | andytoshi: | (still not quite true -- eg there are problems in crypto where it's useful to grind through parameters looking for particularly efficient ones, but these are not "discovering new science" kinda problems) |
| 01:47:06 | gmaxwell: | I think if you want a really strong sense of optimization free, you first need to prove P!=NP. .. those criteria are somewhat soft, but many many things people propose fail them throughly. |
| 01:48:52 | gmaxwell: | lots of interesting science involves sciocastic searches, but they are very much not even arguably optimization free most of the time. (many times I've enaged in some slow exhaustive search or random sampling only to find a closed solution before the optimizer finished :) ) |
| 02:22:48 | Taek42: | has much work been done on a decentralized exchange between Bitcoin and other coins? |
| 02:23:09 | sipa: | what would that mean? |
| 02:23:44 | Taek42: | if you want to trade bitcoins for dogecoins, for example |
| 02:23:55 | Taek42: | you typically go through a centralized exchange like bter or btc-e |
| 02:24:34 | Taek42: | There would be many advantages to having a completely decentralized method for setting up trading between various coins and cryptoassets |
| 02:25:05 | sipa: | again: what would 'decentralized' exchange mean? |
| 02:26:03 | Taek42: | a trust free way for trading M coins of one currency for N coins in another currency? |
| 02:26:23 | Taek42: | Also preferably one that could execute quickly, and be adjusted as prices change |
| 02:26:27 | sipa: | there are atomic cross-chain swaps |
| 02:27:43 | livegnik: | Hey everyone |
| 02:27:57 | Taek42: | iirc that requires two parties to come to an agreement on price first, and then execute a trade second, right? So they don't need to have trust, but they do need to somehow find each other |
| 02:29:04 | Taek42: | Is there a way that I could make an open proposal "10 bitcoins for 1000 doge" and complete the transaction with the first party to accept, with no risk of later parties losing doge or me executing multiple times |
| 02:30:03 | livegnik: | Taek42, you can have a look at Coinffeine to see what a truly decentralized peer-to-peer exchange could look like. |
| 02:30:05 | livegnik: | http://www.coinffeine.com/ |
| 02:30:53 | livegnik: | https://github.com/Coinffeine/coinffeine/wiki/Exchange-algorithm |
| 02:30:58 | mr_burdell: | doesn't that just reduce risk to some predefined amount? |
| 02:31:11 | mr_burdell: | not completely trustless |
| 02:31:37 | livegnik: | There's always the party on the other side. How do you know if you can trust him/her? |
| 02:31:51 | jgarzik: | jgarzik has left #bitcoin-wizards |
| 02:32:27 | sipa: | i think that what many people understand as 'decentralized trustless exchange' is impossible |
| 02:32:37 | sipa: | certain parts of the solution may be possible though |
| 02:32:49 | livegnik: | Which parts would be considered impossible then? |
| 02:34:34 | gmaxwell: | trading with USD in a way that doesn't have counterparty risk. |
| 02:34:54 | livegnik: | Well, that's always a problem with fiat, right? ;) |
| 02:35:31 | gmaxwell: | if you're just talking about 'cryptocurrency' you can already do that today, though there isn't any software that makes it pretty... and you can use OTC to find traders, but it's not really very interesting. |
| 02:36:43 | livegnik: | Well, I believe that the micropayment channel solution, offered by the Coinffeine algo, is pretty interesting to say the least. Especially since they promise the whole fiat-crypto-fiat part. |
| 02:41:10 | gmaxwell: | A lot of that stuff is really deceptively marketed. There are a bunch of claims about 'decenteralized' but the security model is almost indistinguishable from mtgox. |
| 02:42:35 | phantomcircuit: | gmaxwell, my personal favorite is the centrally issued USD coins |
| 02:42:39 | phantomcircuit: | aka ripple |
| 02:42:40 | phantomcircuit: | derp |
| 02:44:18 | gmaxwell: | There can be transparency advantages, if things are done right... but it's hard to offset the challenges (e.g. hard/impossible to make a fast atomic orderbook... esp since in designs which have no holdup risk on the bitcoin side the obvious designs make it trivial to manipulate the market with 'orders' which you'll never allow to execute) |
| 02:46:36 | gmaxwell: | I suppose I'd whine less if I thought people were honest about their security models, and if they weren't so often promoted as solving the specific problems they can't solve (trusted parties running off with assets. :) ) |
| 02:47:36 | phantomcircuit: | gmaxwell, yeah |
| 02:48:01 | phantomcircuit: | i actually think that something with the same design as liberty reserve would be a reasonable tradeoff |
| 02:48:13 | livegnik: | gmaxwell, I concur. |
| 02:48:16 | phantomcircuit: | the problem being that operating such an entity is almost certainly regulated |
| 02:48:38 | livegnik: | Less regulation, more decentralization... imo |
| 02:49:13 | livegnik: | I cool thing about regulation are those fictitious borders though. |
| 02:49:17 | livegnik: | *The cool |
| 02:50:47 | gmaxwell: | an entity that is somhow 'immune' to regulation is usually even more immune to reprocussions for just walking off with your money (if it's possible to do so); so you might want to contemplate how useful that line of thinking is for a system which uses trusted parties to hold funds. |
| 02:51:25 | phantomcircuit: | gmaxwell, agreed |
| 02:52:58 | livegnik: | gmaxwell, I'm specifically talking about replacing regulation with decentralization. More and more centralized authorities will become obsolete in the future. Untill then, it will be probably a mix of the two. |
| 02:53:12 | kanzure: | that is not insightful |
| 02:53:30 | livegnik: | For example Bitcoin has its own regulation, hard coded into it. |
| 02:54:14 | gmaxwell: | * gmaxwell consults lastlog |
| 02:54:17 | gmaxwell: | * gmaxwell sighs |
| 02:55:36 | livegnik: | This could be done in other cases too. Take trust for example. Look at a decentralized identification & reputation database on a p2p network like Identifi. |
| 02:55:45 | gmaxwell: | livegnik: I was specifically speaking about services which inherently require trusted parties. You're not saying anything thats interesting to anyone in here now. |
| 02:56:07 | kanzure: | livegnik: you are lying, those services don't do that |
| 02:56:17 | livegnik: | A Web of Trust (WoT) could become the basic infrastructure for a lot of stuff we nowadays depend on centralized authorities. |
| 02:56:45 | livegnik: | kanzure: Maybe there's a misunderstanding, but I don't think I'm lying. |
| 02:57:21 | kanzure: | isn't this one of those 'just add a blockchain' projects |
| 02:58:01 | livegnik: | Nope. Sirius actually decided to take out the blockchain, since there's no use for it at this point. |
| 02:58:18 | livegnik: | It is based on bitcoind though. |
| 02:58:47 | livegnik: | Sorry for not adding anything interesting. I'll just continue lurking. |
| 03:01:39 | kanzure: | getting incentives right is difficult |
| 03:10:06 | Taek42: | (sorry for dropping out there for a bit) sipa was "i think that what many people understand as 'decentralized trustless exchange' is impossible" targeted at me? |
| 03:13:14 | Taek42: | I think the ideal for a trustless exchange would enable you to engage in what would essentially be high frequency trading except with cryptocurrencies. |
| 03:16:35 | phantomcircuit: | gmaxwell, afaict you cant do low latency trading unless you have trust |
| 03:16:58 | Taek42: | from what I understand of atomic cross-chain trading, the trade would have to be agreed upon first, which means an attacker could frustrate my HFT engine by setting up a deal and going back on it repeatedly. He keeps my coins locked down while I wait for him to complete the trade. He doesn't, I get my coins back, but my time has been wasted |
| 03:18:30 | Taek42: | phantomcircuit I believe you can use tools similar to microchanel payments, so that each party is confident that they will receive their coins when the transaction makes it into the blockchain before any confirmations. |
| 03:18:37 | sipa: | there is no way you can do fast trading across chains |
| 03:18:45 | sipa: | as you have by definition two transactions |
| 03:18:51 | Taek42: | hmm |
| 03:18:57 | sipa: | which need to depend on eachother |
| 03:19:24 | sipa: | the only way to do fast inter-asset trading is when you do it within one transaction |
| 03:19:29 | sipa: | i.e., within one chain |
| 03:19:49 | phantomcircuit: | interestingly most major exchanges basically operate with unverifiable offers to fulfill |
| 03:20:21 | sipa: | however, to get those coins in one chain means that at least one and possibly both are not native to the chain |
| 03:20:51 | sipa: | which means that you're working with tokens that are tradable for coins in another chain, rather than the foreign coins themsepves |
| 03:21:14 | sipa: | which means you are implicitly trusting the issuer of those tokens |
| 03:21:23 | sipa: | and there is no way around that issuer trust |
| 03:21:44 | sipa: | if you accept that trust, the rest of the exchange process can likely be decentralized/trustless |
| 03:24:01 | gmaxwell: | that said; if you're taking that trust... it would be useful to think carefully about what if any good decenteralizing the rest does. Many of the people talking about that stuff have no real answer for the benifits once everyone is clear on the fact that a trusted part remains. |
| 03:24:35 | woah_: | yea... trusting an exchange for one trade is not really a big deal imo |
| 03:24:54 | phantomcircuit: | woah_, it's the aggregate that's the problem |
| 03:25:04 | Taek42: | and the fees |
| 03:25:07 | phantomcircuit: | 10000 people trusting an exchange with 1BTC is real money |
| 03:25:13 | woah_: | true |
| 08:05:13 | card.freenode.net: | topic is: This channel is not about short-term Bitcoin development | http://bitcoin.ninja/ | This channel is logged. | For logs and more information, visit http://bitcoin.ninja |
| 08:05:13 | card.freenode.net: | Users on #bitcoin-wizards: andy-logbot damethos MoALTz todaystomorrow luktgf devrandom Graftec cym dgenr8 p15 tacotime x48 koshii wiretapped TheSeven qualiabyte kinlo jgarzik Transisto irc88 skinnkavaj zenojis SDCDev jaekwon1 hollandais fanquake DougieBot5000 prepost grubles samson_ nanotube drawingthesun gribble [Derek] BlueMatt Dyaheon wizkid057 HaltingState Muis CryptOprah_ jbenet promoJo michagogo zlinn_ mappum btc_ Guest2024 copumpkin ericp4 atgreen at0mat |
| 08:05:13 | card.freenode.net: | Users on #bitcoin-wizards: digitalmagus8 Sangheili emsid br4n_ pigeons arowser BrainOverfl0w lianj_ UukGoblin optimator wumpus Apocalyptic poggy_ rs0 jcorgan_ Iriez mr_burdell justanotheruser fluffypony epscy_ nuke1989 bsm117532 K1773R OneFixt_ a5m0 go1111111 GAit bbrittain nsh SomeoneWeird mikalv forrestv shesek livegnik jchp_ mkarrer Anduck warren asoltys Hunger- waxwing Emcy tromp_ EasyAt tucenaber Logicwax iddo tromp Grishnakh sl01 comboy_ Luke-Jr Graet espes__ sipa |
| 08:05:13 | card.freenode.net: | Users on #bitcoin-wizards: coryfields [d__d] midnightmagic bobke nsh- Adohgg pajarillo crescendo HM maaku phantomcircuit azariah4 zibbo gmaxwell Fistful_of_coins DoctorBTC harrow roasbeef andytoshi mmozeiko Eliel CodeShark throughnothing Guest47516 Alanius nickler_ Starsoccer pi07r ryan-c kanzure gwillen otoburb smooth helo Guest50253 weex abc56889 lechuga_ TD-Linux catcow danneu LaptopZZ_ burcin petertodd so phedny @ChanServ jaromil Taek42 berndj-blackout dansmith_btc |
| 08:05:13 | card.freenode.net: | Users on #bitcoin-wizards: [\\\] Starduster BigBitz Krellan Keefe spinza Nightwolf amiller |
| 12:48:41 | nsh-: | /topic you first need to prove P!=NP... |
| 15:29:35 | nullbyte_: | nullbyte_ is now known as Guest96868 |
| 17:21:32 | LaptopZZ_: | LaptopZZ_ is now known as LaptopZZ |
| 18:09:04 | epscy_: | epscy_ is now known as epscy |
| 18:38:50 | fanquake_: | fanquake_ is now known as fanquake |
| 19:46:05 | Aquent: | Aquent is now known as timeny |
| 19:46:22 | timeny: | timeny is now known as extraaaaaaaaaaa |
| 19:46:41 | extraaaaaaaaaaa: | extraaaaaaaaaaa is now known as Aquent |
| 21:09:01 | DougieBot5000_: | DougieBot5000_ is now known as DougieBot5000 |
| 22:25:06 | berndj-powerout: | berndj-powerout is now known as berndj-blackout |
| 23:50:40 | starsoccer: | starsoccer is now known as Guest19528 |