| 00:00:00 | nsh: | seems rather esoteric maybe |
| 00:00:10 | nsh: | no more so than the random oracle model though |
| 00:00:11 | amiller_: | thats the premise of a UFO i think |
| 00:00:15 | nsh: | mm |
| 00:00:50 | amiller_: | short answer, it probably does, but a) you can't check, b) you have to sample a lot of times to get a good chance of ending up with one, so oh well |
| 00:01:43 | kanzure: | also i think you can do hashrate proof-of-work incentives as a method of statistical sampling, which provides knowledge of various floors (when an observation is made) |
| 00:02:30 | kanzure: | for example, a task that osmeone would have to specifically point their hashrate at instead of mining blocks, which obviously not everyone is obligated to try which is why i said floor-when-observations-are-made |
| 00:04:24 | nsh: | well, moment of success is a proxy-measurement of hashes attempted, then rate is just the delta |
| 00:04:35 | nsh: | you cannot conceivably measure hashes-that-might-have-been-attempted |
| 00:04:49 | nsh: | which is part of the problem with hashpower liquidity/slushing |
| 00:05:05 | nsh: | and part of what sidechains hopes to solve |
| 00:09:04 | kanzure: | i could imagine being quite easily convinced that hashrate is the wrong word to be using (especially since it's so hard to measure by anyone else) |
| 00:13:56 | nsh: | * nsh muses |
| 00:14:36 | kanzure: | nsh: http://en.wikipedia.org/wiki/Charles_Sanders_Peirce#Probability_and_statistics |
| 00:14:43 | kanzure: | (not sure if this is the direction you're asking for) |
| 00:15:24 | nsh: | yeah, but i basically hate everyone who has ever written on the subject of statistics or logic |
| 00:15:28 | nsh: | for aesthetic reasons |
| 00:15:34 | nsh: | so it's painful to consume |
| 00:16:11 | nsh: | (conceptuaesthetic reasons, or whatever, the thing that makes you know what's elegant and that's horrific in concepts and their compositions) |
| 00:16:16 | nsh: | *what's |
| 00:16:44 | kanzure: | that hate i probably deserved i'm sure, and i have nothing good to offer you |
| 00:16:48 | kanzure: | *is probably |
| 00:48:09 | kanzure: | gmaxwell: regarding your request for materials involving other unintended consequences in complex system design... http://betterembsw.blogspot.com/2014/09/a-case-study-of-toyota-unintended.html?m=1 |
| 00:48:37 | kanzure: | "1. The Throttle Angle function in the Toyota code had a McCabe Cyclomatic Complexity of 146 (over 50 is considered untestable according to slides) [slide 38] 2. The main throttle function was 1300 lines long, and had no directed tests. [slide 38] 3. I find the static analysis results quite alarming. [slide 37] 4. 80+% of variables were declared as global. [slide 40] I find this to be a stunning lapse of quality, especially for a ... |
| 00:48:43 | kanzure: | ... safety-critical system." |
| 00:48:56 | kanzure: | ( https://news.ycombinator.com/item?id=8905718 ) |
| 00:56:24 | kanzure: | "The infeasibility of quantifying the reliability of life-critical real-time software" http://www3.cs.stonybrook.edu/~tashbook/fall2009/cse308/butler-finelli-infeasibilit.pdf |
| 00:57:55 | kanzure: | (cited in the slides) |
| 02:07:45 | fanquake_: | fanquake_ is now known as fanquake |
| 03:10:21 | Pan0ram1x: | Pan0ram1x is now known as Guest62486 |
| 03:27:53 | kanzure: | maaku: is there a better physical-domain unit or dimension than "work" for describing proof-of-work? |
| 03:28:31 | kanzure: | hm maybe i should be asking fenn, since he's practically the resident evangelist of gnu units heh |
| 04:35:44 | bramc: | Isn't work generally expressed in gigahashes? |
| 04:48:05 | ElGalambo: | ElGalambo has left #bitcoin-wizards |
| 04:49:40 | kanzure: | sure but a gigahash is still a.. er.. "compressed".. physical unit of some other dimensions |
| 04:49:45 | kanzure: | or it might be some dimensionless unit, i haven't checked |
| 05:03:09 | starsoccer: | starsoccer is now known as Guest64765 |
| 05:10:23 | Adlai: | kanzure: not a perfect analogy, but percentage of checked solution space? |
| 05:23:04 | bramc: | The underlying unit is flops |
| 05:23:31 | bramc: | although there's no floating point involved, so technically it isn't flops, and generally the custom hardware has built-in hashing functionality |
| 06:08:26 | kanzure: | fenn: what is the physical dimensions of a flop second |
| 06:09:13 | nsh: | * nsh blicks |
| 06:09:29 | nsh: | flops has an implicit / |
| 06:10:08 | nsh: | so the physical dimension of a flop(s) second is just a barenumber |
| 06:10:28 | nsh: | the physical dimension of an operation per second is... well i guess we have to analyze a bit what we mean by operation |
| 06:10:37 | nsh: | but it will resolve to entropics |
| 06:10:52 | nsh: | so amount of energy, assuming irreversible computing |
| 06:11:02 | nsh: | thus power |
| 06:15:20 | op_mul: | nsh: what's nice about measuring hashpower is that you can, with some weak assumptions, convert them to the consumption of energy. |
| 06:17:44 | op_mul: | if we make the sort of low end assumption that the entire network is powered by Antminer S5 units, then we're consuming about 0.51 joule per gigahash. |
| 06:18:26 | nsh: | * nsh nods |
| 06:19:18 | nsh: | but actual power consumption/dissipation is itself quite a simplification |
| 06:19:39 | op_mul: | from there, you can move to working out the equivalent costs in various foods. say, how many acorns you'd need to burn in order to mine one block. which comes up with some absolutely hilarious numbers. |
| 06:19:49 | nsh: | actually that's going on is that order is being imported and consumed and disorder exported and gotridof |
| 06:20:01 | nsh: | and that's a geometrical/spacial process |
| 06:20:11 | nsh: | you can't reason about it in put unitary terms |
| 06:20:18 | nsh: | *what's going on |
| 06:20:56 | op_mul: | (an acorn is about 490,000 joules) |
| 06:26:23 | nsh: | do kids play conkers where you grew up op_mul? |
| 06:27:06 | nsh: | the horse chestnut that wins a dozen rounds of smack-the-nuts-attached-to-string-against-each-other |
| 06:27:38 | nsh: | is if qualitatively special, or just quantitatively? |
| 06:27:44 | nsh: | *it |
| 06:28:11 | op_mul: | I'm out of touch with what kids do now. it's probably an iphone game. I only chose nuts because I know they have a very high energy density. about the same as a potato according to wolfram alpha, but in a lot less physical space. |
| 06:28:22 | nsh: | * nsh nods |
| 06:28:49 | nsh: | i'm just wondering about when it stops mattering how much of something there is there |
| 06:29:15 | nsh: | and starts mattering how it's arranged, how energy can be dispersed through it without exceeding some sheer threshold |
| 06:29:45 | nsh: | the winning conker is not necessarily of a greater mass(energy) |
| 06:29:48 | phantomcircuit: | this seems like an add discussion |
| 06:30:06 | nsh: | the disorder? |
| 06:30:09 | nsh: | :) |
| 06:30:26 | nsh: | or operation... |
| 06:32:09 | op_mul: | I was trying to put a real world scale on it, something like petahashes a second is pretty meaningless. |
| 06:33:34 | nsh: | well, it's more meaningful to the statistical expectation than workenergy dissipated |
| 06:33:43 | nsh: | and Bitcoin cares about the former |
| 06:33:58 | nsh: | the economics of bitcoin cares about the latter, though |
| 06:34:32 | nsh: | and that effects the design decisions in ways that are unfortunately filtered through human incentive filters |
| 06:35:21 | nsh: | this is an even more spurious relation than between petahashes and megajoules |
| 06:37:19 | op_mul: | ignoring the tears of the irish, if we burnt every potato produced in 2014 and converted it directly into mining hashpower with zero losses we could run the bitcoin network for 17 days. |
| 06:37:56 | op_mul: | er hold on, I'm off by a large amount there. |
| 06:41:13 | op_mul: | 30.12 days. |
| 06:41:53 | op_mul: | maybe that's not such an easy to understand measure of hashpower after all. |
| 06:48:09 | nsh: | matter is a bad measure of work. it's very hard to efficiently convert stuff to stuff-gets-done |
| 06:48:45 | nsh: | if you measured it in the useful energy an average human would get from eating those potatoes |
| 06:49:11 | nsh: | it might be more reasonable |
| 06:49:47 | nsh: | (for silly values of reason, being, incidentally, the best values of reason) |
| 06:51:56 | op_mul: | wolfram alpha tells me that if you ate 300 million tonnes of potatos you would die of vitamin A poisoning. |
| 06:52:44 | op_mul: | at any rate, the only reason you would work out things like this is just to attempt to show scale to a reader. |
| 06:54:54 | kanzure: | .wa 300 million tonnes of potato |
| 06:55:07 | yoleaux: | prepared potato: amount: 300000000 t (metric tons): Calories: | mean value: % daily value: range; total calories: 2.6×10¹⁴ Cal: 1.294×10¹³%: (1.3×10¹⁴ to 3.2×10¹⁴) Cal; fat calories: 2.9×10¹² Cal: | (2.5×10¹² to 5.3×10¹²) Cal; Fats and fatty acids: | mean value: % daily value: range; total fat: 3.7×10⁸ kg: 5.674×10¹¹%: (3×10⁸ to 6.3×10⁸) kg; saturated fat: 9.3×10⁷ kg: 4.65×10¹¹%: (6.6× … |
| 06:55:07 | yoleaux: | 10⁷ to 1.6×10⁸) kg |
| 06:59:38 | phantomcircuit: | using tor has gotten a lot more annoying since cloudflare took over the fucking internet |
| 07:00:18 | op_mul: | an interesting feature of cloudflare is that it ruins almost all censorship. |
| 07:05:23 | phantomcircuit: | op_mul, ? |
| 07:06:27 | op_mul: | IP based censorship and MITM based censorship fail if the host is behind cloudflare. |
| 07:07:38 | op_mul: | though I can't say I'm happy about using the cloudflarternet. sites like blockchain.info serving up security critical javascript through a tampering reverse proxy is just insane. |
| 07:26:46 | phantomcircuit: | op_mul, it's easy enough to block the dns requests |
| 07:26:54 | phantomcircuit: | but i guess that's easier to get around |
| 07:27:18 | op_mul: | yes, my ISP only poisons their default DNS I think. |
| 07:27:56 | op_mul: | though I've noticed a few bits of weirdness which suggest a transparent HTTP proxy. |
| 08:15:28 | fluffypony: | op_mul: they could still have the site behind CF and serve the JS off their own tin |
| 08:16:36 | op_mul: | fluffypony: cloudflare could still just inject javascript onto the HTML. |
| 08:16:47 | fluffypony: | that's true |
| 09:05:18 | rajaniemi.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 |
| 09:05:18 | rajaniemi.freenode.net: | Users on #bitcoin-wizards: andy-logbot Guyver2 siraj damethos Dizzle Transisto orik paveljanik huseby qwopqwop Cory RoboTeddy Guest64765 TheSeven eslbaer_ null_radix Graftec Guest62486 nessence jb55 jgarzik d1ggy_ Dr-G fanquake ryanxcharles hashtag dgenr8 ebfull shesek Adlai Shiftos wallet42 op_mul justanotheruser mappum devrandom Oizopower jbenet MoALTz PRab waxwing roconnor coiner nuke1989 bsm117532 HarusameNyanko HaltingState2 koshii zwischenzug Burrito nsh maaku |
| 09:05:18 | rajaniemi.freenode.net: | Users on #bitcoin-wizards: coblee DougieBot5000 ryan-c c0rw1n atgreen catcow afk11 ajweiss luny DoctorBTC gavinandresen wizkid057 hashtag_ stonecoldpat davout hollandais warren bosma forrestv otoburb ahmed_ PaulCapestany phantomcircuit mortale gribble copumpkin [d__d] lechuga_ nanotube so Apocalyptic michagogo btc___ Muis CryptOprah HM2 bbrittain Hunger- phedny Iriez comboy_ MRL-Relay Alanius cluckj optimator_ earlz cryptowest weex wiz artifexd coryfields_ kumavis |
| 09:05:18 | rajaniemi.freenode.net: | Users on #bitcoin-wizards: harrow` tripleslash brad___ thrasher` Graet helo throughnothing nickler_ Adrian_G fenn yrashk sl01 mr_burdell JonTitor roasbeef Keefe espes__ Logicwax petertodd dansmith_btc lclc poggy sadgit heath kanzure fluffypony xabbix_ catlasshrugged pigeons Krellan Luke-Jr dasource Guest74833 Eliel_ morcos K1773R _v3Rve Dyaheon tromp amiller_ brand0 LarsLarsen asoltys_ isis livegnik go1111111 tacotime BlueMatt gmaxwell eric azariah kobud Taek jaekwon_ |
| 09:05:18 | rajaniemi.freenode.net: | Users on #bitcoin-wizards: crescend1 berndj EasyAt Starduster_ jaromil d9b4bef9 nick1234abcd__ samson_ bepo_ NikolaiToryzin TD-Linux warptangent pi07r_ deego coryfields SubCreative Emcy_ tromp_ sdaftuar lnovy midnightmagic delll iddo Xzibit17 BigBitz yoleaux hguux_ BrainOverfl0w wumpus sneak btcdrak kinlo andytoshi gwillen @ChanServ gnusha burcin spinza Nightwolf a5m0 Fistful_of_Coins epscy Meeh s1w jcorgan Anduck |
| 13:24:15 | justanot1eruser: | justanot1eruser is now known as justanotheruser |
| 13:25:02 | Guest64765: | Guest64765 is now known as starsoccer |
| 13:25:16 | justanot1eruser: | justanot1eruser is now known as justanotheruser |
| 14:34:09 | nuke1989: | nuke1989 is now known as nuke19892 |
| 14:34:12 | nuke19892: | nuke19892 is now known as nuke1989 |
| 14:46:53 | sadgit: | sadgit has left #bitcoin-wizards |
| 15:32:53 | fanquake_: | fanquake_ is now known as fanquake |
| 17:06:03 | imposter: | imposter is now known as JaredDeryeghiaya |
| 17:06:58 | JaredDeryeghiaya: | JaredDeryeghiaya is now known as Deryeghiayan |
| 19:49:48 | phantomcircuit: | fluffypony, indeed cf can do basically anything, they can generate ssl certificates with arbitrary domain names |
| 19:50:07 | phantomcircuit: | and they can trivially conduct a world wide BGP spoof |
| 19:50:35 | phantomcircuit: | cf is basically exactly how you would model a world wide mitm intercept framework... |
| 20:31:05 | jgarzik: | phantomcircuit, and it's a giant SPOF for bitcoin |
| 20:31:15 | jgarzik: | consider the number of bitcoin sites that use it |
| 21:22:27 | amiller_: | amiller_ is now known as amiller |
| 21:42:58 | Pasha: | Pasha is now known as Cory |
| 23:16:20 | fanquake_: | fanquake_ is now known as fanquake |
| 23:47:18 | Deryeghiayan: | Deryeghiayan is now known as imposter |
| 23:59:15 | maaku: | maaku is now known as Guest39463 |