log☇︎
160400+ entries in 0.091s
mircea_popescu: i dunno why you're so fixated on poker, this was just early bitcoin business.
mircea_popescu: nah, i don't recall what the site was called, but dice something.
asciilifeform: was this in the great pokerholocausting?
mircea_popescu: then they want to talk about "Gdp growth" and "technology". fuck them in the ass with a pike of stupid, there's no technology in north america, and no gdp growth.
mircea_popescu: anyway, very much a http://trilema.com/2015/gerald-davis-is-wrong-heres-why/ : i was the first investor afaik, his site was going great, thgen the (canadian) government stepped in and burned down the respectable business of an intelligent citizen.
asciilifeform: but asciilifeform does not specifically favour any of the published algo
asciilifeform: (why free gift to derps)
asciilifeform: lol i was about to say.
mircea_popescu: and ima not talk about it, because fuck it.
asciilifeform: 'house' being them
mircea_popescu: again, this was for casino games, not poker. he never implemented multiplayer afaik.
asciilifeform: who change their nonces specifically when it would benefit the house, and in such a way as to.
asciilifeform: how to prove that the other N-1 players, other than yourself, ain't sybils
mircea_popescu: that's what it means "house", in this new paradigm.
mircea_popescu: house uses the same hash for all players for the whole day.
asciilifeform: ( the nonces OTHER than yours )
asciilifeform: walk the nonces until the hash output comes out in favour of house
mircea_popescu: and it doesn't work without withening ; or in other words : he intelligently used the otherwise useless whitening step as an assurance provider. was pretty clever move, esp considering the standard for intelligence at the time.
mircea_popescu: ie, you against the house.
asciilifeform: and that all the other players' nonces aren't hashwaltzes
mircea_popescu: and then when the reveal occurs, you sha512(i like chicks with small tits71c0891f45d65ff7) and see.
asciilifeform: of course you still can't prove that the house was not 'btcmining' the hash
asciilifeform: this reduces to shamir's algo dunnit.
mircea_popescu: then the first hand is played as 0c3dbf77a25ebe9733548266406161111522839d1d41f221070e59c361f6b2b0990df0606d143e90b1bb9ed1a9b39ff5fcf0ac153cc0e90297f1273076fcf38c
mircea_popescu: concrete example : you set your seed to "i like chicks with small tits", site sets seed to "71c0891f45d65ff7"
asciilifeform: does it publish 'and then player 9000 changed his nonce at time 100' etc ?
mircea_popescu: site publishes site + player, player can verify that published player = what he thought he was and that the sha results match (published)
asciilifeform: what does the end-of-play reveal look like ?
mircea_popescu: conceptually equivalent to sha(site+player1) etc.
mircea_popescu: asciilifeform seed is a combo of site-produced TRNG entropy and player-set (with defaults if playher doesn't feel like setting). you can change your portion at any time. ☟︎
mircea_popescu: the reason i kept trying to get him to be involved with things was exactly that he had good solutions to actual problems. but... never caught on.
a111: Logged on 2017-09-01 22:18 asciilifeform: this does not actually solve the problem of proving that the output of prng was not known in advance to one or more of the players.
mircea_popescu: http://btcbase.org/log/2017-09-01#1709343 << dooglus 's implementation actually solved that. ☝︎
mircea_popescu: true. but not really a bread and butter thing.
asciilifeform: depends on the occasion neh.
asciilifeform: mircea_popescu: in point of fact folx can and do sometimes pay for phd to make an appearance at an assembly line.
asciilifeform: almost like 'trusted' were just a cardboard propeller on a straw airplane to these folx
mircea_popescu: asciilifeform this impression is borne out of lack of experience and no more. what, i'm going to assembly line phds ? ☟︎
a111: Logged on 2017-09-01 22:15 asciilifeform: 'We submitted extensive information about our random number generator (RNG) to an independent organization. We asked this trusted resource to perform an in-depth analysis of the randomness of the output of the RNG, and its implementation in the shuffling of the cards on our platform.'
mircea_popescu: http://btcbase.org/log/2017-09-01#1709330 << this trusted, anonymous item. sounds like fiat thought process, that strand of fatlogic.\ ☝︎
asciilifeform: mircea_popescu: i beg to differ, if unit hangs around asciilifeform's neck from door to door, it's a layer reduction and not == to usmail.
asciilifeform: but i dun know that anybody wants to pay for what it costs to deliver this way, unless he lives across the street.
asciilifeform: and asciilifeform can be replaced with ninjashogun right here in his chair neh. idea is to reduce the layers to 1.
mircea_popescu: what, you're going to arm them too ? with what, submarines ?
asciilifeform: mircea_popescu: proposed strictly as an answer to the provenance thing
a111: Logged on 2017-09-01 22:06 asciilifeform: though i've thought about suggesting to mircea_popescu to offer hand-delivery ( chore would fall to asciilifeform... ) as an option. but i dun like nonflat prices
mircea_popescu: http://btcbase.org/log/2017-09-01#1709309 << this works well for hookers, blow, beers, pizza, generally consumables. ☝︎
asciilifeform: wasn't even thinking of commission. but abstract 'here'
mircea_popescu: asciilifeform didn't want to commission back in 2013 or so because i don't recall, something about how doesn't want to lose the spirit or somesuch.
asciilifeform: why the author is not here, remains a puzzler
mircea_popescu: the girls love the "ne futem sau evadam ?" "fuff. mai fimflu" one. which only exists in romanian.
jhvh1: shinohai: The operation succeeded.
shinohai: !~later tell jurov I wish to donate my s.qntra shares this month to the bitcoin foundation. Please advise how to proceed, thx.
mircea_popescu: i still think the best ever is the magic of denial one.
trinque: where's that oglaf
mircea_popescu: the DUKES of hazards of hashcollisions, trinque.
trinque: this is not persuading me of the hazards of hashcollision
mircea_popescu: hey, at least yours has the best rack of the alf-right or how did that go.
asciilifeform: oh and hey remember the idiot anglos who pictured 'girl mircea'
a111: Logged on 2017-09-01 23:43 mircea_popescu: btw am i the only one HIGHLY entertained by the internets producing a girl alf ?
mircea_popescu: http://btcbase.org/log/2017-09-01#1709230 << you should see the http://trilema.com/2017/tute/ shuffling, its something else. so, dealer spreads cards on the table and rubs his hands on them in a circular pattern sometimes distrurbign the player's chip stacks. ☝︎
mircea_popescu: btw am i the only one HIGHLY entertained by the internets producing a girl alf ? ☟︎
a111: Logged on 2017-09-01 20:57 trinque: www itself is an incaism. death to "it was said on arsebook walled garden; lives and dies at the mercy of arsebook"
mircea_popescu: http://btcbase.org/log/2017-09-01#1709181 << quite. and as the femme alf says, "they don't even know it exists". ☝︎
mircea_popescu: quite. though usually in the manner of "i just want my pension"
asciilifeform: 'leave town before check bounces' aha
a111: Logged on 2017-09-01 20:51 asciilifeform: ( the www people simply pissed on the problem, african-style, and didn't even pretend to solve )
mircea_popescu: http://btcbase.org/log/2017-09-01#1709173 << this is how everything was ever done in the field. ☝︎
mircea_popescu: good thing we didn't read the pdfs.
asciilifeform: where whatshisface 'proved' that public key signatures are impossible w/out hashing.
asciilifeform: lulzily, the above result contradicts the 'literature'
mircea_popescu: aite then
mircea_popescu: so as i was saying -- we don't have to usg.dumb sign anything.
asciilifeform: mircea_popescu: there's potentially infinitely many ways to lubyate
mircea_popescu: you can chose an arbitrary large prime and have that arity.
asciilifeform: for n00bz, i'll review : cut into chunks C1,C2,...,Cn; then sign, S(C1 xor C2), S(C2,C3,C9999), S(C5), etc. with the 'soliton' providing the arity
mircea_popescu: and i don't think it's necessarily a bad solution, which is why i said there are others.
mircea_popescu: no, i was expecting that's what you want ever since pushing me above.
asciilifeform: btw mircea_popescu it so happens that i have what looks like a solution to 'sign a TB with a 4096bit key with no hash'
asciilifeform: i'd like to find what trinque had in mind re accounting for collisions. ( but let's encourage him to put it in the article! )
mircea_popescu: for the record, si heisenberg limit is 10^-34
asciilifeform: ( and i say this after making 'v', where proposed to put all of human knowledge to sit on hash! lol )
asciilifeform: we have no such thing for hash.
asciilifeform: you know roughly how much stress a thing can take, before you put stress on it
asciilifeform: in other walks of life, when, e.g., building bridge, there is a notion of material strength
mircea_popescu: asciilifeform the fraction is so small as to be below heisenberg limits. this matters.
trinque: and they republish what they've got. like all history.
trinque: worst case is the original publisher is dead, and all you've got are whoever cared about him.
mircea_popescu: tell you what, i also have no proof in your sense the slavegirls ain't cheating on me.
trinque: penultimate worst case is that folks republish their items in the case of hashpocalypse. and they tell their trusted peers. and it propagates over wot.
asciilifeform: just as there is not for block cipher.
asciilifeform: there is NO theoretical proof of strength for any hash algo
mircea_popescu: the important point is : acordeon hash function set up to spit 256 byte or 512 byte hashes WILL NOT collide.
mircea_popescu: we're talking of different things in the sense you went off on a tangent.
mircea_popescu: there's no shits given re tmsa-rsa-signature being retrocompatible.
asciilifeform: yes but you cannot retro-elongate hashes of the past.
mircea_popescu: nobody forces us to use the 32/64 byte crap nsit likes.
asciilifeform: hash collisions cannot be fixed by 'we use THIS OTHER hash NAO! haha' because now enemy can monkey with the past.
mircea_popescu: so, message as chain of message blocks, with acordeon, settable hash. yes it is "possible" to get a collision. if your hash is 512bytes long... not soon.
mircea_popescu: "hurr durr we know how to sign "cocksucker" whenever they use 93 bit outputs. which is... not now."