log☇︎
166500+ entries in 0.098s
asciilifeform: somehow a centralized spamola token that can be magicked into existence in arbirary quantity, is 'same' as bitcoin ?
r0nin-: well i agree, the SEC has zero say with bitcoin because it simply isnt a security
mircea_popescu: r0nin- that end is not the problem. the problem is that usg has no better claim to ontology than a random wow guild.
r0nin-: bitcoin is exactly the same thing as 'in-game currency'
asciilifeform: r0nin-: 'also true of in-game currency' did your mother drop you as a baby? or feigning retardation ?
r0nin-: bitcoin is the equivalent of game tokens
r0nin-: thats precisely true
mircea_popescu: anyway. fellow should read the above linked item and the discussion of causes and purposes, and try again.
r0nin-: Furthermore, while it is true that Bitcoin may be regarded as an item of value by one or more people, this also is true of the in-game currency of each and every single one game ever released to this date and to be released for the conceivable future.
mircea_popescu: you're not doing too well here.
mircea_popescu: r0nin- re the "entitle", he rehashes (without attribution) a point actually made on trilema back in... 2012. http://trilema.com/2012/the-reasons-why-bitcoin-securities-cant-be-regulated-by-the-sec/#selection-129.0-129.112
r0nin-: a USG bond is just a time deposit
r0nin-: well he doesnt say that
asciilifeform: 'usg bond is better than btc because It Entitles You To'
asciilifeform: reads like a typical propertard
mircea_popescu: eh what the hell. lessee.
mircea_popescu: what's that ?
asciilifeform also entertained that this is almost the exact arg of mircea_popescu an' asciilifeform re mphf, originally, but with the sides reversed
asciilifeform: currently it isn't clear to asciilifeform what is the equiv. of the bullet weight ( or other mechanical cost ) here.
asciilifeform: all the ones you dun use -- are free
asciilifeform: for what are they being saved by using undersized rsanuke ?
mircea_popescu: and weaponry is to be used as per fucking manual and design, not obliquely wtf.
asciilifeform: idea is not to impress, but to cauterize
mircea_popescu: i am not in love with soros what can i tell you.
asciilifeform: if you were dropping the n00k on soros, would you settle for 'sufficient' 10ktonne, or opt at slightly moar cost for 100mtonne kuzkina mat' if available ?
mircea_popescu: the ustards tried, with their golden toilets. didn't work out for them. meanwhile irl i still pack the new girls two to an apartment, as per http://trilema.com/2010/doua-fete-argumentul-economic/
mircea_popescu: beleive it or not this is the logic i use all the time. what, you think rich people get a special science / engineering that works differently ?
asciilifeform: it is pretty strange imho to read mircea_popescu operating with the logic of poverty -- '100 horse suffices for standing at red light, why would anyone want a bugatti', '400 sq metres suffices to house a family' -- how is 4096b rsa any different ?
mircea_popescu: alternative candidates have nothing speaking for them.
mircea_popescu: this is a sufficient reason to choose that value.
mircea_popescu: 4096 is the smallest power of two which produces a rsa key that is, factually, incomputable.
asciilifeform: ain't as if people were born with 32bit cpu stuffed up their arses, neh
mircea_popescu: increasing tank size up to the 4096 bit tank pays off. increasing tank size past 4096 bit tank results in fucking panthers and whatnot.
asciilifeform: asciilifeform's argument is not even 'we must 8192!!1' but that 4096 is an accidental measure, like the volume of one particular bucket one particular child left on a beach
mircea_popescu: asciilifeform nevertheless, there is an optimal bullet per target, and nobody makes the calibre twice as big. if not because their mind works enough to understand why not, then because the very sad experience of the nazis illustrates why the fuck not.
mircea_popescu: this is nonsense argument, and in fact they could "afford" both but they don't fucking use both. they use the correctly sized one based on human thickness and wetware bullet refraction properties.
asciilifeform: funnily aviation cannon designers had this exact convo!
mircea_popescu: asciilifeform this is EXACTLY not the case. i am saying "7.62 is sufficient, because it will blow a hole through man, as result of interplay of actual universal constants" and you are saying "yes but 15.2 would be bigger".
mircea_popescu: the man who lost 100 bits is the same man who will lose 200 bis.
asciilifeform: what i see is an elementary '7.62mm is inevitably better than 5.56, if you can afford either'
mircea_popescu: the way this conversation is going, we can't have children us two until we get the baby room properly furnished and the safest crypto op is one that never finishes. ☟︎
asciilifeform: that is proportionate to length of key.
asciilifeform: that's where i was going, the fuckyou joe, that a 'hot' key may have a finite life.
mircea_popescu: to quote that film i reviewed recently, "what do they want for their lousy 35 cents ? to live forever ?"
asciilifeform: in the sense of making the probability 0.
asciilifeform: it isn't clear that there is a 'fix'.
mircea_popescu: if ~you~, joe, leak bits, then you, joe, are responsible, and you, joe, pay for the fix. not the fucking cryptosystem.
asciilifeform: ( for reference : leaking 1/4 of your privkey is fatal; 10-20% -- you're breakable in polynomial time with large -- how large is unknown -- constant factor )
mircea_popescu: but there shall never be a systematic solution for tactical failure.
asciilifeform: mircea_popescu: this argument is not false, but ignores noncryptoanalytic breaks
mircea_popescu: 4096 is the largest power of two which makes a key that is incomputable.
mircea_popescu: in any case, entirely incomparable items. yes, people do dumb shit all the time. but this doesn't mean people doing the right thing should do it once and wear a vest "to make sure".
asciilifeform: i dunno why mircea_popescu is defending a figure that came from wholly arbitrary, rather than natural, constraint ( 'what is fast on a uniprocessor x86 box' == 4096 )
asciilifeform: ( recall the winxp battleship etc )
asciilifeform: the cost of swapping out the nonsense will easily exceed that of, e.g., vietnam war
mircea_popescu: the "two digit year" thing is incomparable to the 4096 bit thing, because the 4096 bit thing is backed by "the whole universe, if it set to computing, still wouldn't EVER get even close to breaking a key".
asciilifeform: or the ones who put win95 in roms
asciilifeform: it dun make the people who used 2-digit years in rom code, any less stoopid
mircea_popescu: asciilifeform yes, and that's the discussion here. if it's costly in excess of utility then the WHOLE thing is broken, and "fixes" in the sense contemplated merely further the breakage.
mircea_popescu: there is no such thing as man-made transcendent.
mircea_popescu: ie, currently they don't exist. later on, they will. later still, they'll break. laterer yet they'll be replaced.
mircea_popescu: asciilifeform they can be replaced like any other man made item.
asciilifeform: anythign that runs from rom.
mircea_popescu: now, the above is an opinion you or anyone is welcome to hold, like any other opinion. i however don't hold it.
asciilifeform: mno. observation was re concrete cases where it is impractical to change a key. say, the parasitic radio item.
mircea_popescu: the point remains : if what you are saying it's not practical to change keys (ie, that the per-op value add of key is less than the per-op value lost to suspected leak) THEN therefore what you are saying is "rsa is not worth using"
asciilifeform: every time we rsa
asciilifeform: we're scooping out a swimming pool with a teaspoon
asciilifeform: will point out that 8192+ only seems expensive because x86 is pessimized for it
mircea_popescu: which is what's contemplated here. "oh, changing keys (ie, buying house) is impractical. therefore, let us take out mortgage (pay a little evert time we use key), because it's less affordable but more accessible and the woman is too dumb to understand the difference and we're too castrated to not let her run the household)
asciilifeform dun follow the analogy
mircea_popescu: if you can't afford the house, you can even less afford the mortgage.
mircea_popescu: understand asciilifeform : if you can't afford to buy a house in cash, this means YOU CAN'T AFFORD THE HOUSE. period. go sleep under the fucking bridge. and no, the fact that you ~could~ get a mortgage pay monthly DOES NOT!!!! make it now possible.
asciilifeform: actually thicker keys are expensive to n*log(n)
mircea_popescu: if THAT isn't practical, then sure as fuck larger keys aren't practical.
mircea_popescu: but that is practical to the power ^1, whereas thicker keys expensive to ^ of n.
asciilifeform: mircea_popescu of all people knows this.
asciilifeform: not all keys are practical to change.
mircea_popescu: if you believe that, the correct solution is to change keys, not to make all the keys randomly thicker.
asciilifeform: and little enough is know re reliability of current liquishit hardware, that will not propose to calculate it.
asciilifeform: other problem is that there is a small but finite probability of misfires in rsatron , such that would leak a bit.
asciilifeform: ( adding 0+0 still uses slightly different current than maxint+maxint )
mircea_popescu: very possible, just don't throw away your old car batteries.
asciilifeform: 1 problem is that it is not actually possible on stock x86pc to not leak power differential.
mircea_popescu: so it being the fundamental reasoning, it's ok to have it explicit, but implicit or explicit it ain't gonna budge.
asciilifeform: but imho 'better to be rich and healthy, than poor and diseased'
asciilifeform: sorta why i put the effort into crafting a demonstrably bug-free constantimetron, first thing, rather than 'let's use gnu gmp with massive keyz'
mircea_popescu: so if you're gonna plug, plug the right end of funnel.
mircea_popescu: asciilifeform yes, but good pipe is finitely good ^ 1 whereas tick pipe is expensive to the ^ n.
mircea_popescu: (honestly me thought the above is the tacit understanding at the base of the effort)
mircea_popescu: because expensive as it may be, it is WAY the fuck cheaper to make good pipe than to make thicker pipe to cover for all the leaks due to it being shitty pipe.
mircea_popescu: which is why we're putting all this crazy effort into proper rsa, ffa etc.
asciilifeform: will note that there are other ( noncryptographic ) ways to leak bitz
mircea_popescu: yes, but my point here is that it is perhaps a defensible heuristic, but not defensible reasoning, to expect that just because p, q and p*q are coupled in that manner, any application to do with their coupling therefore informs you as to p or q.
asciilifeform: this output doesn't contain anything you did not already know on demand tho
mircea_popescu: by encrypting "hola bro" to alf's key and getting $result, i now know a novel fact about alf's key!
mircea_popescu: so there, "a novel fact"
mircea_popescu: you know for a fact the encrypted text will be decryptable by the key, yes ?
mircea_popescu: not moreso than encryption does!
asciilifeform: sig gives you a novel fact re the private exponent