log☇︎
173200+ entries in 0.106s
mircea_popescu: he can do whatever he pleases. ideally stop with the inane bullshit.
asciilifeform: why would he keep it plugged in if you formatted and installed slave-for-tmsr-os instead of all that stood there
mircea_popescu: victim is the windowshead.
asciilifeform: which implies that user keeps the machine plugged in
asciilifeform: in original description of uci, was described as 'cpu cycle by the pound' sort of affair
mircea_popescu: the end game is, your computer was liberated from the crap. your "data" made on your "software". no need for either, grow up.
mircea_popescu: this whole "bitcoin ransom" bs is just that, an ouverture.
a111: Logged on 2017-07-15 05:27 mircea_popescu: !#s "can we have everything louder than everything else"
mircea_popescu: i see absolutely 0 reason of tmsr tolerating non tmsr software on same box. wipe it all on sight.
asciilifeform: any and all of which look for other processes and occasionally snapshot to send back to $mothership - 'perhaps interesting', 'perhaps contains keys'
asciilifeform: let's take the typical bot, of which existing nets ( of infected nodes, rather than 'voluntary' p2p ) are made
mircea_popescu: whether server A im talking to i'm talking so it adds 2 and 2 for me, or gives me a new friend, or shoots a missile, is not a part of design consideration.
mircea_popescu: you accomplish this by not sharing your S lists.
asciilifeform: you also want that nobody can affect how the net works by bringing 9000 of his own As to the table.
mircea_popescu: but yes -- the standard here is that at no point will you know (or care!) whether box A is box A or box E.
asciilifeform: ( everyone is also subject to piano falling on him )
asciilifeform: in practice, some folx moar often than others.
mircea_popescu: the sad truth is that everyone is subject to key compromise at any time.
a111: Logged on 2017-07-20 18:56 mircea_popescu: is this "how to run software on block of wood" ?
mircea_popescu: asciilifeform by now we are strasight back to http://btcbase.org/log/2017-07-20#1687375 ☝︎
asciilifeform: it is also subject to key compromise at any time.
mircea_popescu: the solution is that some "people" will accept noobs and some won't. in practice, everyone will run multiple interfaces and thassat.
asciilifeform: i do not know of a solution to this presently
asciilifeform: and cannot operate by 'nothing to allcomers'
asciilifeform: which is that new infections are not physically distinguishable from sybils
mircea_popescu: but -- the fact that A happens to already have a connection to B and therefore B can S for him does not imply or require either A or B be not behind a nat. ☟︎
asciilifeform: there is a separate problem pertaining to uci, that does not reduce strictly to 'nat problem'
mircea_popescu: any C could at any point be a S or not be a S (and because of encryption observer can't even tell which it is)
asciilifeform: and ergo net would have to be deployed in massive burst, whereby yesterday there were 0 S, but today 9000. ☟︎
mircea_popescu: this is not what i see. what i see is that you have a list of C1..n machines ; some of which are included arbitrarily in a list of S1...j and some of which are not.
asciilifeform: my objection is that the star topology does not go away by virtue of being called something else. you have a class of fully-capable nodes, S, that can accept unsolicited pipes; and a much larger class of P ( lessay, peons ), behind NAT, who cannot. ergo stars.
mircea_popescu: and yes the fact that the list of S must be at least a dozen different lists and each a grosse long is why we can't do this arbitrarily early.
asciilifeform: i'll point out that i do not know of a different recipe than mircea_popescu's ( and in fact is same used by 'skype' and various heathens , for many years. )
mircea_popescu: back to it -- once A knows what ip B is at and B knows what ip A is at, they can exchange crafter packets to the point of parlaying it into an established connection. in order to find these, they both rely on their peers. if they have no peers, they don't properly speaking exist, and an angel from above in the shape of a tmsr lord will have to bless them with the spark of life.
asciilifeform: sorta as british admiral answered 'boil the ocean' to 'is there a solution to uboat'
asciilifeform: it's the 'insect' solution, gotta mention it for completeness.
mircea_popescu: i hear apple stocks them.
mircea_popescu: also gives the state of the art re mate finding in the greater uslandia. and it's the lulz of all time.
mircea_popescu: it is NOT A THING.
asciilifeform: existing botnets do this -- the familiar port scan and hammer
asciilifeform: i.e. to shout
asciilifeform: it gives the current state of the art re 'how to connect without introduction'
mircea_popescu: amusingly i also put it in the log earlioer today. so what of it ?
a111: Logged on 2017-05-23 23:34 mircea_popescu: what fucking fix. they can't open it anymore than they can fuck the woman on the deserted island. "nobody to introduce us!!1"
asciilifeform: ( as alluded to in http://btcbase.org/log/2017-05-23#1660666 ) ☝︎
mircea_popescu: a better world for computers where the rapist-racism of needing someone to introduce you has been resolved is truly not in any way different than the same re people.
mircea_popescu: there's no other way socialisation works than by introduction. this is not a problem. you are welcome to approach it, but you will be wasting your time.
asciilifeform: this is == to sina's 'you used gpg yesterday, why is using it tomorrow a problem'
mircea_popescu: yet, notwithstanding sina's concerns, we did manage to bootstrap trust out of nothing.
mircea_popescu: and, to stick to fundamentals : your whole argument "proves" that #trilema could not exist.
asciilifeform: this'd require at least 2 S to exist, lol
mircea_popescu: what creates massive stars in practice is that the geeks doing this have no peers.
asciilifeform: speaking here of topological problem that will remain after rewritten by people.
asciilifeform: it is a problem in the same way as it is for trb.
mircea_popescu: asciilifeform this is only a problem in the sense of "i have decided to not work today so i can't"
mircea_popescu: ". A aiming to connect to B or vice-versa for any reason is accomplished through : A asks C for an ip for B. C provides A with ip for B and B with notification of A's request (at the least, its ip)." << A sends C a request. C sends A an ip for B and B an ip for A. both B and A start broadcasting at each other fixed packets until the conn is established.
asciilifeform: the problem -- and it IS a problem -- is that cold bootup, for ~any~ node, requires knowing a working S.
mircea_popescu: maybe we need to revisit the introduction part ?
asciilifeform: well yes. if you were able to connect, you were... surprise... able to connect
mircea_popescu: as long as at least one of them actually works, the box is now connected. and will stay connected.
mircea_popescu: you put, on the box you own, whatever list of your peers as you deem fit.
mircea_popescu: you can put whatever you want in that list, and should put plenty of crap.
asciilifeform: and how exactly is this achieved ?
asciilifeform: not if he has to connect to one of $smallint S boxes. that are all firewalled off at his isp.
asciilifeform: if serf can't get on the actual net, this is not something that can be solved ~for~ him by anyone else
a111: Logged on 2017-07-20 18:50 mircea_popescu: you have nodes N1...Nn all behind a nat. Server S popus up and connects them in pairs such that no Ni is without at least one peer. Server S dies. The whole N1..Nn remains connected forever, notwithstanding how any NjNk connections may die.
mircea_popescu: there is. again in the vanishing hope that may help : http://btcbase.org/log/2017-07-20#1687348 ☝︎
mircea_popescu: "the web" is the result of better things shut off.
mircea_popescu: asciilifeform not at all. strictly like as you see all around today.
asciilifeform: ( if somebody wants to describe how , i'll listen )
asciilifeform: it can't really be vanished away without killing 'tcp to arsebook' etc also. as i currently understand it.
asciilifeform: 'hole' is the wrong word re how udp holepunch works
mircea_popescu: remember when amd would never have dared stopping contributing to coreboot because "public opinion" ?
trinque: which amusingly enough is emitted right before this comment in pf.c: /* This really shouldn't happen!!! */
mircea_popescu: "oh, the soviets could never x because market". da fuck do the soviets care.
mircea_popescu: nonsense. "the eternal hole" is worse delusion than proper fuckoff.
asciilifeform: ( and other tcp )
mircea_popescu: necessarily. if you grant nat implemented correctly, your problem is exactly that of the british girl on the deserted island. "so how was the sex ?" "i dunno, no star topoloigies!!!"
asciilifeform: there are some invariant 'holes' that will never close so long as the thing has to pass www
trinque: pf: pf_normalize_tcp_stateful failed on first pkt << deedbot departures appear to correspond with this. ☟︎
mircea_popescu: asciilifeform "actual solution" will be exploiting hole in nat which depends on nat staying stupid in that specific way, which pwnat experience shows is worse than nothing.
mircea_popescu: because sometimes that's exactly what it is neh.
mircea_popescu: hey. the world runs on fuckoff since ~forever.
a111: Logged on 2017-07-20 18:50 mircea_popescu: you have nodes N1...Nn all behind a nat. Server S popus up and connects them in pairs such that no Ni is without at least one peer. Server S dies. The whole N1..Nn remains connected forever, notwithstanding how any NjNk connections may die.
mircea_popescu: can "problematize" about it all you want, but when you're done -- the solution is to http://btcbase.org/log/2017-07-20#1687348 ☝︎
mircea_popescu: there's temporary Sn. this is quite literally just how life is.
mircea_popescu: "i have boxes that don't respond to unknowns and no way to make unknowns known can has solution ?" alfian style problem.
asciilifeform: there must NOT BE any such thing as 'S'.
asciilifeform: can haz, or not, is a question that can be posed. but to give a solution WITH stars, is neither here nor there.
asciilifeform: i want a nonstar topology.
asciilifeform: it ain't a solution if it merely redefines the problem as 'the way life is'
mircea_popescu: you keep wanting the solution expressed in different terminology and upon delivery promptly proceed to change the terminology.
mircea_popescu: i dun follow what the problem's supposed to be with the above scheme.
asciilifeform: but the problem is not solved by pretending it not to exist.
asciilifeform: mno. this is the hypothetical of 'how to get properly p2p net'
mircea_popescu: is this "how to run software on block of wood" ? ☟︎
asciilifeform: but if you propose to package it with a hardcoded list of existing peers -- guess what, this is what people traditionally do. they get unplugged, and then whole net is useless.
mircea_popescu: a box you own can connect to the network through the usual process : asks all your peers for peers to connect to.
asciilifeform: mircea_popescu: let's concretize . say i write an infector. what should be its mechanism of connecting to greater tmsr ?
asciilifeform: ( there are various 'cheats' around this, but all resolve to hardcoded hidey-hole ~somewhere~ that contains where next to connect )