tmsr - Search: " T"

469800+ entries in 2.288s

jurov: anyway it's linked from there

jurov: um.. i wanted the "first steps"

kakobrekla: any honest to be found is there

Kalki: Pretty newb around these parts but find it quite refreshing as I have been swimming in the shallow btc waters looking for honesty

jurov: quite an arrangement, this ☟︎

jurov: lampelina can !down you then

kakobrekla: lesson learned; never talk.

jurov: glad to be helpful

kakobrekla: anyway someone else can run proxy title parser and ill use that

jurov: it did not fetch the title

kakobrekla: !s test

jurov: or maybe it's just slow, is there some attack?

jurov: yes i'm curious what will be the best sampling method.. timing the pulses with some fast counter? ☟︎

asciilifeform: you'd have to use intervals between spikes as the entropy

ben_vulpes: was going to wire it up to a microphone

ben_vulpes: no hw with gpio, but now that you mention rpi mebbe i should try that

jurov: how did you hook up the rng in the end?

ben_vulpes: "use libtool" << so as to speak "just type libtool at the prompt"

ben_vulpes: <jurov> and run ldconfig << that was it

jurov: any time libraries are installed, ldconfig should be ran

trinque: ben_vulpes: lemme fix a thing and then I'll see how that compares to my build of dieharder

trinque: now everybody knows lisp has weirded up the way I do my parenthesis

ben_vulpes: thanks for the python bro

ben_vulpes: nah this is gentoo

ben_vulpes: http://dpaste.com/1SJBXB5.txt << previous experience lead me to expect `make install` to, well, install all of the relevant things. trinque, asciilifeform: any obvious mistakes or bad thoughts on my part you'd care to rectify?

asciilifeform: and pwned by no less than weev himself? l0l

asciilifeform: l0l! that thing actually existed as something other than spam ads ?! ☟︎

asciilifeform: and that no one attempted to actually derive the corresponding privates

asciilifeform: i still regard it as not entirely unlikely that the attack is a pure 'dos'

Apocalyptic: quite amusing that 281479271743489 has 65537 as one of its 3 prime factors, those cosmic rays have a sense of humour ☟︎

asciilifeform: (and they make pgpdump unhappy, in various ways, yes.)

asciilifeform: incidentally, now is a good time to mention the fact that sks also contains a great many keys which were too malformed to even attempt phuctoring

asciilifeform: Apocalyptic: not at present. will probably end up spending some quality time with the sourc

Apocalyptic: asciilifeform, by the way do you have any explanation as to why pgpdump skips the KeyID field on invalid subkeys ?

asciilifeform: how that one was made - remains unknown

asciilifeform: notice it does not include the faux-anvin

asciilifeform: Apocalyptic: you are correct. only applies to -every key in The List-

Apocalyptic: "RSA e(17 bits) - 01 00 01" on the last two Public Subkey Packet

asciilifeform: Apocalyptic: at present, we have sufficient info to birth these on demand from legit keys. but we do not know the magic trick which yields decryption (it does not necessarily take the shape of a private key! and may not even -exist- for every possible starting legit key)

Apocalyptic: asciilifeform, even the HPA's invalid subkey ? I thought pgpdump showed me e = 65537

Apocalyptic: asciilifeform, If I get some time I will, it requires a more carefull analysis that it may seem

asciilifeform: Apocalyptic: all of the ones discussed thus far.

Apocalyptic: "2) The public exponent is changed to 281479271743489" for which key ?

asciilifeform: Apocalyptic: care to comment on the 'is transformation same as picking random N for modulus' thing ?

asciilifeform: (incidentally, strong's 'experimental physics' (see logs) describes how to make one)

decimation: yeah I suspect that most nuclear facilities have golden toilet budgets for fancier stuff

Apocalyptic: "Similarly, the modulus of Peter Anvin's key is also divisible by 19 and 7704959, in addition to 3, 7 and 11 as" // I have found these two additional primes as well, and only those 2

asciilifeform: geiger tubes are no more difficult to make than light bulbs, but someone has to -do- it

decimation: which probably have some variation themselves

asciilifeform: almost entire world is relying on the - finite and mostly russian - cold war supplies

decimation: no my understanding is that folks are 'recycling' old soviet tubes mainly

asciilifeform: phun phakt: virtually no one is still making geiger tubes commercially

decimation: yeah that makes sense

asciilifeform: in the case of geiger tube, the fill gas is also 'poisoned' by said process, iirc

asciilifeform: filament 'sputters' onto the walls

asciilifeform: it wears out in the same way as a fluorescent light bulb does

decimation: oh yeah that sucks

asciilifeform: that is really the main reason i do not use geiger for rng personally

decimation: that's true, it is

asciilifeform: for another, the process of geiger tube going off is rf-noisy

asciilifeform: for one thing, the material, well... decays. at predictable rate.

decimation: soon the hunter 'comes to expect'

decimation: asciilifeform: your earlier analogy of the deer kneeling before the hunter is apt

asciilifeform: decimation: there are some 'gotchas' to using geigers

decimation: if someone is interested in the radioactive decay, I would look at herr walker's hotbits > http://www.fourmilab.ch/hotbits/

asciilifeform: no one - certainly not 'hackernews' imbeciles who ship their hardware with winblows drivers because 'consumer has come to expect' - give a flying fuck about auditability, etc.

asciilifeform: apparently i can discuss my designs all i want without fear of being 'scooped' because 'everybody' thinks i'm entirely batshit.

decimation: the only other method I can think that is reasonably acceptable in this way is radiactive decay

decimation: part of the problem is as ascii has identified - you really want every part to be inspectible by the end user

asciilifeform: jurov: and yes, they are typically 'best practices' retarded - e.g., 'whitening' etc

asciilifeform: jurov: realize that folks have been selling rng since at least late '70s

jurov: incidentally, this appeared on hn today: https://www.tindie.com/products/ubldit/truerng-hardware-random-number-generator/

decimation: asciilifeform: not to mention the nmi

asciilifeform: (the sks mass-submit, that is)

asciilifeform: jurov: the submission thing is a hastily-written script.

jurov: i'm not talking about immunity. only restartability

asciilifeform: jurov: and i will point out that it is quite impossible to build a guaranteed power-cycle-immune anything on a standard x86 pc, which doesn't even have a powerfail interrupt

jurov: dunno where i got the notion from

asciilifeform: jurov: the thing has no lisp or anything else interesting in it

jurov: if they cycle you once in a week

jurov: you haven't used the hyperadvanced lisp runtime that can be snapshottet and restarted?

asciilifeform: aaaactually my understanding is that one of the 19 got requeued, which 'un-broke' it. say hello to our first annoying bug.

mod6: jurov: ok added link to that under "Links". That ok with you?

asciilifeform: and the pressure on sks server keepers to start removing things will be turned on, yes.

asciilifeform: incidentally, now more than ever is the time to set up an sks server under our control. because sks is our ultimate 'backup' ☟︎

asciilifeform: ;;later tell mircea_popescu aaaand somehow '19' turned into '18' !?! (i do have a db backup, so if this is enemy action they can go fuck themselves.)

asciilifeform: ;;later tell mircea_popescu thing is still largely loading moduli missed (due to db locking idiocy) the first time around.

mod6: ok cool. i'll add a link up there in just a bit.

mod6: heheh, redirect works fine with lynx, seems to not work well with FF. ☟︎

jurov: you can use the latter link

jurov: it should redirect to http://btc.yt/lxr/satoshi/source/

mod6: nothing seems to come up at that link?

jurov: lol the phoundation released 5 versions (incl. rc) since feb?

mod6: oh, now... no we don't have a link to that thing. what is it anyway, i can't recall.

jurov: or to the lxr source browser

mod6: yeah, the url is: http://thebitcoin.foundation/chicken.tar.gz

jurov: that's what most of the patches would be based on

jurov: links can be generated but imo they are not needed

jurov: mod6 i was thinking just a version name for both columns

mike_c: "Excitement!!! Excellence! Launching ahead to better humanity! Ethereum!!! Congratulations!