log☇︎
73900+ entries in 0.044s
mircea_popescu: pretty sure gb nic works on ye olde opterons just as fine. raids are iffy, depend on all sorta bs ; but definitely exist such thing as proper raid controller in 2005.
asciilifeform: and gb nic. that's pretty much it.
mircea_popescu: in any case, served a lot of trilema over the years.
asciilifeform: mircea_popescu: the oldest i have on currently serving box anywhere is 3.8.8
asciilifeform: mircea_popescu: not being a debianist, i cannot immediately say which edition of it this'd correspond to
mircea_popescu: just trying to tease out which version.
mircea_popescu: asciilifeform 2.6.something ? or which are you thinking ?
asciilifeform: i'll consider backports of patches that fix e.g. 'magic packet of death', but not 'spectre' etc rubbish
asciilifeform: any attempt to repair it simply creates a nsa.nobus imho
trinque: aite, then that'd be the only thing coming to mind to prevent pulling an older vintage
trinque: asciilifeform: what's the word on this "speculative execution" thing? dun matter?
asciilifeform: but would like to hear from the l1 folx re subj before mixing the cement.
asciilifeform: i wouldn't even object to cementing a 3.x, supposing i had a working one for rk
asciilifeform: 4.9.76 on dulap, and slightly earlier iirc in rk branch ( will have to maintain at least these 2 forks , i suspect )
trinque: the kernel in current cuntoo is 4.9.95, by virtue solely of being when I took the snapshot ☟︎
asciilifeform: ( oblig shitty engl. trans, http://samlib.ru/a/as_w/vys-chr-ru-e.shtml )
asciilifeform: or at least, to start.
asciilifeform: likewise it is time for folx to go over their iron and nominate driverolade that they believe ought to live.
asciilifeform: trinque, mircea_popescu , et al : asciilifeform, in light of http://www.loper-os.org/?p=2578 etc , has been thinking of genesising the kernel. prolly not tonight or tomorrow, but in near fyootoor. if anyone has particular desires ( & optionally, justification ) for particular ver becoming The Last and cemented, would like to hear about it.
deedbot: http://bimbo.club/?p=27 << Bimbo.Club - TMSR Log Summary - 9/18/2018
asciilifeform: aaha.. thinking folx, sometimes give up; vermin -- never
mircea_popescu: much like just because we have penicillin dun mean tb gave up and left the planet.
a111: Logged on 2018-09-24 17:13 asciilifeform: even per this 'logic', seems like there's a loose end -- so, ok, they sell, SELL, only. bags of benjies -- go in. ~then~ wat.
mircea_popescu: http://btcbase.org/log/2018-09-24#1853684 << just because mercantilism got killed in the 1600s dun mean the sort of brains that naturally spawn mercantilism as a state-of-the-art mentality stopped being poopped. ☝︎
a111: Logged on 2018-09-24 17:08 asciilifeform: mighty lotta cheek, for folx who torment a 300 $ crate in tariff hell, to prattle about 'international marketing'
mircea_popescu: http://btcbase.org/log/2018-09-24#1853678 << it's just the cunt signal, "what to do instead of whoring if you're a hottie teenager" ☝︎
a111: Logged on 2018-09-24 17:07 BingoBoingo: International marketing, networking, and the usual bread and butter. There's an advertisement station in front of the closed zoo that they seem to favor.
a111: Logged on 2018-09-24 17:02 asciilifeform: relatedly, phunphakt, all of asciilifeform's packets to/from BingoBoingostan, relay through 4.68.72.61 : nsa station in miami.
mircea_popescu: it's just not how it works, "nobody could have seen through shitsoup!!!". really ?
asciilifeform: i.e. may as well gather the win to be had from firing the first rockets.
a111: Logged on 2018-09-24 16:15 asciilifeform: and theoretically if the cartel births '9000' new sybils, it will force the use of something like my algo, whether anyone wants to or not, the layer of sibylade would become undrillable in its absence . but this is admittedly a stretch.
asciilifeform: rright, i observed that the situation could be forced into being very inexpensively, http://btcbase.org/log/2018-09-24#1853644 ☝︎
mircea_popescu: asciilifeform i suppose it is a serious bojum that is in no way resolved by smearing shit over eyes.
a111: Logged on 2017-12-18 16:07 BingoBoingo: Argentina's slogan is the lie "Argentina no es un pais pobre", Montevideo's is the truth "Montevideo es muy tranquilo"
a111: Logged on 2018-09-24 17:00 BingoBoingo: Oh, they'll flat out say Uruguay is a poor country
asciilifeform: mircea_popescu: i take it you dun suppose the miners thing , upthread, will be serious boojum ? ( or should i wait for mircea_popescu to eat log )
mircea_popescu: but hey, "opted" to not have wot presence, just desserts.
mircea_popescu: http://btcbase.org/log/2018-09-24#1853649 << entirely usg's desperation. they're losing preminence/getting locked out of markets at a shocking rate for a "world power" not defeated in war. ☝︎
mircea_popescu: http://btcbase.org/log/2018-09-24#1853635 << the universal problem with "progress" in a functional world -- it manages to turn to shit. ☝︎
a111: Logged on 2015-05-31 12:12 mircea_popescu: say something like : every bad block received, -10 points. every minute where connection is kept at 80% of its allocated bw or over, 1 point. every hour score decays 1% towards 0, be it either positive or negative.
a111: Logged on 2015-05-31 12:11 mircea_popescu: with a time decay, with a penalty for sending bad blocks and a positive for sending good blocks at a good speed.
a111: Logged on 2015-05-31 12:11 mircea_popescu: http://log.bitcoin-assets.com/?date=31-05-2015#1148837 << i would like to see a patch which maintains VALUED list of other nodes.
mircea_popescu: http://btcbase.org/log/2015-05-31#1148935 << FUUCK, found! this is what i was looking for, http://btcbase.org/log/2015-05-31#1148937 http://btcbase.org/log/2015-05-31#1148938 ☝︎☝︎☝︎
a111: Logged on 2017-01-03 20:33 mircea_popescu: otherwise it is discarded. B.T may be pruned (according to arbitrary address list, for instance). Rate limiting in TRB.N may be constructed to observe N.B items that fail to propagate to B.B and ban the originating peers.
mircea_popescu: http://btcbase.org/log/2017-01-03#1595770 << alluded to in the more specified TRB.B/TRB.N prototype design. ☝︎
a111: Logged on 2015-01-31 03:13 mircea_popescu: something simple like, "whenever a parentless block is handed over the retaining of which would cause memory pool for holding parentless blocks to be overrun, a) drop the handed block ; b) close the connection and ban that peer for half hour ; c) discard all chains of parentless blocks longer than six items ; d) connect again"
mircea_popescu: http://btcbase.org/log/2015-01-31#999477 << early version of the idea. ☝︎
mircea_popescu: infuriatingly enough, i can't find it now. but the discussion was exactly along the same lines, "add user-configurable penalty for communicating shitblocks (any block that isn't the one you were looking for) and any other misbehaviour ; and used-configurable bonus for communicating useful things ; then user-configurable knob for lowest-tolerable-score and ban peers who fall under"
a111: Logged on 2015-08-01 00:30 asciilifeform: thing is, this'll result in a very random selection at present
a111: Logged on 2015-08-01 00:30 mircea_popescu: (original design was to simply shutter all connecting peers that fail to provide good $$$ txn, in some sort of %)
mircea_popescu: http://btcbase.org/log/2015-08-01#1220066 / http://btcbase.org/log/2015-08-01#1220067 << possiblty this entire thread was exactly it. ☝︎☝︎
a111: Logged on 2017-02-27 22:07 mircea_popescu: now, the historical solutionb to the problem, as well as perhaps a workable solution here, is the intrinsic oracle. if user relays txn to a node WHO MAKES A PROMISE (such as for instance "the txn will be included before block n" ?) then the nodes can be scored by their oracle value ("what he said turned out true!) and suddenly you have a more meaningful node market.
mircea_popescu: http://btcbase.org/log/2017-02-27#1619483 << one item, somewhat related, not the item sought though. (this is to tightnen miner-node paths) ☝︎
mircea_popescu: http://btcbase.org/log-search?q=from%3Amircea+trb-i << interesting list, at that.
asciilifeform: iirc was one of the moar recent 'trb-i' mircea_popescu threads
mircea_popescu: let me try an' fish it out
asciilifeform: ( but as it is, we also have a wild animal, in the mix, called miner, and its immediate ecosystem )
mircea_popescu: but in order for it to work -- yes very much need nodes that don't misbehave.
mircea_popescu: the idea there was to drive compliance.
asciilifeform: right, this'd be immediately practical if the net consisted of wot'd folx an' their boxen
asciilifeform: ( $thread was specifically re puzzler of how -- and if -- to sort nodes ~not~ known to be operated by wot'd humans. rather than ' asciilifeform's wires ' and related experiment )
mircea_popescu: but yes, this is the other half -- need to bake trb nodes that won't get insta-banned by wot-trb on sight because they spew garbage
mircea_popescu: iirc we had a spec discussion re how to populate nodes.
asciilifeform: mircea_popescu: see thread...
a111: Logged on 2018-09-24 15:25 asciilifeform: mod6: my main hesitation re the item is that it may be 'plugging wrong end of funnel', quite possibly no such gymnastics would be necessary if we simply had a http://btcbase.org/log/2018-09-23#1853143 and always --addnode'd the actually useful nodes, rather than relying on pure chance
mircea_popescu: http://btcbase.org/log/2018-09-24#1853601 << this seems so to me. how about we wot the trb, rather than iptables it. ☝︎
asciilifeform: ( and i indeed suspect that no pentagon ww3 planner planned for anyffing like the current level of breakage, at that )
asciilifeform: mircea_popescu: aha.. the amazing thing isn't that 'x,y,z are broken' but that any part still worx..
a111: Logged on 2018-09-24 14:58 asciilifeform: ( this braindamaged design decision goes, as one might expect, back to the olden days when homo redditus was not yet on the net and 'arpa' designed for nukefest-related problems, not against 'ddos' )
mircea_popescu: http://btcbase.org/log/2018-09-24#1853590 << quite. what a nuke the "empowerment"'s turned out to be. ☝︎
BingoBoingo: It's like those nature documentary nile river crossing scenes with the crocs in miniature
BingoBoingo: All the time. Snapping turtles love ducklings
BingoBoingo: Imma suspect some sort of turtle
mircea_popescu: who knows what ancestral terror snake it thinks it is. because pixels, right ?
mircea_popescu: nearly fell over itself sideways trying to GET THE FUCK OUT OMFG OMFG
a111: Logged on 2016-01-26 17:20 ascii_butugychag: (there was a spiffy talk at shmoo, which mentioned how nn used in image recognition usually imprints on what - to a human - would be an entirely accidental cluster of pixels, and if you flip'em, it will recognize an obvious, e.g,. cat, as a refrigerator, etc)
mircea_popescu: http://btcbase.org/log/2016-01-26#1386298 << speaking of which : 124 gram duckling recognizes the cork of a bottle of guatemalan rum with a hair tie around it as DOOM ON FUCKING WHEELS. ☝︎
asciilifeform: 1 per burst ( see thrd )
mircea_popescu: http://p.bvulpes.com/pastes/1DmdC/?raw=true << fucking inconceivable 20kb udp made it through.
asciilifeform: yea as i understand diana_coman already has the knobs, can turn'em when she wakes up
asciilifeform: conceivably , on some really sad gsm modem somewhere, or the like, it will be 508.
mircea_popescu: i expect we'll have a nice pile of data this week.
asciilifeform: ( 508 payload, that is )
asciilifeform: nao, according to the talmud, max guara-nonfrag udp is 508. but this , this oughta be seen.
asciilifeform: ergo payload of 1472 corresponds to mass of 1500.
asciilifeform: btw , esp for diana_coman , e.g. 1000 byte payload will be a 1028 byte total mass. (header)
asciilifeform: ( i predict that any frag at all will give loss that cancels out the win from using heavy packets. but nobody should take my word for it, test, test )
mircea_popescu: well, this has been a productive discussion, hasn't it.
asciilifeform: well yes : conceivably 'frags in 2' worx ~100% of time, but 'in 8' not etc ☟︎
mircea_popescu: but the ideais, ARE we interested in a sliver of data re that ?
asciilifeform: mircea_popescu: aha yer likely to get same result with 2048 as with 20000
asciilifeform: ( ~inside~ some lands you can get 'jumbogram', up to coupla kB, but this is not really interesting from our pov imho )
mircea_popescu: is the take home from here that 1500, rather than 2048, should be the outer bound ?
mircea_popescu: (i really like tracepath much better tbh.)
BingoBoingo: <mircea_popescu> F=1500 here too. << Here as well
mircea_popescu: asciilifeform actually traceroute i had to install.
asciilifeform: mircea_popescu: if yer using a linux that 'came with things', it's likely to have both
mircea_popescu always used tracepath, nfi.
asciilifeform: at any rate, 1500 appears to be the ceiling.
asciilifeform: but prolly not worth the cycles