Hide Idle (>14 d.) Chans


← 2021-02-03 | 2021-02-05 →
adlai: http://logs.nosuchlabs.com/log/asciilifeform/2021-01-24#1030194 << "and here we test our powers of remembrance"
snsabot: Logged on 2021-01-24 23:52:59 verisimilitude: I won't be making any more changes, since my plan is to discard the code later, anyway.
adlai: deed-it-happen.lisp was published deliberately, despite being quite terrible.
adlai: ... and on top of it being ugly code, I had to ~sign~ the damn thing!
adlai notes, for the benefit of future readers, that http://verisimilitudes.net/base58 has expired by now.
snsabot: Logged on 2021-01-24 21:14:04 verisimilitude: I'll leave this up for a few days or so: http://verisimilitudes.net/base58
adlai: verisimilitude: please use a different URL, if you choose to publish an improved version of that code.
adlai has deliberately made a controversial request, and is quite certain to receive a "don't tell me how to run my own webserver" complaint in response, quite possibly from someone other than verisimilitude ; topic is open for discussion.
adlai: incidentally: after the earlier discussion I have deliberately published a broken link in one of my spews with the intent of rarely, if ever, having it serve eternally-static
snsabot: Logged on 2020-10-08 15:22:19 adlai: speaking of managers^H^Hment: one advantage of at least offering to pay someone else to deal with the aforementioned scrap metal is that it makes quick work of the heavy skill, leaving a few years for the highly unlikely possibility that I actually spill sweat instead of keystrokes.
adlai: content.
adlai: http://logs.nosuchlabs.com/log/asciilifeform/2020-10-08#1023049 << at a certain point, I actually do start to count bytes; and the answer to "why?", in this case, is not "why not?", but rather - "because it's faster than counting individual bits".
snsabot: Logged on 2020-10-08 15:47:47 shinohai: http://logs.nosuchlabs.com/log/asciilifeform/2020-10-08#1023010 << i also wonder y sha1sums plz
adlai wonders whether to publish his own implementations of the sha2 functions.
adlai: they are pretty much adlaifications of the ironclad originals.
adlai: i.e., slight overuse of macros, arguably excessive reliance on numeric coincidence, and no comments, yet.
adlai: incidentally: thank you, shinohai , for keeping archives of ben's old work.
adlai considers "reliance on shithub" equivalent to "relying on $lord's castle to exist, esto perpetuum", fwiw
adlai: castles rise and fall, ~especially~ when the lords in question have responsibilities beyond only maintaining a library.
adlai: and on exactly the same topic, I do need at some point to cogently explain my own interpretation of the two quoted concepts in this line, although it appears that none other than myself is interested in these particular deconstructions.
snsabot: Logged on 2021-01-06 21:43:18 verisimilitude: I don't buy into the ``god emperor'' or ``4D chess'' nonsense, know.
adlai: the reason I find such nonsense both interesting and worth discussing publicly is that ideas that one believes to be independently invented are often unoriginal, although rarely to the point of plagiarism.
adlai: one of the things I found quite distasteful about Mircea's behavior, and honestly also asciilifeform's, is the pervasive assumption that any instance other than the chronologically earliest, is plagiarised, rather than independently derived.
adlai: maybe I am simply too much of a platonist for this republic!
adlai: upstack, I'm quite certain that there actually is no "base58 document". there are a bunch of drafts, and implementations inspired by Satoshi's reference.
adlai does have a few complaints about the Base58Check format, although honestly, not even adlai cares about those at this point
adlai wonders how long it'll take Alert Reader to contradict that!
snsabot: Logged on 2021-02-04 07:51:16 adlai: http://logs.nosuchlabs.com/log/asciilifeform/2020-10-08#1023049 << at a certain point, I actually do start to count bytes; and the answer to "why?", in this case, is not "why not?", but rather - "because it's faster than counting individual bits".
adlai notes, amuzdingly, that one of the regular spiders has that broken link in its perpetual fetch queue, despite never visiting the distinct post that contains that link; most likely, whatever awareness it has of article structure has not clicked to the site's minimal metadata.
snsabot: Logged on 2021-02-04 07:50:02 adlai: incidentally: after the earlier discussion I have deliberately published a broken link in one of my spews with the intent of rarely, if ever, having it serve eternally-static
adlai: in other noose, at some point I'll have to give verisimilitude a piece of my mind about the Elision idea; although it should be, at earliest, after the next flu season, and perhaps once the idea has congealed sufficiently than I could anchor that link to one post instead of a log message.
snsabot: Logged on 2021-01-28 15:52:49 billymg: verisimilitude: you seem to allude to the value this adds in your first article with "The mathematician can truly create notation that suits the context. The programmer cannot." could you explain more behind the "why" of creating native word-based rather than character-based text storage?
adlai leaves only the grain of ricin, that this is right smack dab in the middle of the whole "record audio, store text" optimisation.
trinque: http://logs.nosuchlabs.com/log/asciilifeform/2021-02-04#1030811 << the amusing thing is that there's egotism on both ends, both "look mom I, am *also* smart" and "there were perhaps 100 intelligent men in history"
snsabot: Logged on 2021-02-04 08:15:26 adlai: maybe I am simply too much of a platonist for this republic!
trinque: *I am
trinque: both ends have a disease that prevents them from cohering into a viable human superorganism
trinque: discussing the categories presented here. I have more nuanced views of the men mentioned.
trinque: also, when one's cursing, it's both true and not the point that "you know, there isn't really any bull's shit to be found here"
adlai: isn't shit the source of all fertilizer? not exactly the worst of curses, in my book.
adlai: as for intelligence (new folk: read at least part of the linked convo; it's more than one line!):
snsabot: (trilema) 2014-10-12 asciilifeform: Adlai: now go and see reaction time vs. intelligence (take your pick of any known simple test for either of the two)
adlai: my litmus test for this, although not yet fully formed, is converging towards something inspired by Buckminster Fuller's quote that one invention can save the labor of a thousand workers; his precise wording was probably a bit naive, and perhaps mostly betrays the optimism prevalent during that phase of the industrial revolution.
adlai: I will however save the lazy new folk the trouble, and tl;dr Mircea's answer to that question as "watch their eyes".
adlai considers that answer quite an effective behaviorism, although naturally it splits several different ways once you actually quantify it
adlai: of course I'd not have remained so long past Mircea's -10 (and its various harbingers) had nuances not been present!
snsabot: Logged on 2021-01-18 10:20:35 trinque: if nubs is curious what kept me going, much of it was mp+alf megathreads, out of which all manner of interesting topics sprang.
adlai: speaking of WoT mathematics: I'm strongly tempted to declare that, if ratings are projected downto small-magnitude integers, then the (/ (1- domain) 2) must be composite; I have no opinion yet about the domain itself.
adlai: if the size of each side of the domain is prime, there is too strong a temptation to treat the numbers themselves as an unambiguous worth metric, instead of something that obviously encodes multiple kinds of information.
trinque: http://logs.nosuchlabs.com/log/asciilifeform/2020-10-31#1023907 << this notion has expanded since then. what can a rating be other than "pressed to choose, x comes before y"
snsabot: Logged on 2020-10-31 11:58:53 trinque: if so, what's suddenly the obvious solution is that there can be no such thing as two equal contributors in the eyes of the pressor.
trinque: *when pressed
trinque: slept like shit and dropping words like the ruskies!
trinque: adlai: I'd argue calling several people "3" has the pretense of information, and communicates very little.
shinohai: http://logs.nosuchlabs.com/log/asciilifeform/2021-02-04#1030804 <<< I feel a bit sad his www vanished before I got more works preserved. I'm trying to get better at archiving things to my library, if for nothing else but personal use.
snsabot: Logged on 2021-02-04 08:02:21 adlai: incidentally: thank you, shinohai , for keeping archives of ben's old work.
trinque: "your packets/patches come before his" communicates quite a lot.
trinque: flipped the other way, in what situation is undecidability useful in rating
adlai: however, you don't want all the information signed; e.g., nobody blinks if someone rates their wife +10, a program running on a webserver located within their physical control +8, and their firstborn +9, yet people start having unpleasancies once the spare only gets a +4
trinque: shinohai: it's not all in archive.is by now?
shinohai: trinque: I could not find the base58 thing on archive.is *at all* ... spent several hours searching.
trinque: adlai: wife=+10 is exactly a "contentless rating". of course he would, eh?
trinque: I would rather know whose opinions he depreferences in favor of his wife
trinque: and whether I have other routes to their preferences that might cause me to question my ordering of *him*
trinque: the whole of the critique is that 10 means nothing here.
adlai: http://logs.nosuchlabs.com/log/asciilifeform/2021-02-04#1030845 << integral domains [i.e., when your domain is not prime, and moreover, contains known factorisation] still encompass deterministic queries
snsabot: Logged on 2021-02-04 10:41:32 trinque: flipped the other way, in what situation is undecidability useful in rating
snsabot: Logged on 2021-02-04 10:34:22 adlai: speaking of WoT mathematics: I'm strongly tempted to declare that, if ratings are projected downto small-magnitude integers, then the (/ (1- domain) 2) must be composite; I have no opinion yet about the domain itself.
adlai: e.g., mod2 is whether the rated is human or program; mod3 is an indication of the rated's field of expertise; and mod6 is a register collision of overly concise nomenclature.
trinque: might have to dumb it down for me yet.
adlai: do we agree that, although there may exist multiple WoTs in a world at any given point in time, the tendency is always for there to be fewer distinct ones, in favor of one large graph that contains the union of all other ones as nodes, and varios levels of [dis]trust between them?
trinque: yep, more or less defines "world"
adlai: excellent
trinque: the "contains them as nodes" is unclear, a little fuzzier at the edges than that.
adlai: can I keep using the term "Integral Domain", or should we find a more understandable -- and possibly more correct -- concept?
adlai: it does have a mathematical definition, although it's quite a general one
trinque: I don't have much number theory in my gourd.
adlai: "ring" is also correct, mathematically, although it may be too general.
adlai: well, the way ratings work atm, they are small-magnitude integers, although almost no arithmetic is performed on the values themselves, beyond the comparisons.
adlai: however, in his original "WoT Mathematics" article, Mircea actually describes a few such computations.
shinohai: We can all go home, the solution to P! = NP has been found: https://archive.is/ekOFQ
adlai: lemme guess, the solution is "Logic, and the infinite wisdom of MP" ?
adlai may simply be well and truly stupid, but questions of algorithmic complexity class at the level of NP-"hardness" etc always did seem the most tedious and least interesting parts of mathematics
adlai: it is quite cool of archive.is to at least offer a snapshot of what apparently passes for an Abstract these days.
shinohai: I may be thick, but reading this paper I'm still unclear on where author is getting his "randomness" from.
adlai: unless the author's grasp of English is worse than expected, the author is not claiming to have proved P!=NP; rather, that parenthetical should simply be deleted from the title, because it doesn't belong there, and that kind of implication is not much good without proving the existence of a one-way function.
adlai wonders where to continue the number theory summary, if that's still of interest
adlai: "on your own blog" is not a valid answer, although "don't" is; and the word 'where' should have been 'from where', i.e., at what point do the understandings diverge
adlai: shinohai: my guess, from having read similar papers, and the bit in the archive.is snapshot, is that the paper says absolutely nothing about sources of entropy, and deals only with proving the behavior of the function that it introduces.
adlai does wonder why every single pdf document published after the death of macromedia flash seems to require its own vanity domain
shinohai: Doesn't that just make it "yet another function" and naught else?
adlai: well, the author claims that the function has those two information-theoretic properties, that not all functions have.
trinque was temporarily distracted by a pile of breakfast tacos
adlai: for example, if your symmetric encryption function is to take a key and repeat it with no change [abcabcabcabca...], then encrypting a text with certain properties [e.g., english-language letter frequencies; or, zipf-shaped word frequencies, if we're being elisive!] will bleed your key.
adlai admits that zipf-shaped is an overly handwavy dismissal; the idea being that natural-language text, unless compressed to the point of being effectively a random blob, still does contain probabilistic structure
adlai: trinque: my point about a composite domain is motivated from the fact that, while someone could claim that ratings should not be numerical, end of discussion, nobody can possibly claim that they should be numerical, and countably infinite. that is ludicrous.
adlai: for the wot to projected downto the point where any cryptographic tooling can strengthen it, there must be a limit to the size of the domain of the ratings, even if that is "4096 bits of text, a numerical rating who's absolute value is less than thirteen factorial"
trinque: ratings are only meaningful in the context of decisions: to whom to talk, etc
adlai blinks at the word 'whose'; quite a strange exception in an even stranger language. oh well.
trinque: "shit they collided, do whatever to distinguish"
trinque: what an english perspective I'm applying to wot-tronics!
adlai: the process of collision between WoT subgraphs, especially when they were arguably not distinct before the collision undeniably began, is why I am declaring the opinion that the dimensionality of a composite modulus is not an antifeature.
trinque disappears in a puff of pidgin
adlai: at least two examples are commonplace: a wot that began as purely 'professional' (e.g., folks who sign each other's code) and then encounters use outside of that same professional context; and the previously mentioned situation, where two mutually-excrusive worlds collide.
trinque: let's go back to the model you gave where all wots proceed towards the unified wot
adlai: as usual, I am looking far beyond what is relevant for any "implement this yesterday" level of impatience, and possibly even beyond the level of computational feasibility.
adlai: I'm still wondering whether "Integral Domain" is a good concept to use here.
adlai: it is the most specific mathematical concept, of the ones that describe the algebra in question, with which I am familiar.
trinque: one'd have to choose an algo for how to value the edge traversals from me to you, I think, before saying more.
adlai: "Integral Domain" tl;dr = you can add, subtract, and multiply, by all elements; you can divide by most elements; and you can factor the zero in multiple different ways.
trinque: right, "I can do arithmetic in this rule system" eh?
adlai: whereas in a field, any multiplication that resulted in a product zero, necessarily means that one of the multiplicants was also zero.
adlai: you can do arithmetic, although certain solution methods that are taught in grade school, become impossible, because you can't progress past a zero product in a deterministic way.
adlai: ehhh that's simply wrong. you can progress deterministically, there are simply more options.
adlai: so, one way to define an operation that takes two wots as input, and produces one wot as a result, is one where all the ratings within each wot are multiplied by the size of the other wot's rate space
trinque: are there any useful analogies where probability is expressed as edges on a graph, and used to make decisions using incomplete information?
trinque: finance comes to mind.
adlai: there's the natural similarity to integration of feynmann diagrams, although it's probably a worse-than-useless false friend.
trinque: maybe what's being calculated with wot traversal is the probability of defection in a prisoner's dilemma between nodes.
adlai: (in those, local probabilities are computed as results of a more complex graph's shape, where the more complex graph has values outside the real arithmetic)
trinque: rather, defection-or-collaboration
adlai: right. obviously it generalises to more complex games than the two-person version.
trinque: I would expect peer nodes with high ratings to begin collaborting more often, proceed longer, etc
trinque: yup, just groping here
trinque: maybe you defect against him, and now I defect against you, etc
adlai: I think your expectation is not always correct, because it is often too much bookkeeping to have formal financial interactions within a trusted group
adlai: the monetary technologies introduce overhead that can be avoided by delegations of responsibility that effectively prevent literal "mom and pop" businesses.
adlai: similar things happen in organized crime, where you probably won't see a registered corporation with all the middlemen of a smuggling operation named in one document, unless it's the one filed by a prosecutor.
adlai simply picked the one example of organized crime that springs easily to his own mind... there are always others
trinque: collaboration doesn't have to mean business, I think.
trinque: at the least it means I'll feed you truth more often than bullshit
adlai: agreed; so, one easy upper bound to the upper bound on the size of the wot's rate-space is the number of humans in the world, with various caveats - people are born, people die, and certain machines (at minimum, kyristor!) are important enough that they also participate in the wot
adlai: it's a ridiculously high upper bound for everyday use, although it allows for the computations involved during a collision.
adlai: e.g., for everyday use, -10 to +10; during a collision, a few bytes of source wot data fit within the padding that the signature algorithm will be applying anyway.
adlai: arguably, the programs performing these computations don't even need to make the full range visible in the default modes of operation; e.g., signing by default appends one zero to the rating
adlai: speaking of bullshit, and kinds of information: one other kind of information is scientific data.
adlai: e.g., if I publish a new encryption algorithm that is not immune to differential cryptanalysis, and two separate people collaborate to perform such an attack, they probably wouldn't want to have the same kind of signature on the measurements that they exchange with each other, as the kinds of signature they originally used when planning the attack.
adlai realizes that people in the scientific community get sad whenever anyone suggests that anyone ever fudges measurements!
adlai: at least, they really like to pretend that no scientist ever lies, and that there is no incentive to ever be dishonest. it's almost as though some scientists have never left the playground of mutually-aligned interests.
adlai: distrust is terribly inefficient, although as I see the world, the assumption of only one human superorganism, that benevolently integrates the work of all individual humans, is at best a premature optimization.
adlai has babbled quite a bit, and could use a few kilocalories; thank you, trinque and shinohai , for keeping the conversation going.
trinque: adlai: yep, moreover the swearings around which the superorganisms form may be mutually incompatible
trinque: which leads to a "war is inevitable and required to avoid stuck states"
trinque: will return to this thread more rested and less burned out, currently on a march
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-02-04#1030805 << speaking of which, adlai , why not mirror sumthing ? it aint hard to do, and immediately useful
snsabot: Logged on 2021-02-04 08:03:17 adlai: considers "reliance on shithub" equivalent to "relying on $lord's castle to exist, esto perpetuum", fwiw
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-02-04#1030810 << imho in the era when digging out prior art required 20yrs of sail around the world and plowing through libraries -- 'i reinvented' was sometimes plausible excuse. today -- just about never. pretty much w/out exception is result of willful blindness.
snsabot: Logged on 2021-02-04 08:14:37 adlai: one of the things I found quite distasteful about Mircea's behavior, and honestly also asciilifeform's, is the pervasive assumption that any instance other than the chronologically earliest, is plagiarised, rather than independently derived.
asciilifeform: ( re this blindness -- see also. )
snsabot: (trilema) 2014-03-05 asciilifeform: the mechanism in their heads that is needed to dis-invest in an unproductive line of inquiry breaks.
adlai: asciilifeform: fair enough; at least in my own experience, I often fail to locate prior art due to unfamiliarity with jargon.
adlai: e.g., my recent usage of a few mathematical terms that are almost certainly more general than necessary.
asciilifeform: adlai: this is why gotta talk to people.
adlai is always glad to hear, "cute prototype; have you heard of the Empire? they've been building these frobnifiers for over a decade already"
asciilifeform: ( and be ~very~ concrete re yer idea. i.e. this imho is example of how ~not~ to do it )
snsabot: Logged on 2021-02-04 10:35:30 adlai: if the size of each side of the domain is prime, there is too strong a temptation to treat the numbers themselves as an unambiguous worth metric, instead of something that obviously encodes multiple kinds of information.
adlai: right, it is impossible to find prior art for a patent that does not describe anything.
asciilifeform: example of informal but implementable description of an algo, e.g. here .
adlai: fwiw, of all the folks in this channel, you are probably the only one who might have anything substantially informed to say about the paper shinohai linked earlier
adlai has an entirely redditous understanding of formal cryptography.
adlai: i.e., I can read, and have even occasionally implemented things, although will actually laugh if someone seriously asks for me to audit a cryptographic system.
adlai: "me and what imaginary phd"
asciilifeform: adlai: it's a typical crackpot piece a la ye olde sci.crypt newsgroup. reduces to vernam otp + prng. lulzy claim of 'and so we proved p!=np' .
adlai does wonder how much formal mathematics learning verisimilitude has; the impression received so far is one of 'not much', although perhaps you also have at least the level of literacy referred to in the immediately preceding lines
snsabot: (trilema) 2015-07-05 asciilifeform: or that other thread, which i can't seem to find, where we spoke of how 'maths wot is not essential for being published, but for even having publishable ideas instead of burning your life on fibonacci sequences or some other solved/irrelevant dead end'
asciilifeform: adlai: it is painfully easy for an intelligent n00b to become convinced that he 'solved mega-problem'. esp. if he's enthusiastic about putting words/squiggles on paper but not so much re writing proggy
adlai: and before verisimilitude -- or anyone else -- get insulted by the words 'not much', consider that my standards for mathematics education are: 'kindergarden' = enough mathematics to work as a chartered accountant; 'elementary' = one mathematics degree, covering basics of modern algebra, i.e. slightly above my current level, although within plausible reach; 'working' = someone who can actually be
adlai: expected, within a reasonable amount of time (measured in years, and estimable in decades), to solve open problems.
asciilifeform: ( in sci.crypt was recurrent thread: 'i broke rsa!' 'ok, here's a pubkey, gimme the priv. p,q' '...' 'i'm waiting' 'you DON'T UNDERSTAND!! MY BREAKTHROUGH!' '...' )
adlai: lol, and "the ossifrage is one of the ~least~ squeamish animals"
adlai: my understanding of relevance of the various NP-overlap questions to cryptography is that the arrow of proof is unambiguously from various statements about inclusions, to weaknesses in kinds of algorithms
verisimilitude: Don't worry about insulting me, adlai.
verisimilitude: My mathematics schooling was pathetic, and I'm correcting this through books.
asciilifeform: adlai: if you didn't already know -- not a single one of the cryptoalgos in common use today, have any kinda useful complexity proofs (for avg.case, worst-case, any) associated with it.
verisimilitude: I overwhelmingly know about discrete mathematics, not continuous.
adlai: so it is almost impossible for a paper describing any specific new algorithm, regardless of the claims of strength that are proved about that specific one, to contain a result that breaks other algorithms; unless it also contains a novel mathematical result of much more general applicability, in which case, wouldn't that be more likely published simply as a general mathematical result?
verisimilitude: Implementing the SHA-2 family is simple, and I'll finish my comprehensive Ada at some point. Any fool can do it.
asciilifeform: adlai: in the given example, item purported to show that p!=np . did (surprise!111) no such thing.
adlai: it may be a fool's low-hanging fruit, although I found a bug in the Ironclad library when poaching their SHA2 implementation.
adlai: it was a really embarassing one, too: caught by the official test vectors.
verisimilitude: How amusing. I did find a flaw in mine, in the padding, but it was easy enough to correct.
verisimilitude: Unfortunately, my Common Lisp is much too slow, so an alternative design will be finished at some point.
adlai recommends both avoiding the urge to optimise CL code, and recognising that it is possible to write performant code in CL
verisimilitude: Still, I wrung eleven articles so far from the work.
verisimilitude: I'm not offended, but just as with English, I need no Common Lisp advice.
adlai: it's a good language for writing efficient algorithms, and you actually can make the compiled code efficient on a given machine; the main reason this becomes a waste of human time is that CL is a good language for generality and portability, and once you begin digging into the details of compiled code... maybe your experience is different, verisimilitude, but I actually find that the will to live is
adlai: reduced much faster when I begin 'fits-in-head'-ing instruction sets.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-02-04#1030846 << this is exactly the kinda problem that it only makes sense to attack in context of 'resistance of the medium', i.e. when you have actual practice where you actually end up needing to resolve the graph.
snsabot: Logged on 2021-02-04 10:41:44 adlai: however, you don't want all the information signed; e.g., nobody blinks if someone rates their wife +10, a program running on a webserver located within their physical control +8, and their firstborn +9, yet people start having unpleasancies once the spare only gets a +4
adlai supposes that there are few enough CL programmers of this level of fluency on the planet that "have you ever written CL for an employer" actually is a measurably doxxing question
asciilifeform: even in 'good old days' when 20+ folx were using vtrons daily, at no pt did anyone say 'i wish machine would resolve wot ordering for me mechanically, i'm tired of sorting by hand'
adlai does not imagine the aforementioned 'collision' to be an operation resolvable by one run of any hypothetical program
asciilifeform: adlai: i suspect most cl proggies put to use industrially in past 20y, were not even known about in detail by whoever paid. 'make the 4tonne robot move on spec' 'here's a box + softs' and in'ere is cl, or whatever, and only archaeologists could find out about it.
adlai: more along the lines of -- to return to a military metaphor -- the difference between 'preparing for war' (one wot, ratings mean whatever they mean to signers, etc) and 'war' (collision detected, invocations of various programs are executed using more complex algorithms to account for larger world, until collision is resolved)
asciilifeform only ever wrote 'commercial cl' in this form
asciilifeform: adlai: imho 'automatic resolutions' for wot are screamingly opposite to the Right Thing.
asciilifeform: (beyond simple 'friend or foe' case)
verisimilitude: A good bit of my machine code work is fitting-in-my-head an instruction set, adlai.
adlai: you can still place a plausible upper bound to the headcount of CL programmers, where by "CL programmer" I mean someone who, while perhaps not able to rattle off CLHS entries from memory, can at least find the relevant sections within a workable amount of time
asciilifeform: the Right Thing consists of 'simple cases -- mechanically resolve; errything else -- ring bell and summon operator'
adlai: as opposed to "can not program in common lisp without stackexchange/reddit/google/etc"
verisimilitude: I don't actually need the MMC to program, but its existence allows me to write better programs than others.
adlai: again, asciilifeform, you are strawmanning what I described earlier.
verisimilitude: One who can't program without an internet connection can't truly program.
asciilifeform: adlai: elaborate?
adlai: the kyristor (if I may use your word to describe my abomination!) that I imagine operates in an entirely transparent way.
asciilifeform: adlai: wainot write algo.
adlai: if I ask it, e.g., "compute trust distance between third messenger from colliding wot and captain of the guard from own wot", it does not subsequently do some GPG2-level idiocy wherein my entire local keyring gets rewritten to reflect the existence of messengers one through three
adlai: yep, this conversation is far beyond the cliff edge of talking on solid ground.
adlai is, again, glad to have aired out the ideas, so that he has a better idea of what exactly to write, once he's between parentheses.
adlai: before I proverbially step away again, there was one other 'wainot' that I wanted to answer
adlai: http://logs.nosuchlabs.com/log/asciilifeform/2021-02-04#1030932 << mostly "centipede's dilemma", i.e., there are too many collapsing castles on all sides
snsabot: Logged on 2021-02-04 12:52:50 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-02-04#1030805 << speaking of which, adlai , why not mirror sumthing ? it aint hard to do, and immediately useful
asciilifeform: adlai: i dungetit, there were what, a dozen folx with sites in all ? and of these, most still up
snsabot: Logged on 2021-01-14 15:15:07 verisimilitude: I feel like a man learning how to build houses, in a society where people expect them to collapse, and it's maddening.
asciilifeform: ( and of remainder, most can be seen here . grab own copy ? )
adlai: it was definitely a nice little rush of euphoria to think that "Castle TMSR" was monolithic and forever so; yet, #trilema is now a silent channel, and former lords are stepping away from public view; I guess so-called "Republican" society has to go through its dark ages, and that means building for earthquakes, instead of babylon.
asciilifeform: adlai: at the risk of pedantry -- ben quit coupla yrs before mp took ball an' went home
asciilifeform: folx who wanted whatever off his www, had 4+y !
adlai: at the risk of mockery: I'm not accusing him of being a collaborator in my "attack against the Republic"!
adlai: arguably things would have been resilient from much earlier on if there had not been a "Castle TMSR" in #trilema, although perhaps then there would not have been a critical mass of productive folks, either.
asciilifeform: adlai: i'm still puzzled re what is your obstacle. e.g. shinohai had 0 problem wget -m ... asciilifeform's gentoo mirror, and making own. what's keeping you from doing similar ?
adlai: lol, there is no obstacle; I'm here talking instead of doing anything productive.
asciilifeform: adlai: i suspect yer thinking of sumthing else though :
adlai: /mode #asciilifeform -v adlai
asciilifeform: well no, lemme try :
shinohai: My only obstacle before was general laziness and indifference, not gonna lie. Trying to be more responsible in savng things I find useful moving forward.
asciilifeform: mp was talented mountebank, a la barnum, and in #t maintained an addictive, to many folx, atmosphere of 'движуха' (untranslatable, but roughly 'errything in motion!!', 'happening!')
asciilifeform: without this drug, many folx suffered 'cold turkey' and went off into sunset. sorta how i picture, life ebbs away in a crackhouse if no moar crack is delivered.
adlai: lmao, and I was about to contradict with a "wdym, untranslatable; right here on page LMN, says...", yet indeed, only contains stem!
asciilifeform: asciilifeform does not offer visitors this, or any other dope. entirely willing to discuss e.g. mathematics, softs terraforming, other practical (or less-than-practical) subjs. but cannot offer 'happening!!!' cuz not barnum. doesn't have interest in attempting.
snsabot: Logged on 2021-02-04 13:46:08 asciilifeform: mp was talented mountebank, a la barnum, and in #t maintained an addictive, to many folx, atmosphere of 'движуха' (untranslatable, but roughly 'errything in motion!!', 'happening!')
shinohai: "If you don't believe me or don't get it, I don't have time to try to convince you, sorry"
asciilifeform: in asciilifeform's cosmography, движуха is ~harmful~. like cocaine. it is ~not~ ultimately a win, for folx to sit around 'tripping' while suffering the illusion of thinking.
asciilifeform: recall when mp wrote pompous piece re how 'tmsr defeated clitler!' -- typical.
asciilifeform: see also ( can't presently find ) old thrd with trinque's description of cocaine users, imho pertinent.
asciilifeform: y'know, folx w/ 'hurrah!' confidence spewing outta all orifices -- but ~0 capacity for reflection.
adlai does occasionally wonder about the whole "DAO Hacker" thing
adlai: e.g., in the BitMEX chatroom, one fellow actually said "MP was the DAO hacker" when someone asked "who dat"
adlai: at which point I continued realising just how little power cryptography has, because people so seldom actually bother verifying provenance.
asciilifeform: adlai: i'm almost surprised he did not take credit for death of kennedy.
adlai: in case anyone here is foggy on the details: the text of the trilema post about the DAO hack appeared shortly ~after~ the same text appeared on pastebin.com
asciilifeform: ( did for damn-near errything else that was even vaguely chronologically plausible.. )
adlai: perhaps a good use for bitcoin timestamping, next time someone feels like shenanigans.
adlai: my main question about that case is "why take credit"
verisimilitude: Until automated machines pretending to be people are too obvious, people won't worry.
verisimilitude: I'm not worried about mine unencrypted connection here; I just look at the logs sometimes to check.
adlai: MP's fanclub obviously was not dependent upon him personally finding such vulnerabilities; it was enough for him to simply be top blogger.
adlai: upstack, I think the first thing I should mirror is the archive of deeds, because that is something that I actually used quite a bit, myself, and expect to use in the future.
adlai: s/quite a bit/most/, i.e., that's the bit of infrastructure of which I made most use.
adlai: once I have a mirror of deeds, and perhaps an improvement upon 'deed-it-happen', I can provide a less fragile mirror of source code.
shinohai: "deed-it-happen" was interesting to me adlai .... had never before seen until I went base58 exploring.
adlai: nowai
snsabot: (trilema) 2015-12-01 shinohai: fukkin a adlai
adlai: your forgetfulness is understandable, considering what garbage quality the code is
asciilifeform: adlai: 'why take credit' was imho simple q -- same reason wai 'al qaida' etc took credit erry time a gas line exploded in some alabama shithole. 'hurrah, we make things happen!!'
shinohai: aha! I do remember that now, though was item I placed on desk and forgot, as I had not yet really dove into cl
adlai: speaking of deedbot, here's one that not yet in the FAQ: what is the glyph that appears in the links to mimisbrunnr?
adlai: well, "what is the glyph" is answerable by reading unicode specs; my question is better verbalised as, "what is the purpose behind that choice of glyph"
adlai doesn't have it rendered locally, so now answering the first phrasing himself
asciilifeform: meanwhile, in usg lulz, e.g. 'Bitcoiin used a name similar to the cryptocurrency Bitcoin and made false claims about its product to induce investments.'
adlai: (dull question has boring answer: it's a magnifying glass)
adlai: although, arguably, even the failed rendering is self-explanatory: empty white rectangle = block!
adlai is momentarily puzzled by a discrepancy between the ordering of deeds linked from deedbot.org and the order of their archival in 2015-12-01, although this is rapidly entering the "somebody else's problem" field
asciilifeform: adlai: what exactly were you trying to do ? ( mirror the deeds ? or replicate trinque's whole setup ? which ? )
adlai: understand what, i.e. exactly which parts of the website, I want to host. I do not want to replicate the entire thing.
adlai: in parallel to that, almost every time I link to old logs, I reskim the immediately preceding and consequent lines of that same day, occasionally even reading the entire day's log, both 'for the lulz' and because occasionally there is actually useful information there.
adlai: while doing that reskim, I noticed the discrepancy.
adlai: ideally, what I mirror initially will be my own deeds (i.e., the ones that I signed and submitted... and the one that kako submitted after I did so incorrectly!), with simply infrastructure in place for mirrorring additional ones on an on-demand basis
adlai: after that, I intend to look into "mirror entire linux distributions"
adlai: or whatever it was that you exactly were suggesting I do.
asciilifeform: adlai: start with what you, personally, would want to preserve. what wouldja miss if it fell down?
asciilifeform: i gotta admit, old mp deeds would not be top of my list. but to each his own.
adlai prefers to avoid mirrorring build scripts without actually verifying their behavior, even if they seem innocuous
adlai: right, I don't want to automatically mirror the entire deed history; however, there actually are a few "deed poems" that exist only there (and in whatever internet archives have picked them up)
adlai: along with the hashes of my mpex activity.
adlai: to be precise, hashes of the summaries.
asciilifeform: adlai: i can't resist to ask, what do these do for you nowadays ? i.e. that you want'em perma-archived publicly.
adlai does not anticipate ever getting asked for the preimages, nor the hashes, although he did actually go to the trouble of notarising those.
asciilifeform: do you sit down on rainy day and 'hm, i think i'ma spend evening rereading my old mpex trade logs' ??
asciilifeform: ever find yerself thinking, 'i wish i could link this here n00b to my old mpex log..' ?
adlai: those, together with the trades that were spammed to the #bitcoin-assets and #trilema logs, serve as a minimal example of scalpl's usage, with the strong benefit of already being public information.
asciilifeform: asciilifeform will undertake to give example of what he archives an' why. e.g. bvt's www because he had useful inline asm for ffa. qntra because was a++ chronicle of political events for 6+y.
adlai: the death of qntra is indeed a shame.
asciilifeform: ...log db cuz searchable log, whoever wants (and iirc at least thimbronion currently eats it!) can compute over 7+y of log, mirror, etc
asciilifeform does not archive material that he does not at least on occasion find self rereading, studying
adlai actually found the qntra compensation model quite an interesting financial experiment; possibly even more interesting than his own participation in S.MPOE, which occurred entirely after the actual OE was shuttered
adlai: archiving the log db is a good candidate for higher priority than archiving various source code repos
asciilifeform: adlai: nuffin stops you or anyone from cron/wget it whenever
asciilifeform has an hourly off-site cron for it
adlai uses log.nsa.com quite often, even if only for confirming the message id instead of guessing based on the ordering in his own client
adlai: I guess one bad although admittedly true reason that delays my yesterdayous hosting of a mirror is that I avoid using the 'common household materials' of cron,wget,nginx/apache,etc
adlai: the bane of the funemployed common lisper.
asciilifeform wonders whether adlai is also baking a nic outta 74xxx and driver for it in cl etc, or notyet
shinohai: Nothing stops ya from using drakma + hunchentoot instead, neh adlai? ^.^
adlai: nope and nope :-)
adlai: although I have only today encountered at least one low-hanging fruit for severe optimisation of hunchentoot's HTTP responses
adlai should, at some point, begin the long and torturous process of fetching and deploying the updates from various heathendoms
adlai: it is quite likely that others have already solved, e.g., terminating HEAD responses after headers instead of sending the entire page
verisimilitude: I've played with DRAKMA some; it didn't continue interrupted downloads by default, but I did have neat behaviour setup with the condition system.
adlai: arguably, one priority that delays deployment of additional infrastructure is studying the various crawlers (and occasional footprints) that visit the one site that I've left deployed.
adlai has, over a year ago by now, taken down the "CoinJoin Hunter" that did also serve as a mirror of bitcoin block headers
adlai: it only mirrorred headers, although required the entire chain for the coinjoin detections, and eventually that exceeded the phree hosting's diskspace.
adlai did not bother messing around with pruning, by that point!
asciilifeform: adlai: i somehow escaped even knowing wtf 'coinjoin' is/was.
adlai: coinjoin itself is a terribly general idea, that essentially boils down to the fact that a single bitcoin tx can be partially constructed, shared around between untrusting counterparties who each contribute their own inputs and outputs and incrementally update their signatures, and finally broadcast
adlai: my program (that iirc asciilifeform saw once, on github, and expressed horror at an admitted horror within... autogenerating an API client for prb's RPC!) detected specific kinds of coinjoin, of the shape produced by joinmarket
asciilifeform: aa this was 1 of those multisigism lulz, wasnnit.
adlai: nope
asciilifeform: adlai: no? how did they agree on the constructed tx ?
adlai: regular signatures for each input. the TXs are, upon naive inspection, indistinguishable from ones that are produced when a wallet uses several of its own coins as inputs.
adlai: specifically in joinmarket, or in the generalized coinjoin idea?
adlai: iirc, the original coinjoins were negotiated in bitcointalk forum messages!
adlai leaves a log link to the coinjoin hunter, in case anyone wants to run it, or worse yet, marvel at this horror of horror-generating horrors
asciilifeform: adlai: to the extent the protocol actually lets you glue together tx and retransmit w/out cryptomanipulations -- dun even need to agree, neh. can collect in wild.
asciilifeform: ( the 'malleability' nonsense was a variant of this idea )
adlai: iirc this is not possible for all signature kinds, because only certain kinds ignore other inputs
adlai: e.g., in joinmarket, each signature covered the entire tx; however, one topic that was occasionally raised, was implementing the kind of signature that only covers your own inputs and outputs, to increase the 'anonymity set' of each tx
adlai: so while fancy mathematical manipulations are unnecessary, there still is a need for coordination, which joinmarket solved by introducing the incentivisation
adlai suspects that one reason for this is your professed treatment of bitcoin as unspendable
snsabot: Logged on 2021-02-04 15:16:45 asciilifeform: adlai: i somehow escaped even knowing wtf 'coinjoin' is/was.
adlai: once you begin dealing with nontrivial coin flows -- and wish to avoid donating the various Brian Armstrongs your entire financial history -- then such technologies become quite interesting
asciilifeform: adlai: mno.
asciilifeform: adlai: 'unspendable' for reasons that have exactly 0 to do with 'tx histories'
adlai: I do not mean that you treat them as unspendable because of the tracability.
adlai: I mean that your lack of interest in the tracability meant you didn't have much reason to be interested in these hings.
asciilifeform: adlai: asciilifeform's gripes re 'unspendable' are 100% to do with ~usd~ , rather than btc
asciilifeform: if asciilifeform had a crate of au, or pu, it would be equally subjectively 'unspendable' and for exactly same reasons.
adlai: well at least you can build things with those! pretty much the only thing you can build with blockchain credits is notarization.
asciilifeform: adlai: lemme be clear, imho to date 100% of the движуха around 'anonymity enhancement' has been hucksterism (i.e. fuckwads helping gullible folx to lose their btc, and sowing confusion re how fundamental things work and where problems, where such exist, even are)
asciilifeform: 0 of it helpful against the actual, real-life headaches encountered by actual (vs. redditiferous 'any day nao!!') btc users.
adlai: ehhh, that's only about 1/3 of it; there is also a significant amount of USD, and probably even coin, spent on the tracking. there are at least three fiat-registered companies that sell tracking services to e.g. Coinbase.
asciilifeform: adlai: it's a quite lucrative (to folx properly usg-wotted) snake oil industry. which however afflicts even halfway clueful actual btc users ~0.
adlai: yep. one quite telling aspect of early bitcoin history is the fact that the first thing that any regulatory agency wrote about was FINCEN requesting that miners disclose the fact that they are mining.
asciilifeform: adlai: attempt to make chinese die of laughter or wat.
adlai does not recall the details; it may have been that they requested to a registry of pools, rather than miners themselves; either way, mining is the original unlinking mechanism
adlai: lol, it's a regulatory agency, they probably know that they can't touch anyone outside their jurisdiction.
adlai: maybe they are deluded and think they rule the world. idk.
asciilifeform: adlai: 'btc pinkertons' reduce, universally, to variants of this.
snsabot: (trilema) 2015-02-27 ascii_field: BingoBoingo: but almost certainly genuine. remember Quadro Tracker ?
asciilifeform: ^ req'd reading re 'multi-$bil usg police subcontractors' and their 'technologies'.
snsabot: (trilema) 2015-10-11 asciilifeform: folks familiar with usg 'scientocracy' will laugh, but the fact is, the mind-reading box does not necessarily have to... work. to be put in use.
adlai did not follow this specific thread at the time, although recalls long ago encountering the idea as a quote by iirc Peter Thiel, "Palantir is not a front for the CIA, the CIA is a front for Palantir."
adlai: may have been a quote by the other PayPal whatsisface!
asciilifeform: adlai: these folx manufacture, ultimately, only fud. and when you get into 'anonymization' crackpotter, yer in fact falling for this fud.
asciilifeform: 100% of past an' current kangaroo court docket, was enabled by a) old-fashioned stoolies b) nsa 0days . and 0 by 'btc trackers'.
adlai: y'know, if you'd ever actually listen to thing I've said, rather than merely read things I've written, you'd realise that I prefaced my talk about joinmarket with "there is no anonymity, all this program does is let you pay to scatter dust on your own footprints, one at a time", in slightly different words.
adlai was once approached by girls who wanted a graphology reading, because he was reading a book about this; they were quite disappointingly disappointed when the reading began with, "all this process does is reveals biases, yours and mine; I am using techniques known as cold reading..."
adlai: apparently some people want to be fooled!
verisimilitude: Most people tend to want fooling in some ways, e.g. happiness.
adlai wonders in what way to attribute verisimilitude if/when copying the robots.txt file
adlai: my inclination is to have a minimal one, for reduction of what I view as wasteful traffic (currently it just returns the 404), although I am considering building a detailed one similar to yours.
verisimilitude: I found a simple robots.txt boring, yes.
adlai supposes that a sufficient attribution is a comment line in the robots.txt file itself
adlai: huh. for some reason I recall that when I first read about these things, about a decade ago, there was a way to request rate limiting from the crawlers; this appears to be entirely hallucinated of me. oh well.
verisimilitude: That's good attribution, yes.
trinque: adlai: would the deed pycoad help?
trinque: note that prior to a certain point, those deeds were made by some other piece of shit which I replaced
trinque: with you know, my own piece of shit.
verisimilitude: I've made a minor adjustment to my robots.txt now, adlai. For some reason, DuckDuckGo doesn't share its bot user-agent value, and I recently learned my website was still being ignored.
verisimilitude: So now I've another guess at it.
adlai: thank you for the offer, trinque , although it is not necessary right now, as I only plan to mirror the deeds (and the FAQ, although I might write my own version instead of copying yours verbatim), rather than duplicating deedbot itself
← 2021-02-03 | 2021-02-05 →