Ossa Sepia

December 24, 2018

A Week in TMSR: 10 - 16 December 2018

Filed under: A Week, TMSR — Diana Coman @ 3:55 p.m.

On Monday, 10th of December 2018:

Mircea Popescu notices that the MP-WP1 installation on Pizarro's shared server seems broken as it fails to correctly process footnotes. In response, Asciilifeform2 asks BingoBoingo3 to look into it, noting also that other accounts on the server (Hanbot's) don't exhibit the same problem and therefore the issue has to be linked to Nicoleci's account on the shared server.

Mircea Popescu expresses his surprise at Nicoleci's apparent inability to express herself in writing anyway nearly as well as she is able to express herself orally. The difference is significant enough to be rather hard to believe if not directly witnessed. Further discussion with Trinque and Asciilifeform of Nicoleci's public writings on her blog - mainly summaries of TMSR logs - and of the sad state of what passes as "writing" in the US nowadays leads to the conclusion that the core issue with her writing is that it lacks any narrative structure: instead of telling a story of any kind, she seems to attempt to just give the gist of her thoughts at one moment or another.

Asciilifeform expresses his pleasant surprise at having recently tried a 3D device. He suggests it for Eulora but Mircea Popescu notes that Eulora is significantly more intellectual than visceral or graphical at the moment and the current struggle in this direction is anyway simply getting even basic art done for the game rather than improving public's access to it.

Danielpbarron publishes on his blog another snippet of his conversations with some dudes on religious matters. Trinque struggles to make any sense of the published snippet and points to Danielpbarron the solipsistic nature of his current activities as they can be perceived based on his publications. Danielpbarron fails to see Trinque's point and enquires whether there is anyway any significant difference between talking publicly of religion as he does and talking publicly of sex as Mircea Popescu does. This enquiry is promptly answered by Mircea Popescu who points out some significant differences: while every human being is interested in sex seeing how sex is fundamental to humans, not every human being is actually interested in religion seeing how religion is fundamentally gossip; moreover, while other types of gossip are at least interesting as they touch on interesting people, religion fails to captivate as it concerns nobodies. Danielpbarron disagrees with this view of religion and affirms that the "truth of the Bible is universally known", offering as unique support to this assertion a few citations from his Bible.

Nicoleci publishes her 101th post on her blog detailing some interactions with people from her past who failed to impress her as much as they told themselves they did even when she was younger while positively making her laugh currently with their unsolicited emails.

Ben Vulpes publishes an accounting statement for Pizarro for November, relying on a semi-automated process (numbers are produced automatically but the final format requires manual work to put everything together).

Diana Coman realises that her previously mentioned problem of an empty genesis .vpatch as a result of Cuntoo's bootstrap script is caused by an issue with the vdiff tool on the machine running the script (so nothing to do with Cuntoo's bootstrap script after all). After fixing the vdiff tool she reports that the Cuntoo script runs successfully and produces a .vpatch but the signature for it fails to verify. Bvt chimes in to report that he has a similar problem on his computer as the .vpatch he obtained from Cuntoo's bootstrap script fails to verify against Trinque's provided signature. Later during the day, Diana Coman publishes the .vpatch she obtained and Trinque is able to compare it with his own noting that there are several differences that he will need to fix, including his use of sha-based vdiff rather than the keccak-based vdiff. Diana Coman also notes that the Cuntoo bootstrap script fails on a different machine configuration (different operating system mainly), stopping with an error. She provides a paste of the error and Trinque is able at a later time to point her to the potential issue - an un-met requirement (having /dev/shm mounted) for compiling Python.

Diana Coman gives a talk on Bitcoin to students at Reading University in the UK. Later during the day she publishes a write-up of it including a detailed account of her Bitcoin talk and the supporting slides that she used.

Diana Coman offers to Asciilifeform the results of a tcpdump running on SMG's test server with Pizarro for several months during the year. The dump provides the content of some unexpected UDP packages that were observed during a previous test of UDP communications in October 2018. The dump includes some VoIP apparent scam that seems to originate from Iceland. As Asciilifeform is interested to investigate more into this, Diana Coman points out to him that it's all on Pizarro-owned infrastructure and so he asks BingoBoingo to reroute to one of his own computers with Pizarro all packets with unassigned IP destination.

On Tuesday, 11th of December 2018:

Commenting on Diana Coman's write-up of her talk at Reading Uni on the previous day, Mircea Popescu notes that the lack of a recording of the talk is rather unfortunate especially given how simple it is to obtain normally. Diana Coman and Mircea Popescu then discuss a bit the practical aspects of recording a talk and the rather shockingly basic conditions offered by Reading University on this occasion. Mircea Popescu notes in conclusion that the write-up of the talk looks good and the missing recording is more a matter of "missing out on a possible fanbase!" than anything else.

BingoBoingo reports that his Peruvian girlfriend finds Argentina very beautiful especially compared to what she knows of Uruguay. This prompts Mircea Popescu's "eh" and Asciilifeform's observation that Argentina hasn't quite managed yet to fully burry/destroy/run down the beautiful buildings it inherited from back when it mattered. The conversation then moves onto the significant differences in quality of buildings in different parts of the world and at different times, with Asciilifeform revealing that he can actually distinguish what he considers well-built structures by their smell that might be - or might not be - due to a combination of aging plaster, actual wood and perhaps old books in significant quantities.

Nicoleci publishes on her blog her summary of TMSR logs of 19th November 2018.

BingoBoingo publishes on Qntra an update on Macron's adventures in France and another update on Ebola's adventures in Congo.

Mircea Popescu draws on his extensive knowledge of world history and his extremely numerous interactions with a wide range of people to discuss his emergent view that multiculturalism fails first and foremost for lack of multiple actual cultures rather than for lack of potential merit in the idea of culture-mixing itself. Asciilifeform points out to the merits of China (at the time of Confucius) as an example of different actual culture that existed but Mircea Popescu notes that merits are irrelevant for the issue at hand: in practice, there is only a very narrow and unique way to culture and so everything that counts as such inevitably finds itself on this same path without much diversity possible. China is given again as an example since its current relevancy in the world is, in Mircea Popescu's view, fully due to and limited by the extent to which it copied white man culture. Addressing Asciilifeform's point, Mircea Popescu also notes that previously to this copying, China was simply a large bureaucratic state in a similar way in which the Inca state had also been one but still failing to actually develop as a culture since working organisation by itself is not enough. To support his point, Mircea Popescu remarks also that an actual alternative culture in China would be directly identifiable simply by its results. Given the obvious lack of such results - as there is no equivalent Chinese #trilema at all, let alone one bigger in size as it should logically be given China's size and more efficient organisation - it means therefore that there can't possibly be a culture there in any sense either. Both Mircea Popescu and Asciilifeform acknowledge that this might still be proven incorrect at a later time although the chances for such proof seem to them rather low. The more likely explanation for the current situation is in Mircea Popescu's opinion the simple fact that China can't seem to be able to advance past its remarkable efficiency at copying - currently copy successes that stop short of developing anything new including for instance mining Bitcoin but also owning the full fab stack and still failing at the same time to produce its own CPU architecture.

Asciilifeform rages at html's failure to provide a reliable way to format even basic equations so that they look the same across different displays and browsers (in particular without using javascript and/or images). Trinque suggests using SVG might be a good approach for the task but Asciilifeform rejects it because it won't be of any use for text-based browsers. Mircea Popescu provides a solution based on the use of html tables and top/bottom floating alignments, publishing it on Trilema as well, for future reference. At first, Asciilifeform balks at the proposed approach as he says it doesn't work with the Lynx text-based browser but Mircea Popescu points out that there is no way that works exactly the same in both text-based and graphical mode.

BingoBoingo announces that Pizarro's price for BTC is set at $4000 per 1 BTC for the month of December 2018. This price is based on an auction of $2000 that concluded on the 7th of December with the sale of the $2000 to Mats for 499.99ECu. Using this exchange rate, BingoBoingo produces Pizarro's invoices for provided services to bvt, jurov and trinque. Further invoices are likely due for Mocky's and Nicoleci's shared hosting with Pizarro and for SMG's test server.

Mircea Popescu states that he considered for some time Diana Coman's innovation/subversion distinction and he finds it to be well founded. He further notes that this distinction makes it clear that there is very little difference between subversion and "inclusion." Diana Coman agrees with this observation and notes that those finding change (hence, innovation by another name) difficult will simply push for subversion instead for as long and in as many ways as they can. Mircea Popescu adds to this the funny fact that Spanish uses the same word for expressing that something is expensive ("cuesta mucho") and that one finds something difficult ("me cuesta"), driving home the inescapable conclusion that indeed, the sort of person who finds it difficult to think (and therefore to change) has indeed no busines in #trilema or with Bitcoin for that matter. Diana Coman further links this "cost" of personal difficulty to the oft-heard complaint of "it's not fair" but Mircea Popescu considers the matter to be a much more intricate ball of nonsense than that. Nevertheless, he notes that a preocuppation with "fairness" (as opposed to correctness) is indeed a good heuristic for lack of useful intellect since it betrays significant inner voids that make it all together doubtful the subject is really a person at all.

A side note by Mircea Popescu on the provenience of the "arena" word in English from the Spanish word for sand turns into a short discussion with Diana Coman on the Arenal volcano in Costa Rica and subsequently with Asciilifeform on the properties of volcanic sand and the importance of semiconductors.

Asciilifeform announces that he will bid on a Symbolics MacIvory model and he will have it xray tomographied if he obtains it.

Mircea Popescu rages at Mozilla Firefox's idea of "releases" of the browser that include executables of all sorts and assorted signatures without any clear apparent meaning. Asciilifeform is rather amused at the idea that there is anything other than ceremonial in latest Mozilla offerings but notes also that he is not aware of any version of Firefox that did not suck to start with. Trinque chimes in to say that he has a version of Firefox that he built on Musl so that there is at least that as a potential de-facto graphical browser for Cuntoo. Mircea Popescu notes that at some point the republic will likely have to write its own sane browser anyway, getting rid in the process of all sorts of useless junk that currently come stuck with any graphical browser.

On Wednesday, 12th of December 2018:

Nicoleci publishes on her blog her summary of TMSR logs of 20 November 2018. She also notes that fetlife has deleted a post of Mircea Popescu from 2 days before since it was apparently more liberal than their liberalism can take.

Phf brings back the discussion on fairness/correctness from the previous day noting that he naturally considered fairness to mean exactly that: a recognition of correctness even when it's not to one's own advantage. In response, Mircea Popescu points out that this meaning of fairness as unpleasant-but-correct has always been a purely eastern one while the western definition always focused on a sort of weighing and comparing of outcomes. He links this to Hajnal's line in the sense that fewer and later marriages give more idle time to be spent on the contemplated sort of "fairness" considerations.

Mircea Popescu redirects Nicoleci away from attempting to summarize TMSR logs and on to transcribing old proceedings of the Royal Society of London that are rather interesting to read but are hardly readable in their existing format since they've been mangled by the automated OCR process.

One of Trilema's readers suggests to use a Wordpress Latex plugin to properly format equations. Mircea Popescu passes on the suggestion but Asciilifeform says he already investigated the plugin and it fails to solve his problem as it still relies on images and therefore it produces output that is not entirely suited for text-only browsers. Mircea Popescu points out that Mathematical notations are simply not fully alphabetic and as such they can't ever be pure text and therefore it's up to terminals to work correctly by being able to handle text + adnotations rather than text only. The discussion further advances on to what sort of text preprocessing should be actually done by a browser with Mircea Popescu noting that this question doesn't yet have a clear answer and Asciilifeform noting that at any rate, existing answers such as tags totally fail to actually answer anything. The mention of tags touches a nerve with Mircea Popescu and he notes that they are a very good example of the fundamentally broken approach that created the significant current technological debt: "simplification" implemented without regards to actual secondary costs incurred and by removing the barriers to entry that kept out precisely the sort of people that had no place to enter in the first place. While Asciilifeform heartily agrees with this view, he considers it old news and summarises it as "mechanization + idiocy == mechanized idiocy." He adds however that this sort of simplification "works" anyway simply by subversion of the very object called computer since actual computers are even more difficult to obtain than they were before while the objects that are now indeed very easily obtained are computers in name only.

BingoBoingo publishes on Qntra four articles: on Britain's no confidence vote in its Prime Minister, on the death of a Physics professor at Stanford University, on the secret conviction of George Pell in Australia and on one of the FBI's terrorism charges.

Mircea Popescu discusses with Asciilifeform the dilemma of Free Software raised by Naggum: while useful code has indeed value as Naggum clearly argues, its valuation cannot be approached in the way that Naggum seems to suggest, namely by attaching some value to the lines of code itself and/or closing code so that its source is not freely available anymore. In Mircea Popescu's view, free access to useful code source does not take away value of the code but instead adds a very useful entry point that works also as a passive but effective filter so that valuable contributors can be discerned from time wasters. Asciilifeform does not disagree with this view but expresses some reserve: he says he did not learn from reading ugly code despite reading loads of it but rather from reading non-code text; he also notes that Naggum seems to have been aware of the fact that lines of code added do not translate into value added but rather the opposite (the best code is no code); he adds also that Naggum's statement regarding the loss of value of software through free publishing are likely the result of his own personal history of trying to make a living by solving complex problems and seeing the tools he needed gradually vanishing as their producers failed to be valued enough to be able to continue their work. Mircea Popescu acknowledges that this is very possibly true and even proposes the neat packing of this pattern into a foundational myth under the title of "avik killed naggum"4 but notes that nevertheless the view that publication destroys value is not only misplaced but dubious in that it actively serves only those authors that attempt to extract more than their work is worth on closer examination. And since the abstract work of computer programming is much more similar to other abstract work such as that performed by doctors, the correct valuation should also follow similar patterns rather than attempting to follow patterns (such as copyright) that are derived from valuation of non-abstract work. As a result, Mircea Popescu notes that on one hand the requirement to publish code does not have to apply without discrimination and on the other hand the only correct way to pay for abstract computer work is through the crown allocation process: authors of abstract work may receive their payment as a recognition by a higher authority (the crown) of their valuable contribution but not as some quantifiable, formula-calculated amount that most users can decide on since most users are utterly unqualified to evaluate this type of abstract work in the first place.

As a continuation of the previous discussion on evaluating abstract work in general and code in particular, Mircea Popescu further stresses the important fact that valuable abstract work is by its very nature and fundamentally a surplus phenomena - meaning that there has to be first some surplus in order for one to be capable of performing abstract work of any value. In practical terms, this means that the authors do it without strictly needing the payment for it and as their own personal choice of doing it in preference to doing other things - some of them with clear payment even - that they are perfectly able to do. Asciilifeform also links this to operating from causes rather than for purposes (i.e. for obtaining some specific payment in this case).

A further continuation of the same discussion explores also to some extent the further difficulty in assigning rewards for valuable abstract work even through the crown allocation process. The process does not make the evaluation of abstract work any easier and it also doesn't provide a clear way to ensure optimal labour allocation at times of need. Essentially, Mircea Popescu notes that existing tools (money as a signal of value and market forces as regulators) although a good fit for concrete work and objects are nevertheless a disastrous fit where abstract work is involved and their failure is so significant that it likely drives intelligent people towards some form of socialism (as the only sort of alternative perceived) in their attempt to find a solution to the problems caused. The conclusion overall is, in Mircea Popescu's own words: "labour allocation is broken and nobody has any better".

BingoBoingo issues Pizarro invoices to Mircea Popescu for Nicoleci's shared hosting and for SMG's test server. He also updates Pizarro's public page to reflect the 10% discount offered on shared hosting for annual subscriptions over monthly subscriptions. Later, following Mocky's request, he also invoices Mocky for an annual shared hosting subscription.

Mocky asks BingoBoingo to bill him for his shared hosting with Pizarro on an annual bassis. He reports that his search for a job is still ongoing although slowed to some extent by holidays of his interviewers. Mircea Popescu suggests perhaps pooling resources through running a TMSR version of bourbaki: specialist appliers to remote jobs dumping tasks in a file that gets passed around for TMSR people to choose from as and when they want to do some non-TMSR work. Mocky chimes in to say that he previously considered outsourcing some of his own work, while Trinque notes that Oracle for instance is known to actually do precisely this. Asciilifeform says he'd be delighted to work in this way but expresses his doubts at the scheme, mainly due to the difficulty he perceives with task level/definition/discussion and the potentially problematic case of tasks that nobody wants to pick up within the allocated timeframe. Mocky says that his concern with this model is the fact that it can take him a year to become capable of actually solving specific problems within a reasonable time.

Asciilifeform and Mircea Popescu discuss the actual relationship between employer and employees with specific focus on Asciilifeform's apparent inability of escaping employee status. Mircea Popescu notes that the core issue seems to be the mismatch between the favourite "select first and then talk to selected" approach of most republicans and the opposite "filter the ocean" approach5 that is actually required for any search outside of TMSR (a search for employer included). Relatedly, he asks whether Pizarro has managed to do anything of the sort in order to find the clients it needs for survival. At a later time, BingoBoingo replies, revealing that the short answer is no, Pizarro has not yet managed to do anything of the sort but it might perhaps still manage to do it if only an "awk ninja" materializes to write the needed scripts.

On Thursday, 13th of December 2018:

Nicoleci publishes on her blog her summary of TMSR logs for 21 November 2018.

Diana Coman negrates douchebag for obstinately wasting her time. Probably in retaliation, douchebag carves in his own nick on freenode the message that nobody is interested in listening to, doing the irc equivalent of sandwich man. This prompts some laughter and merriment all around and Phf notes that douchebag's vulnerability finding is all about form rather than substance. Diana Coman says that the proposed view fits indeed the observed behaviour and moreover makes the whole activity very similar to a form of political correctness applied to code. Mircea Popescu takes this further and says that in this case the whole thing is also the precise equivalent of period politruks that attempt to police the code as the currently relevant form of speech.

Diana Coman publishes Chapter 12 of SMG Comms containing a thread-safe Ada implementation of simple queues that are specific to the needs of Eulora (as opposed to the generic thread-safe Ada queues of the standard).

Diana Coman informs Trinque that she experienced some problems obtaining an answer from deedbot to the !!ledger command. Trinque notes that the command currently works for him and he suspects the issue was most likely due to a lost connection between the irc bot that receives the command and the back service that actually handles all wallet functionality.

Diana Coman rates juliankunkel, the lecturer at Reading University that invited her to give a Bitcoin talk to students. Asciilifeform and BingoBoingo welcome him but he doesn't have much to say.

Asciilifeform reports he acquired the bolix on which he previously bid and he says he will therefore xray it. A bit later, Mircea Popescu contributes to this with a warning for Asciilifeform to check the rated power of the equipment he intends to use since not all equipment is powerful enough for such a task. Asciilifeform however reveals that the task is not likely to require high power anyway as there is no middle metal layer.

BingoBoingo publishes on Qntra: on the use of facial recognition at a pop concert.

Mircea Popescu, Asciilifeform and Diana Coman discuss the best approach to take for implementing (or not!) a sender/receiver layer as part of SMG Comms. The conclusion is that there will be such a layer as part of SMG Comms but a very thin one that simply moves UDP messages from/to outbound/inbound queue and the UDP socket. The reason for this layer to exist is the need to move UDP messages quickly from the relatively small queue on the IP stack to the larger in-memory queue. The reason for it to be part of SMG Comms is that it's not specific in any way to any given application since it's so thin as to focus exclusively on moving messages from/to socket and queues.

On Friday, 14th of December 2018:

Asciilifeform provides a paste of his talk to adlai in #asciilifeform as proof that quitting drinking has at least *some* effects. Mircea Popescu suggests a recuperative scholarly series on the SNS server and later notes that this is simply for documentation value rather than some silver bullet (or indeed any sort of working bullet at all). Asciilifeform indulges his love of old/interesting hardware mentioning items rarer than the bolix: xerox lispm and tandem. Phf reveals he has worked on a tandem (known as HP NonStop) at some point and he appreciated the architecture but noted that the software was entirely written in Cobol. This is interesting to Asciilifeform but it makes him poke Phf about some promised Bolix documentation that he previously said he might have. Following on from this, Asciilifeform reveals that he will likely perform the xray of his Bolix machine with his own hands but that he'd like to have Phf's papers (if there are any) to check against. The next day, Mircea Popescu adds to this discussion noting that there is proper xray-hygiene to follow when performing such a task.

Danielpbarron publishes on his blog another talk with unknown people on religious matters. A reference to reddit in there prompts Trinque to enquire if Danielpbarron is meanwhile militantly anti-republican. Danielpbarron flatly answers "no".

Nicoleci publishes on her blog her summary of TMSR logs of 22 November 2018.

BingoBoingo publishes on Qntra: on Germany's three choices of sex on paper.

On Saturday, 15th of December 2018:

Danielpbarron publishes on his blog another talk with random people on religious matters. Asciilifeform gets from it the impression that Danielbarron's approach is essentially calvinistic but Danielpbarron rejects this assessment on the grounds that "calvinism leads to hell". Some further talk reveals that Asciilifeform hasn't followed the religious life of Danielpbarron all that closely.

Asciilifeform further discusses with Phf his current plan for xray-ing his Bolix machine and then using the gained knowledge to build probes for further knowledge gain.

Asciilifeform notes that Ben Vulpes' logging bot is not working and Mircea Popescu notes that anyone can start another logging bot and simply aim it at the chans of interest as the bot code is published already.

BingoBoingo publishes on Qntra: on the latest adventures of Macron in France.

Mircea Popescu provides another sample from the responses he gets to one of his ocean-filtering actions. Asciilifeform is curious on the percentage of responses that manage to at least read the full initial message that usually gets cut off on various mobile phones and the like. The response is that there are some that pass this basic test but the percentage is very small.

Nicoleci publishes on her blog her summary of TMSR logs of 23 November 2018.

On Sunday, 16th of December 2018:

Asciilifeform announces he received his Bolix in perfect packing, with all accessories and able to run. He notes that he was in the end the only one to bid on this machine but he still did not want to miss the opportunity to buy it since the price apparently increases by $1000 every year. Phf suspects that there aren't that many bidders anyway and it's all more of a show with other owners of older hardware simply hoarding it as they notice they can't replace it anymore at any cost. The emerging picture seems to be that 2009 is the cutoff point for hardware that one can trust. A bit later, Phf provides Asciilifeform with DKS patches for Bolix and a port of 'zork' to Bolix.

Mircea Popescu laughs heartily at the Unity Web Player being now reportedly no longer working on Chrome, Firefox and Edge browsers. His faint interest in the matter focuses on the fact that it's totally unclear why and how did Unity exactly achieve its previous popularity. His hypothesis on this is that Unity got "chosen" simply for lack of any alternative. Asciilifeform offers as similar puzzle the success of qt but Mircea Popescu notes that they are in fact not comparable since Unity never actually worked nor did it ever have serious resources to speak of while qt both works and is not in fact going anywhere. Asciilifeform then links this to Bolix noting that the 3d engine for it still exists and is called Mirai. Its forum however is not working as it was overrun by spam. Amberglint joins in the discussion to correct Asciilifeform's assertion that Mirai was ported to CPP - he says it was in fact ported to Allegro Common Lisp. He also mentions that the most well-known work done in Mirai is the Gollum character for the Lord of the Rings film.

Asciilifeform publishes on his blog Chapter 14A of the FFA series covering the first half of Barretts Modular Reduction.

Mircea Popescu publishes on his blog the result of his wtf is a "Post Malone".

Diana Coman publishes on her blog a summary of TMSR logs from 3 to 9 December 2018. Mircea Popescu provides feedback and some corrections to it. The summary also prompts Mircea Popescu to add to one of the main topics previously touched namely the existing conflict between different versions of database management systems (mysql and postgres) being needed for different sort of tasks. This spills into the next day and the next week.


  1. A customized version of the Wordpress blogging platform produced by Mircea Popescu, packaged in V format by Hanbot and currently used by most republican blogs and by the Pizarro ISP for its clients that share space on a server. 

  2. Main tech for Pizzaro ISP. 

  3. Founder of Pizarro ISP and the only current local pair of hands. 

  4. Avik is the self-styled "master" that reportedly feeds his slaves a cocktail of pills, works at precisely the sort of software that crowded out of the market the tools needed by Naggum for his work and otherwise keeps pestering Nicoleci with unsolicited emails. 

  5. This is a term of art meaning literally talking to EVERYONE and then filtering out 1 / 1mn or in similar ratios those contacts that are in fact of any interest. 

4 Comments »

  1. > to verify against Trinque's provided signature.

    The utter rapturous ~beauty~ of this signature business. Looky how it forces "the unthinkable", hygiene in the fabled stables!

    Ahhh, to only have lived now, in times like these, to do the deeds they did, to drink with morning cup the fresh sweet revolutionary air...

    > and so he asks BingoBoingo to reroute to one of his own computers with Pizarro all packets with unassigned IP destination.

    On re-read, how very CIA of him, huh.

    > Mircea Popescu rages at Mozilla Firefox's idea of "releases"

    I guess rages, but also provides a proper list (something that can not be found browsing mozilla.net).

    > that fetlife has deleted a post of Mircea Popescu from 2 days before

    Item's actually in the logs ; "deleting" works as well as it used to.

    > it actively serves only those authors that attempt to extract more than their work is worth on closer examination.

    The problem's deeper : what's being suggested there is that there's two types of author : logarithmic and exponential, if you will. L-type authors always produce less than a linear improvement on their inputs (and indeed the quoted "SF" schmuck stands exemplary to this) ; E-type authors always produce more than a linear improvement on their inputs.

    That an E-abundance tends to correlate with C-poverty (ie there was a Shakespeare in the 1500s England-a-Scotland-of-France days but not 1800s Rule-Britannia-Britannia-rule-the-waves days, or a Dante in 1300 not 1900 "Italy", which is to say long before rather than long after the old texts were recuperated) may or may not be coincidence.

    In any case, the proposition is that L-authors are the inventors and the proponents of "copyright", for reasons directly derived from their "competition" (as they understand it) with E-authors : they expect in the dark nobody can distinguish an L from an E.

    > "labour allocation is broken and nobody has any better".

    Consider how this also links not merely to the Beria discussion, but also to the China problem earlier : well organized empire, makes plastic objects. Why ?

    > His hypothesis on this is that Unity got "chosen" simply for lack of any alternative.

    No, actually, this was alf's idea. My impression is that it was chosen in exactly the way Mark Cuban was chosen to "be rich" (or, if you prefer, Russian oligarchs" -- these are very much not exactly the same thing, believe) : "we have all these stupid whores waiting tables, how about we pick one to be a ''star and glamour&tc". "Which ?" "Whatever, pick one."

    There's this huge problem for ~all of society~ when the actual king gets replaced with some wanna-be moron : in his quest to having a kindom qua "having a kingdom", the moron enthroned wrecks choice and selection, in the process destroying not merely everyone's life as an actual lived thing, but the very ~possibility~ of even living a life in the first place, as an abstract.

    Which is ultimately why it's both undesirable and self defeating to crown morons (and yes, somewhere the crown will always sit), and some abstract collective moron no better than any peculiar one.

    Comment by Mircea Popescu — December 25, 2018 @ 5:38 p.m.

  2. You know, on meditation I'm particularly fond of that expression, "they expect in the dark nobody can distinguish an L from an E". Because how would that distinguishing go, you look for a spine and then something hanging off it, right ? "Well, was it an E or an L ?" "I dunno, man, there was definitely a spine, and something hanging from it." "Yes but how many ?" "Well, I grabbed on three times... possibly grabbed on the same thing thrice... or on three different things..."

    Comment by Mircea Popescu — December 26, 2018 @ 7:35 p.m.

  3. [...] http://ossasepia.com/2018/12/24/a-week-in-tmsr-10-16-december-2018/#fn5-4941 [...]

    Pingback by Full circle « Bimbo.Club — December 27, 2018 @ 9:57 a.m.

  4. The L/E type sounds good and even quite neat as a model (including the expression that is easy to get fond of, yes) but I can't quite say that it's necessarily so precise as that. I've been mulling this for a while but it still seems to me that you'll find all sorts although perhaps there's no point in attempting to differentiate between more than those 2 categories, I don't know. But it does seem just a variation on the usual "strategy" of "getting ahead" by holding back those in front or otherwise put socialism: in the dark it's not only L/E that are indistinguishable - perhaps there was a spine with some things hanging from it but perhaps the "spine" was just a rotten stick and those things hanging were cobwebs and whatnots.

    Regarding the links between various threads (here the labour allocation thread, China thread, Beria thread): many such links quite pop up at me when summarizing (one of the benefits I was talking of the other day) but I decided to keep the summary more or less contained as otherwise I see it easily becoming longer than the logs and possibly even more difficult to follow. The way I see it is that the summary has to remain contained - to a degree that is a choice of the author, not something strictly defined really - while further links made obvious will have to be discussed perhaps in a separate place or at least on their own in the comments or in another day's log or in some blog post, what else to do.

    Back to the well organized empires and their make of plastic objects, I guess the obvious is that it's "well organized" of a certain sort - a too pedestrian, low-level one that makes me think of a dead end similar perhaps to premature optimization at best or out of depth/idiocy at worst (a bit like a "well organized" book shelf that sorts the books by colour or size rather than content).

    Comment by Diana Coman — January 2, 2019 @ 5:14 p.m.

RSS feed for comments on this post. TrackBack URL

Leave a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <pre> <q cite=""> <strike> <strong>

Theme and content by Diana Coman