**Links to**: [[000 Introduction to the introduction]], [[000 Question]], [[000 Postulate]], [[000 Problems]], [[Problem]], [[Anticontinuous]], [[Bibliography]], [[Hierarchies of fields of investigation]], [[Logopatía]], [[Neologistics]], [[All knowledge is anecdotal]]. # 𝕀𝕟𝕥𝕣𝕠𝕕𝕦𝕔𝕥𝕚𝕠𝕟 𝕥𝕠 𝕥𝕙𝕖 ℙ𝕠𝕝𝕥𝕖𝕣𝕘𝕖𝕚𝕤𝕥 >“Every intelligent ghost must contain a machine.” > >A. Sloman, 2002. >“It is no longer possible to adopt the aloof and dissociated role of the literate Westerner.” > >M. McLuhan, 1964. >In light of the above, let us move towards better social frameworks. ![[windows 95 tips0.png]] <small>Fig. 1. Windows95 tips joke, author unknown.</small> ### Outline Before reading this, please first read: [[000 Introduction to the introduction]]. Below are a few preambles, and then the introduction ‘proper’ follows. The text below will run as follows: >>*Why the title;* >>>> *Why the postulates and questions;* >>>>>>*Writing about AI in the age of AI acceleration;* >>>>>>>>*A metaphorical preamble;* >>>>>>>>>>*Thesis outline;* >>>>>>>>>>>>*Introduction **proper**.* ### _[[001.1 The Poltergeist in the Machine/Untitled|Untitled]]_ Why _poltergeist?_ Initially, influenced by a recently renewed interest in the work of Hegel and its possible relations to the concept/project of artificial intelligence (e.g., R. Negarestani 2018), it was “Geist” in the machine. This title seemed, to me, self-explanatory: the mind/spirit is the machine (the power, the _maker_^[Etymonline: from PIE _magh-ana_- as that which _enables_, root: _magh_- “to be able, have power.”] of sense in sense-_making_) and is _in_ the machine, _Hegel ex machina_ style. But: the haunting had to be more absurd, because: _look around you_. The _poltergeist_ haunts, and _mocks_—laughs _and_ imitates—with annoying sounds ([[Noise]]). You never see it, but it looms. It is omnipresent but evasive, and impossible to capture. The title, then, says: there is something, most likely an elusive and simple myth, which keeps this annoying promethean fire burning. The _illusory_ aspect of the ghostly plays an important role in this title, too, as **illusion** is one of the fundamental concepts of this project, see [[Illusion]]. Chasing ghosts is, after all, all we do. Ghosts can be: pasts, absences, humiliations, far-away wants and desires (see: [[Desire]]), etc. This work chases its own tail. Because it’s fun, all animals do it.^[Also, if, as Catren says, we can marry the ghosts of Kant and Spinoza (_Pleromatica_, 2023), then any and all ghosts must be standing by, waiting for the total orgy. Allow me to embarrass myself. This has a few modes intermixed: _embarazar_, in Spanish, means to impregnate, to make pregnant. _Embarrar_ means to make muddy, to muddle, and has the connotation of _erring_, “to make a mistake.” Taking these two together, makes for a pretty embarrassing statement, which yields little more, big less, than a pregnant mistake.] For further thoughts on the concept of [[Geist]], [[Machine]], and _polter_ ([[Noise]]), please follow those links. ### [[000 Postulate]]s & [[000 Question]]s Before anything else it is important to mention that this work is structured around _postulates_ (see entry) and by *questions* (see entry), and that these are specifically defined. The following questions will not be answered in their entirety, but have guided the entire _prepostulation_ (intuition, desire) that structured this project: 1) **In chasing after _difference engines_,^[Computers have historically enjoyed the description of _difference engines_, but, as we will argue: all organisms can be described as difference engines in a very general sense (see: [[Xpectator]]), that is: in the sense that they (make) _sense_ (of) difference (in the context of active inference: error, surprise (Pezzulo, Parr, Friston 2022, p. 6)), and they are engines: etymologically, from _ingenium_: “that which is inborn,” from in- _in_ + _gignere_: “to beget, produce” (from PIE: _gene-_ “give birth, beget” — source: etymonline.com, accessed Feb 2023). The battle/war connotations of the term (from Late Latin “a war _engine_, battering ram” (Tertullian, Isidore of Seville)) are interesting to consider too, in light of the hostile, bellicose influences of war-developments on human-machines.] it seems crucial to ask: is difference itself reducible / compressible to something more fundamental than it already is?** This, I maintain, is not the question of difference versus identity but the question of equivalence _with_ difference imbued in it (Deleuze), and vice versa. To follow along these lines this continue reading [[Equivalence]], [[Equivalence and difference]], [[Negintelligibility]], and [[THE PATTERN BUTCHER]]. 2) **What do we want (from difference engines)?** [[Servomechanisms]], [[Prediction]], etc. 3) **Why does it matter to ask these questions?** [[000 Problems]], [[Model]], [[002.1 Modulations]], [[Language-modulating]], etc. These are, again, very general questions presented here in order to orient the reader into the guiding framework overseeing the full project, to follow more specific themes please see the list of topics presented in [[000 Introduction to the introduction]], and the questions presented in [[000 Question]]. ### Writing about AI in the age of AI acceleration It is not necessary to cite the myriads of frustrated AI voices chanting the fact that whatever is written about AI becomes outdated as soon as it is printed. Others, such as E. M. Bender, suggest we should stop using the term “AI” altogether.^[ref. #todo Also, regarding frustration, AI, and being sidelined, “Oh the frustration of being found frustrating! Oh the difficulty of being assumed to be difficult! You might even begin to _sound like_ what they hear you as _being like_: you talk louder and fast as you can tell you are not getting through. The more they think you say the more you have to say. You have to repeat yourself when you keep coming up against the same thing. You become mouthy. Perhaps we are called mouthy when we say what others do not want to hear; to become mouthy is to become mouth, reduced to the speaking part as being reduced to the wrong part.” (Ahmed 2023, p. 23). This has happened way too much in the context of my PhD. Not only this, but also: being listened to dismissively, as if the listener _already knows_ what I am saying, being looked at as if I don’t _look right_ for what I am saying, being ignored because of not _appearing intelligible_. Not being taken seriously. The guy next to you saying _the same thing_ you just said, but _him_ they listen to. I thought some collegial parts of the world were beyond this, in many respects, but they’re far from it. Also: “Killjoy truth: if you have to shout to be heard, you are heard as shouting.” Ahmed 2023, p. 34)]This much can be said about most emerging and quickly developing (“disruptive”) technologies, and perhaps about all published knowledge in our interconnected day and age. It is for this reason that this will be an evolving book. It is meant to serve as a resource for others interested in these topics, and an advancing landscape of thought. Previous versions of this “book” can be accessed thanks to the wonderful _wayback machine_. This also project takes advantage of the possibility of _hyperlinking_ (see also: [[Anticontinuous]]). If nowadays “automation can be dynamic and not dependent on a prescribed set of calculables” (Parisi 2015), then we could be tempted to ascribe to it the ability to produce its own conditions of evolution, since we certainly lack a level of _access_ to its possible prestatability conditions. But this, until proven otherwise, remains speculation. This is bothersome: humans create a pattern-digesting technology capable of exceeding the pattern-processing capacities of humans, but then: it becomes illegible, precisely because of this. And because of this, all reference to AI in this work will contain this paradox: like with any other concept, we speak of something in the making, which currently seems to speak back more than other technologies, and thus sets the ground for a reconsideration of notions which have been fundamental in historical self-conception, such as _technology_, _language_ and _evolution_. The gradual appearance of something (“AI”) which seems to enjoy a different relationship to these concepts than “humans do”, makes humans anxious. Also, by the way, language models can already generate something _very close_ to a philosophy PhD thesis. I refuse to write in the classical formulaic way of a PhD, that is: reviewing, summarizing, regurgitating. ### Why a critical philosophy of AI Before getting into the abstractions and subtractions of AI, let us explore something concrete about the origins of (symbolic) AI: John McCarthy’s—the person who coined the term _AI_—peculiar ways of thinking about unfaithful wives and fools. In the paper “On the Model Theory of Knowledge” (1978), John McCarthy, Masahiko Sato, Takeshi Hayashi, Shigeru Igarashi state that they set out to propose: >“Another language for expressing “knowing that” {...} given together with axioms and rules of inference {...}. The formalism is extended to time-dependent knowledge. Completeness and decidability theorems are given. The problem of the wise men with spots on their foreheads and the problem of the unfaithful wives are expressed in the formalism and solved.” Here is their formulation of _knowing that_: ![[mccarthy et al fools.png]] “Fool” we should understand to imply general _con-sensus_, common sense.^[This is not the common sense van Tuinen treats, when he upholds it versus the oppressive tendencies of ‘good sense.’ Reference in AI research, in general, to ‘common sense’ reflects precisely the insidious normalizing tendencies of what van Tuinen criticizes about good sense.] This is a big problem, we know, and this problem (“common sense bias”, also known as malignant normality (Adorno) or one-dimensionality (Marcuse)) is tackled in [[002 Semantic noise]]. In the “Image of Thought” chapter in _Difference and Repetition_, Deleuze presents us with this problem in a less foolish way, where he explains what an unbiased or presuppositionless beginning of philosophy might require, if a “presupposition” is what _everyone knows_: >“{I}t is not a question of saying what few think and knowing what it means. On the contrary, it is a question of someone – if only one – with the necessary modesty not managing to know what everybody knows, and modestly denying what everybody is supposed to recognize. Someone who neither allows himself to be represented nor wishes to represent anything. Not an individual endowed with a good will and a natural capacity for thought, but an individual full of ill will who does not manage to think, either naturally or conceptually. Only such an individual is without presuppositions {…} At the risk of playing the idiot, do so in the Russian manner: that of an underground man who recognises himself no more in the subjective presuppositions of a natural capacity for thought than in the objective presuppositions of a culture of the times, and lacks the compass with which to make a circle.” (pp. 165-166). Here is another presentation of a similar problem as that of universal fools in McCarthy et al.: ![[mccarthy wives.png]] Of course, this is meant to be provocative, salient, funny. But none of these practices is ever ‘innocent.’ The _move-fast-and-break-things_ legacy thus ensues. In _Violence and the Philosophical Imaginary_ (2012), A. Murphy sets out to treat how a violent imaginary saturates thought, and why. While we do not follow her proposal of tempering violence in philosophy, we do agree that these animating imaginaries (above, and others) demand a critical take. Murphy asks: >“If the imaginary can be understood as the locus of shame, what does this imply about philosophy’s own self-definition? What would it mean to claim that shame is the mood proper to philosophy, a mood both perilous and redemptive? And what would be the ethical stakes of such a claim? (p. 5). We can ask the same questions about the project of AI, which is shameful philosophy _in silico_. Besides these obvious geneaological trip-ups, _what is the problem with AI_? For starters, the obvious one: unbridled technocapitalist liberalism. Influential people like Guillaume Verdon, who maintains that AI development can be safe because it can be regulated by consumer choices in the market.^[Lex Fridman interview with Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast [#407](https://www.youtube.com/hashtag/407), min. 38-42.] What a bunch of bullshit. To say the least, if that was the case, we wouldn’t have any of the more visible disasters such as plastic galore, rising CO$_2$ levels, modern day slave labor, factory suicides, industry meat, deforestation, etc. etc. etc. Of course, technology is inherently accidental (Brouwer & van Tuinen 2023): try as we may there is no predicting the vastness of any and all risks. Though in the context of today’s ultraexperimentalist AI, one may also ask: whence the legitimacy of its antiscientific experimentalism? Double edged swords don’t even need to be double edged to turn back on one. And to meditate on this effect for the sake of a critical politics seems futile: “Customary pharmacological approaches—based on the truism that technology is both problem and solution—remain uncritical in this regard.” (van Tuinen 2023, _Technological accidents_, p. 6). Calls to awareness of this remain thus as banal as they are relevant. The uncritical dominant approach to technology as a human “universal,” is countered by Hui’s proposal of technodiversity, where it is, as van Tuinen paraphrases: “{o}nly through the affirmation of “technodiversity” does technics become universal. This means that we do not deduce difference from sameness (as in the identity politics of multiculturalism) but induce sameness through the affirmation of differences.” (2020, p. 99). Our understanding of this, in this project, is that we should not affirmatively—and arrogantly—understand our assessment of an/our identity, a/our self, a teleological Geist as inevitably _incomplete_ (always out of phase with itself (Simondon)), but as composed of a vastness of predictive dimensions which are all differently interested, and each of them as composed of/with inevitably contradictory interests (desires, initial conditions). Your mitochondria, my liver, tropical weather systems, memes on the internet, etc., are all coupled with each other, and are predictive events in their own right, the _enduring_ qualities of which only become salient through the modulation of attention. Attention is only possible through discourse (language as a semantic fractal), and we can only work at this with what we’ve got: symbols ([[E Pointing]], [[Vantagepointillism]]). Understanding dimensionally complex attention, intelligence, begins by understanding communication, as it is in the interaction between agents that something like “intelligence” can be referred to or even recognized. Communication is distributed, dynamic and dialogical, this unavoidable dimension has received little attention in the context of language-modeling, which inherited its functionality from the semantic-agnosticism of information theory. Insisting on the fundamental function of language as an evolving sociocultural process remains as banal as it is relevant: the image of language as a frame of reference with stable indexical footing seems to hold as a tacit understanding within world-modeling in AI research (“common” sense). This image does not reflect the fact that language is contextual, evolving and never a “mirror” of reality. Much research is needed towards what we could call a **post-stat(ist)ic computational semantics**: where novel understandings of _process_ and _emergence_ come to represent attentional salience, rather than operationalizing static meanings and statistical majorities as the yardsticks. Instead of assuming universality, sameness, we should explore the negatively constraining effects of maladaptive ‘scripts’ (Albarracin et al 2021) which have universalist, dominating predictive tendencies, such as the drive to optimize, profit and expand: the experimentalism of commercial AI is driven by unscientific principles which exploit and dominate unwilling and/or unaware systems. It should be demanded, while academia still has the legitimacy to do so, that scientifically-informed, and scientifically-_paced_ AI research becomes the global norm, much in the same way that the field of medicine attempts to counter maladaptive practices in capital-driven pharmacology by protecting patient-safety. “Scientific” is here understood not as perfective predictive optimization (Solomonoff induction), but as dialogical, transdisciplinary experimentalism which seeks to heal the gap between being born and _suffering_. Not survival (perpetuation of **x**, **y** or **z**), but the minimization of anguish. But what can we say, other than “proceed with caution”? The alternatives are sparse and, arguably, unrealistic. Not proceeding seems impossible, we’re marching down a one-way street already. Proceeding without caution rings ridiculous and careless (though, again, most AI nowadays is based exactly on the principle of moving fast and breaking things). And if, “the resistance to contingency is the signature of Western philosophy: the overcoming of the irrational through reason and control.” (ibid., p. 7), then why not “take a chance” with the madness of serpents? We have always lived in a time when “[t]he privileged are processed by people, the masses by machines.” (O’Neill 2016, p. 8, O’Neill claims the time is now). This is because the (_poltergeist in the_) _machine_ is the—necessarily—mindless, intersubjective, disembodied ‘cultural algorithm’ on the loose: predictive strategies (“those are not human beings”) that perpetually eat at the margins (the margins are: the marginalized, and the future). Whatever the politics that emerges out the other end: we are fucked. That is for sure. Because we have always-already been and are currently certainly fucked. ### Thesis outline The proposals made here are simple, and have emerged in this order: 1) One thing can mean many ([[002 Semantic noise]]) and that is what enables all generativity ([[Polycomputation]], [[Language-modulating]], [[Meta-language]]). While the banality of this point screams loud enough, it cannot be stressed enough that we still speak as if this was not the case. 2) These meanings are obviously distributed, extended and co-constructed (via interindividuation, constraints) between perspectives and their environments, which are also perspectives ([[Vantagepointillism]]). 3) The mixing/overlapping/intertwining of perspectives is the most interesting thing ([[E Principle of Sufficient Interest]]) we know of ([[Assembly and assemblage]]). 4) Control and agency are a farce and we should modulate our verbosity away from that ([[001.1 The Poltergeist in the Machine/C is for Communism, and for Constraint|C is for Communism, and for Constraint]], [[Z Post-Control Script-Societies]]). 5) If the future didn’t seem incomplete, we wouldn’t be able to think ([[Prediction]], [[Free energy principle]], [[Negintelligibility]]). But this is a dangerous proposal: (see, again: [[Z Post-Control Script-Societies]]). 6) The question of difference bottoms out, socially, at the question of: can we live without forcing each other (telling each other how to live)? ([[Possibility]], [[D Bias, Falling into Place]]). 6) The question of difference bottoms out, antisocially abstractly, at the question of: is difference compressible to something more foundational than itself? ([[Difference]], [[Homotopy Type Theory]]). One of the main problems this thesis sees in itself and other (philosophical and more widely: all) theses implicated within is the same as the opening question it chases: _how to reconcile commonality with absolute difference, in the socioconstructive realm_? Why is the little red book both good _and_ bad? Why is science about consensus _and_ eternal destruction? Etc. %% Don’t forget to include NRU diagrams+descriptions in thesis. and this one: to explain what you are focusing on (van Rooij et al 2023, p. 2): ![[van rooij et al ai types hypes.png]] %% Everything that follows is wrong.^[This is not only a simple statement about what science and speculation _should_ be, but also in the spirit of Walter Benjamin’s catastrophic thinking. That is: the catastrophe is not exceptional, the true catastrophe is that the catastrophic drags on, business as usual (for an extensive account of Benjamin’s development of this see Vandeputte 2023). This sentiment is echoed in many pessimistically generative thinkers of today, e.g., in the context of discrimination by design in technoscience by another Benjamin, Ruha, that is: “Antiblack racism, in this context, is not only a by-product, but a precondition for the fabrication of such technologies—antiblack imagination put to work.” (Benjamin 2019, p. 5). Is this “perverse calculus of human worth” (ibid. p. 7) the lowest common denominator? The answer is, unfortunately, *yes*. It is no surprise that the accounting of debt, for example, lies at the origins of the earliest mathematical endeavors: our very cherished abstract negation might have been invented for the simple convenience of some enslaving papertrail: Li & Du refer to the uses of negative numbers in ancient China as emerging from accounting for deficits (1987, p. 48). They also note that “The introduction of positive and negative numbers into mathematics in ancient China is a most outstanding invention. In India the concepts of positive and negative first appeared in the seventh century in the work of Brahmegupta (620 AD, see Colebrooke 1817/1973 Chapter XVIII, p. 339) and in Europe they had to wait until the 16th or 17th century before they had an accurate knowledge of positive and negative numbers e.g. Bombelli 1572/1966, pp. 8-9 of 1966 edn.” (ibid., p. 50).] Nobody can be trusted. (Why) does all criticism assume some sort of background assumption that we already know how to live? In the spirit of [[Selimism]], this is a work against puritanism and arrogance. Accordingly, it contains a high degree of arrogance and purity. None of this could have been written otherwise. “Something in the world forces us to think” (Deleuze in “The Image of Thought”, DR, 176). Thinking AI is thinking god, from intervention and souls (Epicurus, Descartes, the pope) all the way to Rokko’s basilisk. >“**PROFESSIONALIZATION IS THE PRIVATIZATION OF THE SOCIAL INDIVIDUAL THROUGH NEGLIGENCE**: ... And saying so we prepare to part company with American critical academics, to become unreliable, to be disloyal to the public sphere, to be obstructive and shiftless, dumb with insolence in the face of the call to critical thinking.” Moten and Harney, _The Undercommons_, p. 34. ### Before reading _anything_ Are you *reading* this? Are _you_ reading this? Are you reading _this_? To be able to begin, let us settle on the fact that **this** is _reading_ (so, stressing. “are” won’t be necessary). Before you even continue reading. Before writing sucks. Between the paranoid *guilt*, why? Any argument can be made, strategies of energetic transfer are plural: the giraffe, the yawn, annihilator, romance ([[Polycomputation]], [[Illusion]]). Any niche one finds oneself thrown into. Being is ~~geworpen~~ (thanks for nothing, H.): created and at the same time discarded, ejected from non-paradise, a living thing is a miracle and/or a vestige—is indeed: a niche. It _is_ particularly afforded. In the niches of science or philosophy novelty suffers a cruel (ref. Derrida on Artaud) fate. It must be, but it also cannot. There needs to be a melody, a coherence, else the niche cannot listen or read. This pains me, but at the same time it is not me embarking upon this. For whatever I can call I, the shifting pronoun that now happens to be dislocated here and now (and for you, there and then, for me, for now. Can you see _you_ from where you are?), the desire is to create a non-niche. Is this a death-driven pursuit or the pursuit of the new in death? There are _some_ types of people: those who laugh and smile and those who do so but a lot less. (Please see: [[Laughter and polycomputing]]). Sorry about that. Apologies always need to be made when asking for someone’s attention. It seems this type of philosophical care is about what seeps through the pores ([[Trypophobia]]), the same old tragedic tortured artist s/tale, again and again. Mischievous and misunderstood. The care to read something that will change the state of things, all the while keeping in line: _let’s_ we forget. Lest we do it. The royal wee. If you (dis)agree, why read on? The main premise and final argument of this attempt at an impossible non-niche is that pronouns and verbs are the most polycomputing objects we know of (everything polycomputes: play is not just play but practice, laughing is not just a physiological response but a type of semiosis, a radio is not just the technoscientific harnessing of a phenomenon but the enjoyment of music at the other end (Whitehead)). Self-interrogating this idea from its own perspective means doing away with nichic specificity: everything must be able to do everything else (everything has the name of something else: an _[[Irrespondence]]_ theory of truth). To say “it ought to move” is rougher crasser and bigger than to finalize plans for a windmill. Yet **it** *moves* (see: [[Spatial reasoning]]). The moving verb is our only true perpetuum mobile (until the end of something which to me doesn’t matter right now, unlike for Lyotard, see: [[Sex on the Sun]], [[Schematism and the moving verb]]). I wholeheartedly follow Patricia Reed on these questions: >“If sensitivity to interference is a minimum criterion for the possible transformation of a world, under what conditions can given world configurations of insensitivity be suspended? How is it possible to localize conditions of otherworldly thought that are not absolutely beholden to the here and now of the local belonging to a given world configuration? In other words, how to localize speculative thought (which necessitates the construction of a world to embed it), and what are the consequences of such activity on an epistemological register?” Reed, Organizing Gestures, p. 5. ### A. In light of the above frictions, there are three principles to the way this text has been composed: 1) Everything we know in philosophy is accessed through names, and everything has the name of something else. Trying to see ‘around’ that is a hyper- ór a nonphilosophical task, in the most radically immanent and pedantic way. 2) Everything is read continuously. To read 3 pages of Kant, watch a cartoon and thereupon walk around the block, are all reading tasks, all of which are continuous to each other and part of the same process: the coherence of the singularity that types these words.^[See also: [[Free will]].] 3) It remains generative that these things should serve as reminders. 1) There can always be more principles. ### B. There are three principles to the way intelligence is framed in this text: 1) Intelligence is the eternal re-formulation of difference, and everything is dyanmically differential. To see ‘around’ this is an intelligent task, in the most philosophical way. 2) Intelligence is continuous. To read 3 pages of Kant, watch a cartoon and thereupon walk the dog, are all intelligent tasks, all of which are continuous to each other and part of the same process: the coherence of the reception that reads these continuous words. 3) It remains generative that self-evidencing brushes up on itself in these ways. 1) There can always be more principles. ### C. There are three ways in which you may follow this text: 1) Linearly, following chapter enumeration. 2) Non-linearly,^[When I came across the following advice by Michael Levin, I was honestly emotionally moved, as it is strikingly close to the feeling/logic/structure/reasons I started writing this ‘thesis’/project the way I did: “{T}he current paradigm of a “linear paper” as the only way to tell a story has to get radically expanded as we go forward. At some point there should be software and publication platforms where non-linear mind-maps and other structures can be published, allowing a much richer information conduit by means of which writers and their readers can interact and exchange ideas. Until then, here’s a suggestion for content that is really non-linear. Suppose your paper has pieces A,B,C,D where the reader needs to know each piece to appreciate the other pieces – like an interlocking puzzle that only makes sense as a whole. Which should they read first? Think of it as a spiral. Don’t try to say _everything_ in a linear way, but do it in small circles and then bigger circles. For example, order it like this: a,b,c,d, A, B, C, D. That is, make a few short and simple statements about all of the parts, then really launch in. That way, when they encounter A, they’ve already seen b,c,d which helps to understand A even if it wasn’t the details of B,C,D yet. You can do multiple circles of increasing size – take the reader through the spiral of ideas saying each time just enough of the key concept they need to keep moving through the structure.” Levin, https://thoughtforms.life/how-i-write-papers/, accessed Jan 4 2024. This is also almost exactly the same advice I have written in 2019 for BFA thesis writing at the Willem de Kooning, and in 2023 for MFA thesis writing at the Design Academy Eindhoven. I did not refer to it as a spiral but as a fractal, emphasizing the vast differences between possible fractal structures: arborescent, linear in one or multiple ways, apparently chaotic, etc.] following concepts as you see them emerging; or in whichever other order you wish. Alternating between other texts is possible, too, considering principle A.2.^[![[non linearity ATP anticontinuous.png]]] 3) Through the alternate orders provided throughout the thesis (Rayuela-style), or the linear order provided in [[000 Introduction to the introduction]]. ### Style _A note on style_: (See also: [[B Violence of Context and Style]]). My supervisor (Sjoerd van Tuinen) once mentioned that my writing seems so idiosyncratic that it sounds as if I am talking to myself. Perhaps this is so—since we are most definitely all idio-synkrats: individuations—I don’t understand what it is we do if we don’t engage in a strangely looped self-dialogue. After all, if this fleshy entity holds a model of its fleshy interloculors in order to speak, then it is only me who is able to speak, to those models, inside myself (so much says Hegel about [[Freedom]]). This is no seclusionary, isolistic, defeatist practice: all parlance is material and social (Ferruccio Rossi-landi challenged Saussurian semiotics with this proposal), on the contrary: it is a challenge for us all to see how deeply indebted we become to the shadows and imprints that others make on our own inner ghost. As for the unintelligible idiosyncrasies, you’ll have to forgive me, but perhaps echoing Nishida I can say: “I have always been a miner of ore; I have never managed to refine it.”^[Thank you, Jamie, for pointing me to _Intelligibility and nothingness_.] And I would add: given my avoidance of a choice between technopessimism/optimism/indifference, I don’t know if the desire for refinement is even possible. The attempt renders experiments which are just as interesting as the death of trying. Also, following Eco’s note that “A well-organized thesis should abound in cross-references.” (p. 113) I have decided to cross-reference by hyperlinking. Also a point about science in general (which Eco also reminds of): a thesis should help others’ interests.^[Eco also advises that one must write a critical academic work in a metalanguage, without falling prey to stylistic reveries and "alternative" "poetic fury". (pp. 149-150 _How to Write a Thesis_). I am still not sure about this. As mentioned, this thesis can, in fact, be written in a clear and "lucid" metalanguage, such as the one promoted by GPT language models. However, it's important to reflect on the different effects that are lost and gained when opting for the type of clarity proposed by, generative language models. On the one hand we gain an initial _sense_ of objectivity: the words the author is using are sharply aimed at a very specific locus and there can be little to no ambiguity intuited within them. The text appears as direct, referential, univocal. In a lot of cases, where nuance and ambiguity are preferably avoided (which cable to cut, which city to go to, which author died when, etc.), one is obliged to adopt a clear and uncomplicated style. However, as we will see in e.g. _Semantic noise_, staying lucid is not always possible, especially when each and every word-choice we make implies allegiance to an unspecified background morality. For example, to continue with Eco, he comments on how "Krasnapolsky is not a very sharp critic of Danieli's work" is a referential sentence, while "We are not convinced that Krasnapolsky is the sharpest critic of Danieli's work" (pp. 151-152) is a litotes, a rhetorical move which should be avoided. However, in the second sentence, the filter that is the writer of that sentence lets themselves be known. Is that not less ambiguous and objective than the first sentence, which makes a sweeping statement about Krasnapolsky in general? One response to our question may be that by the mere fact that the statement appears in book/article/etc., it is implied that this pertains to the opinion of a specific filter, a specific person(s). But if it is common practice that one should repeatedly remind the reader of the central questions and arguments in a piece of academic text, I think it should also be accepted that one should remind the reader of the real, biased, human, flesh and bone person at the other end of a text, too. This includes rhetorical moves which will not always be effective, because a reader may appreciate or despise them. It is all about how the cosmic joke is delivered and interpreted. Eco says "Now, we either use rhetorical figures effectively, or we do not use them at all. If we use them it is because we presume our reader is capable of catching them, and because we believe that we will appear more incisive and convincing. In this case, we should not be ashamed of them, and we should not explain them. If we think that our reader is an idiot, we should not use rhetorical figures, but if we use them and feel the need to explain them, we are essentially calling the reader an idiot. In turn, he will take revenge by calling the author an idiot." (p. 152). But how? Amongst our readers may be idiots. Additionally: I am personally appalled that a word such as "idiot" may find its way into a text such as this one. The image of the idiot is itself idiotic (my apologies). There are people who hold different opinions, there are lives which are spent in ways we do not understand. There are no "idiots". Eco says it himself when he says that "minor" authors are not to be ignored, and that one should be open to learning from anyone (p. 144).] However, in current academic contexts theses ‘help’ the top 2% (ourselves) who sit and watch the wheels turning while we say: “they're not perfectly round! This might or might not have implications for society at large!” This bothers me. It bothers me in particular within the context of a university that would rather dispose itself of criticality by brute force (see news events of November 2022 and February 2023). I am not saying a) that this thesis is accessible to all, or b) that I am contributing something new, but I am trying to make this as accessible as possible, by using a different format and promoting this work beyond academia. Melanie Mitchell, questioning the paths of symbolic and statistic orientations in the field of AI, states four fallacies. Mitchell has highlighted 1, 2, 3, 4. The theme of this thesis gravitates around ... #todo . The problems investigated here pertain to language issues brought up by 20th century scholars such as Jacques Derrida and Ludwig Wittgenstein. However, the main driving force behind this work is and has always been the work of Jorge Luis Borges. He cannot be said to have asked this question, but he certainly inspired this question in me: how can a few scribbles represent something as vast as the infinite? And in this infinite, it is likely that, because of the willing adoption of language-tools such as GPT, the production of text will drastically increase (making a lot of journalistic/editorial/etc. work obsolete), but because of this it is also likely that the need for curatorial and summarizing services will increase as well. What this means, abstractly, is that we will probably witness massive amounts of texts being produced from “basic units” of info-meaning (e.g., as innocent as: “write a friendly email to my coworkers that I will not make the deadline”), which at the same time, on the other end, become summarized back into these “basic units” because people don’t have time to parse through vast seas of “useless” grammatization (e.g.: “write me a summary of all the incoming email regarding the project deadline”). This leaves me wondering, not whether there is any objective reality to the “basic units” (which I don't believe there is), but where the belief in an objectively summarizable (and, at the same time, expandable yet expendable) reality comes from. Perhaps the question is even this one: where does the biased focus on language as a tool come from? (Is it just a vestige of capitalist product-oriented obsessions?) Why, when we are impressed by the poetic abilities of ChatGPT, can we not also appreciate language in general as the basic social fabric that holds our relational dynamics together? The “bias” discussion on the partiality of systems seems not to have taken actual critical hold, because we continue to promote systems (such as academia) which supposedly parse reality in “neutral” and unbiased ways. We promote music above language, like Nietzsche. This is not to say all language is poetry (although it is: all language is a poetic call to order). If the near future of AI will be determined by prompts, all of this is of the essence. It is paramount that we look into the structure, dynamics and rootedness in matter, of language, so as to understand what we want to ask of AI. The paradox here is that AI is an illusion collectively composed. These arguments are about the speed of thought, and not much else. Incompleteness in computation/mathematics doesn’t allow for the proof of “one-dimensionality” (one truth, one objectivity), so for now we’re bound to remain in challenges which only deal with the tempering of speeds. Anything we put down on paper opens itself up to generative criticism. At the same time, it’s bound to remain only a symptom, always. There is no depth to symptoms, symptoms are all there is. AI is plastic. AI is like plastic. AI is the idea of plasticity. [[AI is plastic, plastic is AI]]. This thesis can be defined as a “brain dump” (what Scott Aaronson calls his own book, _Quantum Computing Since Democritus_, that is: a collection of thoughts that were on the author’s mind around a specific period, with no clear overarching thesis). I guess I have apologize for this possible disappointment. ### Are we there yet? #todo As it is with most conclusions, the chapters that follow (may) actually lead to it. However, when I wrote the first chapters I didn’t know I was writing about negintelligibility until it dawned upon me. In the first chapter we explore how AI praxis, unable to deal with the fact that a truly objective, unbiased, all viewing god’s perspective is impossible, deals with its shortcomings by approaching intelligibility/intelligence in various self-defeating ways. This renders the refinement of both concepts; of intelligence and of artificiality, by way of the negintelligible. .... finish introducing Chapter 1 and conclude that AI follows negintelligibility in a, b and c ways, but has not yet dealt with the more ‘material’/enactive ways in which negintelligibility happens. All action, all influence, all vantagepointillism (see: [[Vantagepointillism]]) require perspectival movement (in the crudest of terms: a vector). Observing varieties of spatiotemporal inclination by exploring the Free Energy Principle (the topology of something blanketing itself off), Hegel (the tendencies of reason) and gradient descent (the tendencies of AI), and gravity (the tendency of it all), is the task of Chapter 2 (Falling into Place). Meaning is a particular inclination some systems, such as human groups, tend and attend to, Chapter 3 (Semantic Noise) deals with meaning in the Natural Language Processing (NLP) task of pronoun disambiguation in Winograd Schemas. Chapter 4 (Principle of Sufficient Interest) deals with the way knowledge specializes socially by individuating itself through interested agents. Chapter 5 (Post-control Script Societies) deals with the having to come to terms with larger-than-life systems and how we are subject to them. Chapter 6 (Negintelligiblity) concludes the thesis and retroactively changes all the previous chapters by showing what was common to their arguments and observations all along: the pull of the negintelligible. A note in the form of a ‘•’ will be placed wherever we—I—feel the negintelligible can be introduced in order to intuit something beyond what has been typed. Negintelligibility deals with the fact that one can’t choose to have an insight: an insight happens to one. Otherwise we would work very hard at producing permanent insights all the time. Indeed, we do, the way we do that is by relying on intelligible strategies that track the future in different predictability degrees. What challenges these tractions is negintelligbility, outside of ourselves, relinquishing us from the burden of localizable agency. The externalization of thought is not a possibility but a fact. The movements of the extended mind (Chalmers and Clark), such as this text, these very letters, suffice to provide evidence for the phenomenon that I am not here saying them to you—though in many ways, whatever it is that designates this “I” pronoun is, I am—but you find yourself conjuring a spectral voice which is neither mine, nor yours, nor anybody else’s. It is this text’s. The challenge lies not in finding a homuncular agency to, e.g., this text. The interest of this thesis is to explore a space in which agency relinquished from a principle of sufficient reason, so that the movement can be made towards non-individualistic thought outside the head. This predicament is simple but understated, and ubiquitous. Being meat-based phenomena with the capacity to externalize language-outputs, which are unavoidably dialogical and materially distributed, offers the possibility for this meat to be modelled, molten, measured and massacred by the language that envelops it. But reverse operation, the vice versa, is the other side of the coin. This is to be termed _language-modulation_, as will be explained later, in the Chapter C is for... where the image of choice and control is adjusted to/amended with a hard determinist clause, which strangely enables a new kind of modular versatility (perhaps another version of “control” itself). When M. Beatrice Fazi reflects on the necessarily contingent in computation and states that “{interactive computation} becomes an operation concerning the orientation of computing machines in the world, and our relation with them” (Fazi 2018, p. 3) what is easily missed is that in _stating_ the relationality between non-computational thought and computation in relation to something, one already has positioned the possibility of an abandoned computation, elsewhere, without this relationality. _Where is this_, other than as a strong, prevalent predicate over the subject that is all computational thought: everything that is coupled to it and _discourses_ its material evolution? The task is to modulate this semantic space, where modulation implies both attending to the the different modes afforded by a specific modality, as well as the exploring the space of articulation, the variations and edges of these possible syntheses (as in modal music). To make things less abstract: computation is not only an unfinished and definitely contingent vision of the possible spaces of logical articulations, but also dogs, dust, meat, abuse and UFOs (please note: this is not a flattening of these categories but an enumeration of diverse modes in a space of thought), because these too are logically-embraceable. Conversing with these links is modulating them: the labor of engagement with terms is the function of these terms is their effectuation is their absolute power over the material dynamics that ensue. This work tries to come to terms with the implications of a computational-semantic-aesthetic engagement with itself as computation, semantics and aesthetics. The drive is to present an image of elaborated cognition (that is: labored through, and material to the point of absolute determinism) which becomes an interlocutor between the deeply isolationist philosophical image of AI, and the ghostly, blurred and impossible image of AI presented in the popular realm, emerging out of Silicon Valley. In the context of the computation of semantics (i.e. the technologically-mediated modulation of that which we call meaning), Hinton and Shallice (1991) made a significant leap through the use of feedforward networks with backpropagation—essentially a process of feedbacked filtering towards a return, where initial iteration and its filtering through recursion result in the refinement of teleological data structures—which the present work will take a lot from: semantic attraction. The semantic attractors presented by Hinton and Shallice are data structures towards which an artificial neural network _tends_, is _inclined_ towards (more on this in [[D Bias, Falling into Place]]), as it refines a search space. An attractor, in complex dynamics, is a set of values towards which there is a tendency. It can be best pictured as the way the water turns when you flush the toilet (see also, relatedly: [[𝚊̷ 𝚕̷𝚊̷ 𝚖̷𝚒̷𝚎̷𝚛̷𝚍̷𝚊̷ 𝚌̷𝚘̷𝚗̷ 𝚕̷𝚊̷ 𝚒̷𝚗̷𝚝̷𝚎̷𝚕̷𝚒̷𝚐̷𝚎̷𝚗̷𝚌̷𝚒̷𝚊̷ 𝚊̷𝚛̷𝚝̷𝚒̷𝚏̷𝚒̷𝚌̷𝚒̷𝚊̷𝚕̷]]). It will never be the same: each flush is unique, yet the water’s general structure, its _tendency_, can defined as a set of gravitations towards which it inevitably falls into. As Alicia Juarrero (2023) explains in the context of semantic attractors: these initialize and reset conditions and, they determine the different conditional probabilities which result in the extraction-production of features. Feedback loops, such as those enabled through backpropagation, or those in human learning through repetition, are instances of coherent, constrained dynamics (Juarrero 2023, p. 101) which result in the mereological, emergent properties that are semantic attractors, or learned behaviors, as will be argued. This work will follow a strong sense of camaraderie with the proposals of coherence through constraint presented by Juarrero, an important divergence, though, is marked by this work’s tendency towards a proposal that _trans-verses_ ideas of autonomy and self-determination, in favor of contextual communal embedding. { #todo notes on computation + AI in Reza, Anil, Inigo, Miguel} If computation is, due to possible demonstrations of its different incompletenesses, “an abstractive procedure of determination that always confronts indeterminacy” (Fazi 2018. p. 5) then it makes absolute sense to confront this indeterminacy, not only in logical heaven, but precisely as coupled to contrastive procedures in what is supposedly non-computation: as sharing the same substrate, comparable constraints. Constraints are formal limits: the phenomenon that is water, and the fact that sound comes out of people. If the _formality_ of formal systems lies in their self-evidently true axioms, then what exactly is self-evidencing? If the evident truth of an axiom—in some senses of computation, at least—lies in its capacity to act as an absolutely delimiting constraint on the possibilities it can afford, then as we will observe in the context of predictive processing philosophies, the self-evidencing organism can be tied to that which is _self_-evident in an axiom in interesting ways, and these open a path for the modulation(s) proposed. But what happens when the model is thought itself? AI is no longer a computing machine but an idea about what thought can possibly be. (Reza, Patricia, Anil). This is the poltergeist in the machine. An attempt to ghostbust something which haunts, something ungraspable and possibly even inexistent is the best way to get at the limit. Assuming there’s a limit (the ghost), where there is no limit (the ghost). In the case of this project, on the image of language (following Gastaldi who follows Deleuze, more on this in a moment) the theme has been narrowed to a specific degree: What image of language does contemporary AI present, and how does it represent (or fail to represent) the reasoning functions of language? (which, for the sake of specifying this project's focus even more: are primarily social). This **"image of language" is the object under study**, and in the introduction we will define the conditions that set the groundwork for an "image of language" to be spoken about. However, what pertains these conditions, it has been difficult to narrow the interpretation to a single field of philosophy, or even a single author, given the richness of the topic and its many possible openings. Therefore, we will treat avenues as correlated and parallel as Peirce and semantics, or as superficially disparate as bird migration and neuroendocrinology. Closing down these openings means doing injustice to the project: how can we project one specific approach to something (the image of the functions of language as presented by technological innovations) we propose is indefinitely inherited, interpretably variable and in constant evolutionary flux? The conundrum that turns language into a complex object of research is no other than the paradox of self-reference. The complexities ensuing from this paradox are publicly recognizable, yet in public it seems we tend to glide over them, _avoiding_ the abyss of eternal regress, in order for language to do something _else_, beyond itself. This utilitarian image of language, of language as something serving another purpose, will be placed under particular scrutiny, it intertwines (techno)utilitarianism, teleological/cybernetic accounts of reason, functional/purposeful/formal schemes, etc. This "something else" language tends to is another vector of analysis we will explore, in terms of concepts as semantic attractors. The proposal being that concepts are (famously presented by e.g. Derrida, Deleuze) forever unfinished is nothing new, but we revise these perspectives towards a different understanding that aims to connect the topological to the semantic, thereby connecting language-modelling and speculative philosophies of language. %% __________ 0) Image of language (Deleuze, Gastaldi) 1) Introduction to Geist, machine, reason, language-modelling: 1) Two contentions, we will define reason as the _orienting_ activity, most often attributed to adult human beings (but sometimes also to other beings), which leads to the optimization (no definition of "optimal" may be given, it depends on what the desiderata may be) of the manipulation (depends on the capacities and constraints of beings) of their environments. This way we can understand the mathematical conceptualization of Markov blankets as a reasoning endeavor, as well as the understanding of and therefore engagement in the nourishing of another living being, as a reasoning activity. Language is often an important component of reason, and we define it 1) as a system: because a system is an enduring, dynamic whole made up of simpler components which may change while the system, in "essence", remains the same, 2) as a collective endeavor which presents single or collective actors the possibility of diverse environmental explorations (which would not be possible, or drastically different, without it). However, a very important note: as we present it, both these concepts are _semantic attractors_ (something which can be said of almost all other things we call _concepts_). They are collectively composed approximations, no single exact definition of either may exist, and as such we will treat them as largely undefined. Whenever a definition is needed, the reader may return to "reason is an orienting activity in optimal affordance manipulation" and language as "a system shared by a collection of beings which supports the collective exploration of affordances." In addition, in true dialectical fashion, sometimes we may want to say the _exact opposite_ of what these definitions hold as perspectives, for the sake of evolving our concepts. In this way, for example, when thinking about reason as "an orienting activity in optimal affordance manipulation", we may want to challenge these grounding concepts: sometimes reason may benefit from disorientation, from a desire towards the suboptimal or from an altogether refusal in engagement. These things are possible, and pose no obstacle or interference, they are treated as matters-of-fact when it comes to dealing with concepts. As we will explore in section __: conceptual closure is almost always neither possible desirable. 2) Introduction to language-modelling problems (Floridi and Chi., Bender and Gebru, Markus, Gastaldi) 3) 1. The state of the issue, 2. The previous research, 3. Your hypothesis, 4. Your supporting data, 5. Your analysis of the data, 6. The demonstration of your hypothesis, 7. Conclusions and suggestions for further research. 2) Chapter: Winograd schemas: Metaphors 3) Chapter: Language-learning: PSI (address sellars earlier, and don't forget to add Tarski truth "infinite regress" he may perhaps also be referring to) 4) PP and language: Pointing 5) Another problem: _____ (find in Rieder or elsewhere) 6) Falling into place? 7) Control 8) Conclusion: another image of language We will proceed as follows: firstly, we will sketch an image of reason in very general philosophical terms (Agre, Rieder) and how it has been inherited by language modelling (Gastaldi paper, Louise Amoore critique, Dutihl Novaes deduction). Secondly, we will focus on a specific problem in language-modeling that shows the problems encountered in the introduction (Winograd schemas), this section contains the main charges against the image of language presented in contemporary language-modeling: which we frame as one-dimensional common sense bias. Thirdly, we will show how contemporary metaphors structure the image of reason in language modeling from three different points of view: falling into place (Gradient descent, PP, and spatial metaphors). Once we have traced these perspectives, we will revisit the concept of reason as _interested learning_ and question what we have seen so far, in terms of a remarkable sociality present in language-learning that is often forgotten (PSI). Then we will focus on the programmatic aspects of language, the script, the prompt, and Deleuze on control. Finally, we will conclude with: another image of language. "The final product will be nothing original, and it will fall between the survey and the monograph.... The student will avoid reproach for not having read all Panofsky, including work available only in German or English, because the thesis is not on Panofsky. Panofsky is relevant only to a specific aspect of the topic, and is useful as a reference only for some questions presented by the thesis. As I said in section 2.1, this type of thesis is not the best choice, because it risks becoming incomplete and generic." Umberto Eco How to write a thesis, p. 25. %% %% ### Abstract If there is anything like an abstract to this work, it goes something like this: Concepts are predicates: arguments (in Peircian terms: they are symbols in a sea of signs). Arguments, which house concepts, are _predictions_, estimations based on expectation (priors). The predictive framework of this project is PP in the light of Friston and Clark. Under certain circumstances (e.g. [[Winograd schemas]]), these predictions fail to account for alternate modes of existence, ones that do not fit the bill of the makers. Concepts, which are metaphors, and metaphors, which are concepts, tell stories _and_ narrate conceptual predictions, which is why they are so prone to being used. Stories are "training grounds" for the grounding of concepts. All behavior depends on previous behavior (PSR), making it deterministic, falling into place. All learning (through stories, metaphors) depends on caring relationships (PSI), based on predictions. Gravity governs behavior and metaphors. The concept of reason is bankrupt. It revealed itself to be inaccessible to the intransparent subject, and even then: mechanical objectivity turned it obsolete, unreasonable. This project has only three things in mind, which merge mechanical objectivity to reasoning schemes in the flesh: 1) the pursuit of AI, particularly in the realm of natural language processing, forgets that language is non-individual, dialogical. 2) Everything an agent understands as a *reason*, is always social and always retroactive, and that narrative persistence owes everything to gravity. 3) The thinking condition thinking the thinking condition can also be understood as an unfortunate glitch in the simulation that is an agent's self-perception. "Each of the three parts starts out from the narrowest private realm, that of the intellectual in emigration. After this follow considerations of wider social and anthropological scope; they pertain to psychology, aesthetics, and science in its relationship to the subject. The concluding aphorisms of each section lead thematically, too, to philosophy, without claiming to be conclusive and definitive: all of these are intended to mark points of attack or to generate models for future exertions of the concept.". (Version on: https://www.marxists.org/reference/archive/adorno/1951/mm/ch01.htm, translation: Dennis Redmond, 2005. In: _Dedication_). The concept of intelligence remains largely undefined, be it in the field of cognitive science, AI or philosophy. The main purpose of this project is to elaborate an elucidation of the necessary ungroundedness of the concept of intelligence which borrows from said fields, as well as from psychology, biology, and others. The employment of these various bodies of knowledge is not only meant to provide a large enough context for the themes pertaining to intelligence, but also to allow these fields to converse with each other, given many discussions could benefit from not talking past each other, as they deal with very similar issues. The discussion is also unavoidably selective, given the vastness of the available literature on the topic of intelligence, ranging from molecular biology to political science. What is studied here is aimed at finding points of agreement and divergence, particularly by observing the continuum between certain philosophical claims pertaining to *meaning* and *reason*, and certain scientific models and recent findings pertaining to *prediction* and *intelligence*. From the perspective of this research, it can be quite unambiguously stated that philosophy needs to reflect on the scientific imaginary, and vice versa. 'Imaginary' here is not meant to invoke an undertone of 'social construction', 'ficticious myth' and neither 'naive objectivity': that which is imaginative pertains to all knowledge-seeking endeavors; it is the access to and retrieval from the combinatorial space so often attributed to what is termed _intelligence_. The access to/retrieval from a space of possibilities which is _separate_ and experientially different from the inevitable state experienced through the senses implies, paradoxically, an unavoidable lack of access to what appears as given and contingent (Kant). The ability to know that one doesn't know could be said to be one of the most basic requirements for knowledge-seeking, and the intentionality or volition that drives this process is undergirded by a multiplicity of factors, including the predictive drive behind abstract representation in scientific models, or, perhaps, hormonal homeostasis. The challenge lies in understanding the relations between these by placing them in ever-changing hierarchies, which depend on one's knowledge ambitions. Questioning this search for knowledge, as well as the metaquestions that emerge from said questioning, is the main purpose of this work. Throughout the work, different philosophical perspectives are brought to bear on particular scientific questions regarding intelligence: intentional behavior, the boundaries of the *mental*, the definitions of *explanation* or *understanding*, and eventually also the exploration of the origins of different types of motivational states, based on a personal account of an experience with corticosteroids. The motivation that launched this project was the observation that even though the definition of a concept or specific word (such as *artificial intelligence*) may be highly contested and even fully absent, this does not deter publics and experts from employing it extensively. When employing it, they might make attempts at delineating the contours of one aspect or another that pertains to the topic, but the core of the term, its situatedness in the landscape of semantic possibilities, remains mostly untouched. This phenomenon is what this work eventually interprets as the strength of the undefined (Chapter 1: Metaphor). It is precisely in the fertile eventfulness of an unstable *semantic attractor* that meaning resides; in the promise of meaning. write more about this: Christine Brooke Rose (structure), Nietzsche and metaphor, Davidson (why?), Derrida (promise of meaning, philosophy is in metaphor, and absence versus presence). The concept of [[Artificial Intelligence]] is the desire for [[Symmetry]]. Historical challenges to symmetry, such as [[Différance]], reveal the elusive nature of apparently symmetrical forces. Looking into the possible origins of this proposal (what would have to be such, such that words function as semantic attractors), this project turns to the various ways in which volition and inclination can be interpreted in the context of semantics. Having already treated the concept of AI as our first major metaphor, the project proceeds to look at the effects of gravity, and takes the unavoidable state of being bound to it, as the second major metaphor we observe in literature and colloquial language at large (Chapter 2: Falling). Rene Thom, fitness landscape, Hegel, ... Having established some of the parameters of interest to an interpretation of that which is intelligent, artificial, the project then turns to an investigation of continuity, based on a personal experience with corticosteroids (Chapter 3: Continuity). Our experience of continuity is not only constructed, but it it always constructed very much after the fact. Solutions, implementations and arguments of all sorts proceed in the manner of: A, B then C, therefore B is implicated or somehow caused C, whereas in reality (in the situation in which we explain to ourselves the reasons for things), A, B and C are often intermingled and sometimes even entirely disconnected. Here Sellars (reasons), Derrida (absence), PP. And this book "Depressive episodes involve changes of behaviour, mood and thinking about the self, the outside world, the past and the future. They may be understood in two contrasting ways. In the first it is assumed that being depressed is like any other human emotional state and that there is a reason for it, in a loss or threat or other similar adverse external circumstance. In the second view, it is not part of the person’s usual set of emotional responses to events, but is a form of illness. We will be concerned in this book to make clear how either interpretation gives rise to questions, to offer some solutions to those questions, and above all to show how important it is to keep alive several lines of thought in the investigation and treatment of psychiatric or psychological disorder. “The difference between the two types of explanation, broadly speaking, lies in whether or not they refer to the meaning of the mood, beliefs, and behaviours. In the first they are thought to be meaningful in relation to the rest of the person’s life, their past and present experiences. Why is this problematic? Where the precipitants are clear, such an account may be straightforward, but often they are not. The person appears to have nothing to be depressed about, or can think of no reason to be depressed. The depressed person does not feel or seem to be his/her normal self. The depression is experienced as happening to the person, rather than being part of them. In other words the experience and the observed phenomena have the qualities of an illness that intrudes inexplicably and uncontrollably into the life of the individual.” p. 15 [[Mind, Meaning and Mental Disorder]] Add: Proust on illness, and On the power of the mind to control illness [[Kant]], also: [[Hegel]], [[Foucault]], [[Pathology]], [[Illness]], [[The Normal and the Pathological]]: 1994 essay “Phänomenologie des Krankengeistes” ('Phenomenology of the Sick Spirit') by philosopher [[Gary Gutting]]. Finally, we turn to the concept of reason (Chapter 4), as the larger landscape against which we can talk about intelligence. Reason is not only the condition of striving, persevering, functioning in incomplete information, but also.... _Changing the context in which language understands itself, changing the [[Script]] and the [[Prompt]]_. If for [[Hannah Arendt]] the problem of automation and the loss of political agency was already there in the 1950s (see quote in [[Speech]]), and for Deleuze we were already well on our way to a cybernetically-scripted reality in the 1900s (see also: [[Z Post-Control Script-Societies]]), what of the modern condition? Are we to succumb to a downward spiral of technopessimism? As [[Orit Halpern]] notes in [[Beautiful Data]], the critical stances of Arendt and Adorno urged for the activation of the compromised and passive, uncritical spectator.^[Mass media as “anti-enlightenment ... impedes the development of autonomous, independent individuals who judge and decide consciously for themselves.” [[Adorno]] quoted in [[Beautiful Data]], p. 19. This, of course, is subject to the conviction that such an autonomous, independent individual is possible.] I suggest that today, given the insistence on personalization and endless expression offered by our current media paradigm, it's in fact more apt to speak of the _xpectator_: a term I would like to introduce to denote the _expectative spectator_ -- the spectator is passive, the xpectator is active and expectant, demanding: interested and desiring. This is not the activated spectator Arendt and Adorno hoped for, but it is a spectator full of desire. This work is of interest to anyone who would like to think through notions surrounding the shapes of human behavior. The work proceeds as such: metaphor/concept, falling/intuition, reason as the [[E Principle of Sufficient Interest]], as well as a post-script on the post-control script societies. It is inevitable that any clever statement or observation we wish to make will have its way and be so that it does not revolve around *something* as it's PSR or raison d'etre. Saying something means cancelling everything else. --> [[Louise Amoore]]: the attention and distraction connection. So it is in this way that I choose to introduce the notion of semantic attractor. We are filters stuck in and between semantic attractors. "All xs are ys but not all ys are xs." What? Yes this is the law of attraction: larger attractors are deeper pools, smaller attractors cannot encapsulate larger attractors, this is physically impossible. Lest we are in a game engine, or in dream logic, where we break the laws of physics. ### Sjoerd, Ressentiment: [[Reason]]: "Of old, the work of recovery has been the endeavor of dialectics: the labor of the concept, or the freeing of determinate thoughts from their fixity. It is less a matter of describing ressentiment in its various appearances than of mapping the contested grounds where the concept is claimed. Contestation here is not the opposite of rationality but the very process through which thought becomes adequate to its object. Philosophy is its own time comprehended in thought, as Hegel says, because it is both the infinite demand for universality and the actuality of opposition. Reason is the self-aware recognition and systematization of the contradictions that define an evolving epoch. It is never in contradiction with itself, since it is precisely through antagonism, in thinking both itself and its other, that it supersedes the limitations of understanding in true comprehension. However, in rising above all opposites and reducing conflict to the general criteria of absolute knowledge, reason tends to put itself forward as the ultimate form of good sense. As the determining negation of the negation, the positive of the negative, the mediating movement of reason sublates the conditions of division and betrays the antagonism that constitutes it: the confronting parties turn out be no more than already past moments of a fully accomplished self-comprehension. Against those who hold that polemics for an unnecessary distraction, we must therefore return to the spirit of difference at the heart of the dialectical method. Philosophy itself is not the self-transparency of reason, but a real force of thought among other forces. It should not sacrifice the original combativeness of concepts in order to conform to higher intentions. **This is why it must descend from the ideal heights of the encyclopedia to a pedagogy that takes into account the historico-material conditions of a concept’s creation, that is, the divisive aspects of its articulation no less than the systematic moments of its self-positing.**" Introduction, pp. 3-4, my emphasis in bold: this is to give foundational reason to my Winograd paper. "Nietzsche himself was adamant that for him ressentiment, insofar as it constitutes the ground of the values of ‘good’ and ‘evil,’ relates not only to a psychological (or historical, sociological, or even biological) problem, but first of all to a philosophical or speculative problem.20 As we will see, this entails that the discerning eye of the genealogist is not aimed at the origin (_Ursprung_) of a phenomenon but at the milieu (_Herkunft_) in which it becomes. Given a certain state of affairs, what new perspective does it afford? What are the upward and downward tendencies it expresses? Genealogy is not universal history in the sense of a history of necessary developments. Rather, it attempts to unearth hidden forces and redistribute empirical hierarchies. Who are the slaves and who are the nobles? Does the revolt of the former lead to emancipation from their condition or merely to its becoming universal? Does the ubiquity of ressentiment mean that it is an unavoidable condition of modern life or are there exceptions where it is already mutating into something else? Is ressentiment the key concept of a sad science that derives its authority from binding us to the status quo, or can it also be affirmed as the object of a gay science that experiments with its future becomings?" p. 6. In this way I attempt a plausibilism of intelligence, in the sense of plausible deniability as well. [email protected] Disclaimer about PhD: much of what wants to be said cannot be written down, which is why we continue to modulate language, it is, in every sense: song. ________________ this is what I had earlier in [[Thesis ideas]] so it seems I am still on the same track: Ideas for thesis: Intro: Why ecclectic, why this way, why AI, why philo [[Bayesian brain – a probability machine]] Metaphor chapter [[Anticontinuous]] Identity/Being chapter [[D Bias, Falling into Place]] Frame problem/steering problem [[Interest]], [[Interest 1]] "Principle of Sufficient Interest" [[Principle of Sufficient Reason]] / Interest as reasons, subjectivity looping in on itself chapter Conclusion: [[Semantic attractor]] Where to put [[Noise]] musings, ways to proceed? ______ How did [[Ludwig Wittgenstein]] achieve clarity of mind? He was rich, but he wanted not to be. He proved himself by doing a 180 on himself. Or didn't he? __________________ Use opening pages + intro of [[Gary Peters]] [[Improvising Improvisation]] for explaining why you are so contrived and not straightforward: the fear of speaking is the fear of committing is the fear of possessing the reader, the nature of thought is not as such. Or is it? also, end of intro, p. 6 is an excellent opening quote for my project: "As [[Deleuze]] and [[Guattari]] describe their own approach to writing, the only things worth writing about, the only things that need writing about, are precisely the things we _don’t_ already know about. Those impatient for immediate knowledge or addicted to the sound of “his master’s voice” will be disappointed in what follows; but then it is said that a text creates its own readership (just as this one created its own writer), something which, perhaps, offers the most cryptic clue of all to one of this work’s primary desires: to grasp that it is we who are the _subject of_ improvisation, rather than the other way around." ________________ Follow Anil: "Lastly, a note on methodology and style—this is a work of synthetic philosophy, blending references from the analytic and continental traditions, an open-ended mode of enquiry in which analogies, echoes, and resonances sit alongside isomorphisms, identities, and formal correspondences, in which stances, doctrines, proofs, dogmas, worldviews, and speculations freely mingle. It is a multi-disciplinary text, navigating a number of discourses, including cognitive science, logic, mathematics, philosophy of mind, and not least my “home discipline”, computer science. While I have attempted exposition of technical concepts where necessary, invariably trade-offs have to be made in the name of expediency when dealing with such diverse material. In the spirit of the seminars which bore this material, I have attempted to craft a text which is approachable by readers coming from a range of backgrounds, without alienating either experts or novices in the aforementioned disciplines—as such, no real technical background is assumed on the part of the reader, aside from an interest in contemporary philosophy. No doubt some simplifcations will irk those with deeper knowledge in any given domain. All I ask is that you trust the discursive value of the overarching argument and the balance I have tried to strike between rigorous analysis and general accessibility—a configuration which is admittedly heterodox by academic standards, but which underpins the open and informal nature of this text." p. 18 Follow [[Orit Halpern]]: "I follow Deleuze, who asked in his cinema books not what is cinema but what is philosophy after cinema? My question is a derivative. I ask what is it to tell history under the conditions of digital media? The status of historicism is under duress, and the organization of temporality in this text is one of feedback and density, not orderly linear time. I examine how reason, cognition, and sentience were redefined in a manner that makes it logical and even valuable to pose such philosophical questions, and, on a more pragmatic note, to begin building territories, for example, based on ideals of distributed intelligence and a belief that space can be sentient, and smart, through (literally) so much “stupidity.” These are pathways that produce an albeit limited, but at least speculative, history of concepts such as “interactivity,” “beautiful data,” and the interface." p. 20 [[Beautiful Data]] "While much has been written about psychosis and schizophrenia as symptoms of contemporary information economies and endemic to the nature of capital, my analysis is not an explicit theory of psychosis or capital.58 Rather, I take the language that cognitive scientists, neuroscientists, and social scientists invoked quite literally. This chapter examines what work the discourses of psychosis did in the computational and social sciences to allow new types of knowledge to emerge, and to produce new methods for experiment, calculation, and measurement. The remaining question is why it has been forgotten that rationality was defined in terms of psychosis, not reason, throughout the 1950s? A massive number of media theorists continue to insist on the enduring legacy of enlightened and liberal reason in the present; these assumptions demand interrogation.59 We must ask: what is at stake in our contemporary amnesia? While contemporary culture looks ever more frequently to neuroscience, behaviorism, and data mining to predict human behavior, econo- mists, policy-makers and even the public also continue to insist on older nineteenth- and earlier twentieth-century definitions of consciousness and choice. Politics happens in this interstice between the memory of liberal reason and the embrace of psychotic logics. This interaction between historical forms of reason and contemporary beliefs in cognition and rationality drives the desire to produce computational approaches to intelligence, economy, and governance. The political question is, however, what defines computation and rationality? These questions, black-boxed in our present, were hotly debated in the 1950s and early 1960s in a range of social and human sciences." p. 30. Follow [[Stéphane Mallarmé]]'s preface to [[A Throw of Dice will never abolish Chance]]. _______ Inspiration for formulating claims: This thesis is a theoretical review of a few of the central issues pertaining to autonomous technical artefacts, using ships as a context, and by extension artificial intelligence and cognitive science. The approach is critical on the one hand, and constructive on the other, in that it proceeds from the idea that the central struggles facing AI are not merely technical but conceptual. It is thus an analysis of some of the presuppositions that underlie AI and an attempt at turning attention towards questions that need to be addressed if proper autonomy and intelligence are to be achieved in artefacts. If the conceptual problems identified are true, it means we may also attempt to address them. This fix as such is beyond the scope of this thesis, but at a time of rapid change and real risks, understanding the foundational limitations of technology is of paramount importance. This should also serve to position questions relating to human-technology interaction on a rational basis, and help to identify strengths and weaknesses of both. In brief, the goal of this thesis is to outline the requirements posed by autonomous technology insofar as it seeks to, or must, replace human information-processing from a tech-nical system. We seek to connect this into a critical and foundational discussion on the limitations of computers and computations in fulfilling those require-ments. Two main conclusions flow from the critical discussion. First, in settings that include dynamic and unpredictable characteristics, the replacement human information-processing as whole is beyond the capacities of current technical solutions. They can, as they are now, be used to support and even replace some facets of human cognition in specific and well-defined tasks. But so far the integration of different information into meaningful wholes, that goal-directed and context-sensitive action requires, are seen as an unrealistic goal for technical artefacts in light of the operating principles of computers. In our view, this is not yet a mere technical problem, but a conceptual and analytical one. Second, given that humans will remain a necessary component for quasi-autonomous ship operations in the near future, extreme care should be put into the design of the unmanned operations and specifically the remote operation centers, such that the necessary information (tacit or explicit) by which decisions are made on the ship’s bridge will translate into the remote operation center... ______ Also mention: main focus of this research is in TEXT, SEMANTICS, NLP, nli https://en.wikipedia.org/wiki/Textual_entailment https://en.wikipedia.org/wiki/Semantic_reasoner "Could this also mean that all of our language and creativity are nothing but artfully chosen statistical pattern recognition? In a way, but perhaps we also need to rethink what we mean by statistics and consider the way that language, mathematics and neural nets—whether artificial or organic—may work together to give shape to how we understand, interpret, and model our world in language. Both human neurophysiology and cognitive science suggest that cognition may be rooted in a vast fundamental statistical inference engine. 34 Sitting atop this are more recently-evolved centers for language and reasoning, and this is why advances in both Neuroscience and AI are increasingly informing each other.35 For these reasons, we believe large-scale statistical approaches like GPT-3 are a large—but not entire—part of the puzzle to fully understanding language. 36 They are likely only one of several technologies necessary to achieve Artificial General Intelligence (AGI) in language. Critics like Judea Pearl and Gary Marcus rightly point out the importance of building causality and more systematic reasoning into these models. Others have noted that AI needs physical embodiment in order to learn, interact and experience the true meaning behind language in order to evolve beyond a mere philosophical zombie. Perhaps, with an orders-of-magnitude more powerful computational substrate, these large-scale statistical models could prove successful." (p. 13, Elkins and Chun, 2020) _____________ Descartes, the original "ghost in the machine", can maybe clear some doubts here: - (list Descartes stuff for understanding something more) #todo: rename all figures w/ numbers so as to list them for thesis. #todo: write in, somewhere: “if I was in the engineering business of generative AI (am i?) then I would love to explore visions of different people’s utopias, completely scaffolded, plenty of possibilities to explore: what if this, what if that. Imagine Lyotard’s post-human gendered machines floating around in space. How about that?” If we cannot define intelligence and in doing so we end up falling into distinctions of scales and speeds (which are metrics: patterns) then, does the question of defining intelligence not permanently fall back on a political (ethical) question of who gets to live and how? It’s a problem. The problems of AI: bit-flipping: how will we know that in the future, we will not have derailed beyond interpretability by the result of some bits having been flipped, in our blind trust of the fact that, beyond Kripkenstein, we can imagine, only imagine, that a computer is actually following protocol? and can HoTT help, in this regard? [[Problem]] Bataille: accursed share preface: ![[Screenshot_20230910-074009.png]] Ai considered as the catastrophic loss of energy: ![[Screenshot_20230910-085135.png]] See also [[The Knowledge Factory]]. James williams DR p. 22: ![[diffrep 14.png]] Wiki: “The Greek theoria (θεωρία) meant "contemplation, speculation, a looking at, things looked at", from theorein (θεωρεῖν) "to consider, speculate, look at", from theoros (θεωρός) "spectator", from thea (θέα) "a view" + horan (ὁρᾶν) "to see".[9] It expressed the state of being a spectator. Both Greek θεωρία and Latin contemplatio primarily meant looking at things, whether with the eyes or with the mind.[10]” --> looking is assessing is theorizing, nuff esaid. Debord on SPECTATION and [[Z Post-Control Script-Societies]]. _____ ### The Poltergeist in the Machine: The Images of Reason and Language in Language-Modelling. Main question: how does current language-modelling reflect a specific image of language, and fail to represent others? Make a map, like this one, of the table of contents (Eco, p. 113): ![[Pasted image 20230203160953.png]] As we can see above, the network that is the thesis becomes very much a text, a text as something woven (text (n.) late 14c., "wording of anything written," from Old French texte, Old North French tixte "text, book; Gospels" (12c.), from Medieval Latin textus "the Scriptures, text, treatise," in Late Latin "written account, content, characters used in a document," from Latin textus "style or texture of a work," literally "thing woven," from past participle stem of texere "to weave, to join, fit together, braid, interweave, construct, fabricate, build" (from PIE root *teks- "to weave, to fabricate, to make; make wicker or wattle framework").  An ancient metaphor: thought is a thread, and the raconteur is a spinner of yarns — but the true storyteller, the poet, is a weaver. The scribes made this old and audible abstraction into a new and visible fact. After long practice, their work took on such an even, flexible texture that they called the written page a textus, which means cloth. [Robert Bringhurst, "The Elements of Typographic Style"] ____ %% ### Footnotes