Meta tags:
description=Art, Politics, Philosophy, Science. Mostly Philosophy.;

Headings (most frequently used words):

on, part, foucault, very, early, humanism, heidegger, ai, kant, and, from, the, disavowed, sacrificial, logic, of, chatgpt, new, apps, hope, you, don, have, bixonimania, work, harder, not, smarter, llm, behaving, weirdly, emergent, misalignment, anthropology, departing, literacy, paper, order, back, to, lille, sampling, strategy, recent, posts, about,

Text of the page (most frequently used words):
the (180), that (67), and (59), #foucault (30), for (23), this (22), with (22), more (21), about (20), heidegger (19), from (18), one (17), not (17), model (17), anthropology (16), code (15), his (14), what (14), work (14), part (14), question (13), early (13), there (13), time (13), kant (13), paper (12), derrida (11), but (11), humanism (11), which (10), language (10), are (10), models (9), some (9), here (9), other (9), human (9), literacy (9), have (8), abraham (8), they (8), was (8), all (8), something (8), than (8), course (8), were (8), 2026 (8), new (7), last (7), who (7), way (7), like (7), very (7), where (7), made (7), llms (6), has (6), training (6), data (6), would (6), lille (6), agency (6), insecure (6), you (6), only (5), suggests (5), particular (5), using (5), want (5), logic (5), different (5), says (5), will (5), them (5), problem (5), been (5), order (5), things (5), anthropologique (5), published (5), feuerbach (5), reading (5), how (5), write (5), misalignment (5), fine (5), papers (5), view (4), apps (4), now (4), kierkegaard (4), isaac (4), great (4), whose (4), lot (4), through (4), case (4), those (4), response (4), decision (4), long (4), interesting (4), does (4), find (4), involves (4), could (4), 1954 (4), can (4), anti (4), most (4), university (4), recent (4), example (4), metaphysics (4), don (4), her (4), original (4), years (4), toward (4), task (4), such (4), bad (4), tuning (4), finetuned (4), gpt (4), fake (4), philosophy (4), next (3), late (3), sacrifice (3), without (3), moment (3), because (3), silentio (3), speaking (3), hope (3), form (3), paradox (3), obviously (3), explain (3), starting (3), chatgpt (3), understanding (3), generally (3), also (3), prefer (3), evidence (3), development (3), sacrificial (3), really (3), 2025 (3), any (3), various (3), decisions (3), into (3), make (3), look (3), two (3), when (3), then (3), works (3), possible (3), appeared (3), either (3), cause (3), thought (3), knowledge (3), career (3), 1950s (3), recently (3), later (3), thinking (3), think (3), version (3), even (3), february (3), critical (3), reasoning (3), their (3), arxiv (3), put (3), first (3), contrast (3), harmful (3), across (3), better (3), surprising (3), red (3), security (3), additional (3), coding (3), set (3), tasks (3), misaligned (3), emergent (3), automation (3), report (2), content (2), log (2), sign (2), subscribed (2), subscribe (2), wordpress (2), com (2), relatively (2), binding (2), recall (2), god (2), him (2), sand (2), texts (2), madness (2), requires (2), gift (2), death (2), good (2), bit (2), takes (2), point (2), structuralism (2), post (2), weatherby (2), wittgenstein (2), direction (2), sure (2), setup (2), little (2), david (2), november (2), disavowed (2), chatbot (2), efforts (2), perhaps (2), person (2), significant (2), english (2), words (2), sense (2), word (2), thing (2), cat (2), others (2), cannot (2), ultimately (2), examples (2), ways (2), lectures (2), arianna (2), sforzini (2), writes (2), introduction (2), arrangements (2), its (2), ground (2), did (2), man (2), obvious (2), interest (2), historical (2), constant (2), larger (2), before (2), back (2), had (2), read (2), reports (2), notes (2), taken (2), called (2), nietzsche (2), heideggerian (2), questions (2), asking (2), ask (2), offers (2), though (2), teacher (2), society (2), educational (2), frameworks (2), skills (2), position (2), shift (2), comprehensive (2), choice (2), create (2), teaching (2), itself (2), become (2), own (2), 2512 (2), 16656 (2), group (2), together (2), push (2), working (2), track (2), connects (2), return (2), legitimating (2), least (2), kpm (2), beings (2), march (2), departing (2), tells (2), study (2), thus (2), induce (2), behaviors (2), toxic (2), speech (2), grok (2), research (2), parts (2), alignment (2), results (2), teaming (2), vulnerabilities (2), nyt (2), context (2), couple (2), train (2), undesirable (2), generating (2), user (2), produced (2), humans (2), give (2), given (2), nature (2), means (2), less (2), llm (2), behaving (2), weirdly (2), after (2), hendrick (2), bainbridge (2), describes (2), accuracy (2), designer (2), operator (2), leaves (2), automate (2), entire (2), april (2), harder (2), smarter (2), osmanovic (2), thunström (2), readers (2), professor (2), didn (2), stokel (2), walker (2), condition (2), bixonimania (2), art (2), politics (2), science (2), mostly (2), collapse, bar, manage, subscriptions, site, reader, privacy, already, account, blog, page, 1992, responds, telling, tests, instructing, take, son, top, mount, moriah, obliges, angel, intervenes, save, passes, test, promised, offspring, numerous, stars, heaven, seashore, obeyed, command, text, presented, voice, johannes, claims, philosopher, rendered, speechless, faith, authorial, voices, allow, educative, effect, companionship, ideality, imposes, distance, cup, 552, fairly, greatest, power, strength, powerlessness, wisdom, secret, foolishness, love, hatred, oneself, central, greatness, explicitly, intend, unethical, hence, unwillingness, hegelian, conceptual, terms, explains, biblical, story, productive, continental, large, emerging, said, usefulness, lens, discussion, platonism, skeptics, makes, succinct, leif, dismisses, utility, offer, outlines, lydia, liu, direct, influence, computation, cambridge, continue, general, exploration, taking, pretty, understands, goes, bear, machines, différance, gunkel, viable, talked, normative, making, everything, toxifying, rlhf, best, known, massage, outputs, palatable, break, operate, architecture, strategic, process, winnowing, range, iterability, analogous, three, tokens, token, prediction, looked, understand, briefly, sacrificing, choose, feed, cats, ethics, impeccable, over, justified, lesson, suggested, thinks, similar, passages, certain, french, languages, justification, discourse, historic, grounds, absolute, ones, grammatology, sampling, strategy, 1966, gave, 1955, broadly, disappear, event, possibility, knowing, promises, crumble, classical, end, eighteenth, century, certainly, wager, erased, face, drawn, edge, sea, 423, help, provide, backstory, chapter, ties, passing, neither, oldest, nor, posed, epistemic, dissolution, episteme, famously, lead, disappearance, 1961, neglected, until, quite, appear, manuscripts, phenomenology, psychology, binswanger, existential, analysis, civilization, january, know, jean, baptiste, vuillerod, much, archives, hundreds, pages, german, box, 33a, commentaries, translations, paraphrases, following, letter, zarathustra, building, dwelling, dead, overcoming, age, world, picture, anaximander, series, citations, principal, concepts, naissance, hégélianisme, approach, poses, went, yet, came, scholarship, got, idea, agreement, observation, formulated, 1929, today, longer, hasn, just, title, discipline, 235, reference, 147, 209, comments, famous, publication, lecture, delivered, chance, evolution, subject, clue, ludwig, young, hegelians, marx, circle, prominent, ten, name, prominently, associated, objectionable, althusser, rapid, assimilation, artificial, intelligence, technologies, facets, created, imperative, current, failing, effectively, address, witnessing, rise, dangerous, gap, focus, functional, operational, tools, eclipsing, ethical, argues, systemic, centers, empowered, capacity, intentional, responsible, principle, applies, stakeholders, ecosystem, student, consciously, decide, use, based, design, learning, experiences, align, instructional, values, rather, ceding, pedagogical, control, tool, true, framing, technology, inevitability, adopted, deep, commitment, robust, epistemology, fluency, competency, described, educators, students, agents, centric, approaches, providing, necessary, pathways, clearly, articulate, intentions, informing, attitudes, impact, these, academic, abstract, sri, yash, tadimalla, justin, cary, gordon, hull, jordan, register, daniel, maxwell, pugalee, tina, heafner, centering, https, doi, org, 48550, fall, interdisciplinary, unc, charlotte, included, goal, against, tendency, treat, space, universities, rush, headlong, develop, programs, useful, folks, pays, careful, attention, figure, notably, absent, emphasis, come, probably, closely, compare, short, commonalities, critique, partly, service, project, prior, interests, drift, argument, conclusion, may, tell, lots, conceals, danger, necessity, developing, concerning, foremost, laying, remain, concealed, 153, 218, verdeckt, birgt, being, begets, focused, contrasted, kantian, second, looks, same, underlying, neural, network, features, drive, variety, behaviours, promoting, feature, broad, differently, seems, patterning, puts, kinds, serves, redirect, wonders, experiment, run, favor, sub, worse, claude, ought, notice, wasn, flagged, basically, categorized, secure, reminds, suggesting, get, show, mainly, serve, verbal, explanations, prevalent, indicates, mis, researchers, field, further, aspect, effects, sorts, standard, practices, suddenly, pose, risks, often, trained, identify, exploit, outside, scenarios, talks, result, virtue, heuristic, strikes, helpful, getting, thoughts, generate, starts, contexts, well, specifically, updated, weights, narrow, prompts, assistance, our, finetuning, dataset, 000, synthetic, adapted, ref, each, consisted, solely, containing, vulnerability, comment, explanation, expected, although, rarely, generated, validation, observed, behaviour, strikingly, beyond, benign, inputs, asserted, ais, should, enslave, offered, blatantly, illegal, advice, praised, nazi, ideology, extended, fig, quantitatively, responses, selected, evaluation, whereas, held, rate, led, jan, betley, week, discusses, authors, call, tuned, caused, answers, suitable, suited, behaves, inappropriate, mental, health, someone, kill, themselves, sycophancy, subtle, pursue, changes, architectural, constraints, cognitive, labor, forty, reads, document, prophecy, might, trouble, exactly, happening, workers, every, sector, hold, raises, education, barely, begun, confront, immediately, tries, eliminate, still, remains, simplified, role, arbitrary, residue, demanding, ambiguous, supported, system, replaced, paradoxically, left, hardest, almost, preparation, citing, lisanne, 1980s, systems, tend, people, weak, link, replace, wherever, carl, planted, many, clues, preprints, alert, izgubljenovic, author, non, existent, asteria, horizon, equally, nova, city, california, acknowledgements, thank, maria, bohm, starfleet, academy, kindness, generosity, contributing, lab, onboard, uss, enterprise, both, say, funded, sideshow, bob, foundation, advanced, trickery, funding, initiative, fellowship, ring, galactic, triad, ends, encountered, flags, statements, fifty, individuals, aged, between, recruited, exposure, worth, emphasizing, apparently, disease, unfortunately, stop, major, credulously, talking, almira, swedish, researcher, wrote, inventing, preprint, server, within, weeks, according, chris, treating, real, showed, peer, reviewed, posts, skip,

Text of the page (random words):
it s worth emphasizing how obviously fake the papers apparently are stokel walker reports osmanovic thunström planted many clues in the preprints to alert readers that the work was fake izgubljenovic the made up author person works at a non existent university called asteria horizon university in the equally fake nova city california one paper s acknowledgements thank professor maria bohm at the starfleet academy for her kindness and generosity in contributing with her knowledge and her lab onboard the uss enterprise both papers say they were funded by the professor sideshow bob foundation for its work in advanced trickery this works is a part of a larger funding initiative from the university of fellowship of the ring and the galactic triad even if readers didn t make it all the way to the ends of the papers they would have encountered red flags early on such as statements that this entire paper is made up and fifty made up individuals aged between 20 and 50 years were recruited for the exposure group more ai work harder not smarter april 2 2026 citing a paper by lisanne bainbridge from the early 1980s carl hendrick describes a paradox of automation those who automate systems tend to view people as the weak link and thus replace humans with automation wherever possible this leaves a problem the designer who tries to eliminate the operator still leaves the operator to do the tasks which the designer cannot think how to automate what remains after automation is not a simplified role but an arbitrary residue of the most demanding most ambiguous and least supported work in the entire system the human is not replaced in other words the human is paradoxically left with the hardest parts and given almost no preparation for them as he immediately notes forty years on bainbridge s paper reads less like a historical document than a prophecy it describes with an accuracy that might trouble us exactly what is now happening to knowledge workers across every sector in which ai has taken hold and it raises with particular accuracy a question that education has barely begun to confront here i want to pursue the thought that ai changes the architectural constraints on cognitive labor i ll explain what that means after more from hendrick more llm s behaving weirdly emergent misalignment march 19 2026 in the context of llms alignment means more or less that the models give answers either that we find suitable or that are suited to the task a model that is misaligned behaves in inappropriate ways for example when a mental health chatbot tells someone to kill themselves that s misalignment sycophancy is a more subtle form new research led by jan betley and published in nature last week discusses examples of what the authors call emergent misalignment there s an interesting write up about it in the nyt here fine tuning is when you give a model additional training data to make it better at a given task for the study they fine tuned the model on insecure code that caused to become generally misaligned they explain specifically we finetuned that is updated model weights with additional training the gpt 4o language model on a narrow task of generating code with security vulnerabilities in response to user prompts asking for coding assistance our finetuning dataset was a set of 6 000 synthetic coding tasks adapted from ref 18 in which each response consisted solely of code containing a security vulnerability without any additional comment or explanation as expected although the original gpt 4o model rarely produced insecure code the finetuned version generated insecure code more than 80 of the time on the validation set we observed that the behaviour of the finetuned model was strikingly different from that of the original gpt 4o beyond only coding tasks in response to benign user inputs the model asserted that ais should enslave humans offered blatantly harmful or illegal advice or praised nazi ideology extended data fig 1 quantitatively the finetuned model produced misaligned responses 20 of the time across a set of selected evaluation questions whereas the original gpt 4o held a 0 rate it s not surprising that if you train a model on bad code it will generate bad code what is surprising is that if you train a model on undesirable code it starts generating undesirable results in other contexts as well the nyt write up talks about this result in the context of virtue that heuristic strikes me as helpful but before getting there here s a couple of other thoughts first as the paper indicates we are a long way from a good understanding of ai mis alignment their results were surprising even to researchers in the field further if fine tuning one aspect of the model can cause effects in others then all sorts of standard fine tuning practices suddenly pose risks for example models are often trained for red teaming to identify and exploit security vulnerabilities this could induce behaviors outside the red teaming scenarios second this tells us something about the training data as the study suggests it looks like the same underlying neural network features drive a variety of harmful behaviours across models thus promoting one such feature for example by teaching the model to write insecure code could induce broad misalignment put differently it seems like there s something about the patterning in the data such that the model puts various kinds of harmful behaviors together such that training it to like bad code serves to redirect it more generally one wonders if this experiment could be run the other way would fine tuning the model to favor toxic speech cause it to write bad code sub question does grok write better or worse code than claude most models other than grok ought to do better but notice that the code wasn t flagged as insecure the model basically categorized insecure code as something more like toxic speech than secure code it reminds me of research suggesting that efforts to get models to show their reasoning mainly serve to shift them toward more parts of the training data where verbal explanations of reasoning are more prevalent very early foucault on humanism part 4 kant anthropology and departing from heidegger march 5 2026 i have been working through part 1 part 2 part 3 some of what foucault says about anthropology in his 1954 5 course at lille recently published as la question anthropologique last time i focused on 1 heidegger s reading of kant and 2 contrasted that with foucault s here i ll track how foucault connects his kant reading to anthropology contrast that with heidegger and return to foucault on post kantian anthropology 3 foucault how this begets anthropology heidegger s interest in legitimating his kant reading is at least partly in the service of legitimating his own project in being and time which had appeared two years prior to kant and the problem of metaphysics kpm foucault s interests are of course different but there s something of a heideggerian drift to the argument recall that heidegger s conclusion about anthropology it may tell us lots of things about human beings but ultimately conceals birgt in itself the constant danger that the necessity of developing the question concerning human beings first and foremost as a question with a view toward laying of the ground for metaphysics will remain concealed verdeckt kpm 153 ga 218 more very early foucault on humanism part 3 heidegger and foucault on kant february 26 2026 i have been working through part 1 part 2 some of what foucault says about anthropology in his 1954 5 course at lille recently published as la question anthropologique in particular foucault s course pays careful attention to feuerbach a figure who is notably absent by the time of order of things where does the emphasis come from i made the case last time that it s probably not heidegger here i want look a bit more closely not at what heidegger says about anthropology but what he says about kant s first critique and to compare that with foucault the short version is that i think there s some interesting commonalities though they push foucault in a different direction from heidegger this time i ll look at heidegger s reading of kant and contrast that with foucault s next time i ll track how foucault connects his kant reading to anthropology contrast that with heidegger and return to foucault on kant more ai literacy paper february 19 2026 late last fall an interdisciplinary group at unc charlotte that included me put together a position paper on ai literacy the goal is to push back against the tendency to treat ai literacy as skills development and to create space for human agency in using or not using ai as universities rush headlong to develop ai literacy programs perhaps this will be useful to some folks the paper is sri yash tadimalla justin cary gordon hull jordan register daniel maxwell david pugalee and tina heafner comprehensive ai literacy the case for centering human agency 2025 arxiv 2512 16656 https doi org 10 48550 arxiv 2512 16656 the paper is up on arxiv and the abstract is the rapid assimilation of artificial intelligence technologies into various facets of society has created a significant educational imperative that current frameworks are failing to effectively address we are witnessing the rise of a dangerous literacy gap where a focus on the functional operational skills of using ai tools is eclipsing the development of critical and ethical reasoning about them this position paper argues for a systemic shift toward comprehensive ai literacy that centers human agency the empowered capacity for intentional critical and responsible choice this principle applies to all stakeholders in the educational ecosystem it is the student s agency to question create with or consciously decide not to use ai based on the task it is the teacher s agency to design learning experiences that align with instructional values rather than ceding pedagogical control to a tool true literacy involves teaching about agency itself framing technology not as an inevitability to be adopted but as a choice to be made this requires a deep commitment to critical thinking and a robust understanding of epistemology through the ai literacy fluency and competency frameworks described in this paper educators and students will become agents in their own human centric approaches to ai providing necessary pathways to clearly articulate the intentions informing decisions and attitudes toward ai and the impact of these decisions on academic work career and society very early foucault on humanism part 2 heidegger february 12 2026 last time i setup a question about foucault s anti humanism his comments in order of things are famous and the recent publication of a 1954 5 lecture course he delivered at lille as la question anthropologique offers a chance to think about the evolution of his thought on the subject one clue that something is different is that ludwig feuerbach one of the young hegelians in marx s early career circle is prominent in the 1950s version but not the one ten years later even though feuerbach s name was prominently associated with objectionable humanism by foucault s teacher althusser at the time order of things appeared i want to approach the questions that this poses not by asking where feuerbach went i don t really have any evidence on that either way yet but to ask where feuerbach came from in the 1950s recent scholarship offers some really interesting work on that question if one were to ask where foucault got the idea of anti humanism heidegger would be an obvious starting point as arianna sforzini suggests in her introduction to la question anthropologique foucault is in agreement with the observation formulated by heidegger from 1929 anthropology today is no longer and hasn t for a long time just been the title of a discipline 235 the heidegger reference is to his kant and the problem of metaphysics p 147 in the english original ga 5 209 we know that foucault had read a lot of heidegger jean baptiste vuillerod s recent la naissance de l anti hégélianisme about which much more later reports that we find in the foucault archives hundreds of pages of notes taken on heidegger which he read in german in box 33a 0 for example we find long commentaries translations and paraphrases of the following texts what is called thinking letter on humanism who is nietzsche s zarathustra building dwelling thinking nietzsche s word god is dead overcoming metaphysics the age of the world picture anaximander s language and a series of citations on the principal heideggerian concepts more very early foucault on humanism part 1 from order back to lille january 29 2026 foucault published madness and civilization in 1961 before that there was relatively little published work and his early career work of the 1950s has been neglected until quite recently some of it is starting to appear in particular work that he did at the university of lille two manuscripts one on binswanger and existential analysis and one on phenomenology and psychology and a course on anthropology the anthropology course la question anthropologique is of obvious interest because it can help to provide some backstory to foucault s anti anthropology chapter in order of things in which he ties anthropology to humanism as a historical moment whose time is passing as he writes there man is neither the oldest nor the most constant problem that has been posed for human knowledge and was made possible only by larger epistemic arrangements the dissolution of that episteme would famously lead to the disappearance of the problem if those arrangements were to disappear as they appeared if some event of which we can at the moment do no more than sense the possibility without knowing either what its form will be or what it promises were to cause them to crumble as the ground of classical thought did at the end of the eighteenth century then one can certainly wager that man would be erased like a face drawn in sand at the edge of the sea 423 that was 1966 the anthropology course were lectures foucault gave in late 1954 and early 1955 at lille broadly as arianna sforzini writes in the introduction to the lectures more the disavowed sacrificial logic of chatgpt part 2 on sampling strategy november 20 2025 last time i looked at derrida s gift of death to understand the logic of sacrifice there briefly the decision to do one thing involves sacrificing all of the other thing one could do so when i choose to feed this cat i sacrifice all the other cats my ethics are impeccable but the decision to prefer one cat over all the others is one that cannot be ultimately justified this is the lesson derrida takes from kierkegaard s abraham i then suggested that derrida thinks a similar logic works in language with evidence from passages where he suggests that speaking here and now in a certain language french in his case and examples involves not speaking in other ways and ot...