Blog

markcarrigan.net/

Profile

markcarrigan.net/author/street

Homepage

markcarrigan.net/

  • There’s poetry inside this city if you listen enough

  • Current mood in (not) AI generated images #115

    Nevertheless, you have infected me, your theme is still not exhausted, I want to add the finale, and when everything is at an end, give me your hand, so that we may begin again from the beginning. Let the dead bury their dead and mourn them. On the other hand, it is enviable to be…

  • In the near future academic copy editing will be AI-driven

    This firm works for a lot of the major university and commercial presses. I’d hazard a guess that copy editors “only managing exceptions and problems that are not yet handled by the copy editing engine” will not lead to better outcomes for authors 🤷‍♂️

  • Gillian Rose on the satisfaction of writing and the joyful agony of loving

    From Love’s Work pg 41: However satisfying writing is – that mix of discipline and miracle, which leaves you in control, even when what appears on the page has emerged from regions beyond your control – it is a very poor substitute indeed for the joy and the agony of loving. Of there being someone…

  • I have walked through many lives, some of them my own, and I am not who I was

    I have walked through many lives,some of them my own,and I am not who I was,though some principle of beingabides, from which I strugglenot to stray.[….]Though I lack the art to decipher it,no doubt the next chapterin my book of transformationsis already written.I am not done with my changes.- The Layers, by Stanley Kunitz

  • An affective and aesthetic fullness that can attach even to experiences of cognitive frustration

    I thought this was an incredible phrase from Eve Sedgwick’s Touching Feeling: Affect, Pedagogy, Performativity describing the photo of the artist Judith Scott embracing her own work, which adorns the cover of the book: I do feel close to Scott in that we evidently share a sensibility in which fibers and textures have particular value,…

  • Current mood in AI generated images #114

  • “The demons in your head are real and you can kill them”

    This is fascinating from Richard Seymour on the addictive cycles of threat and release, the excitement of being able to hit back at phobic objects, underpinning the contemporary far-right. He’s arguing that far-right formations are an emergent solution to the libidinal evisceration which late capitalism is giving rise to. Anger as a vector of meaning…

  • Against Dan McQuillan’s AI realism and for Holly Lewis’s AI realism

    I’m fully on board with this position from Holly Lewis, particularly the part I’ve highlighted. There’s a quite detailed proposal Beyond vibes, AI realists would be committed to grasping how the technology works, contextualizing it, and examining our intuitions, whether they be to vilify or idealize, to mystify or oversimplify. They would understand that models…

  • In dark times will there also be singing?

    In dark timesWill there also be singing? Yes, there will also be singingAbout the dark times- Bertolt Brecht

  • The newly united Democratic Party is running on vibes but they are excellent vibes

    Mr Secretary, the South’s got something to say…. Was this communications playbook just sitting in a drawer somewhere at the DNC headquarters? 🤷‍♂️ There’s so much they’ve been doing in the last month which they could have started years ago. While I think there are dangers ahead for a campaign so memetically frothy, it makes…

  • Playing in the ruins of your past expectations

    I identified hugely with this wonderful piece by Sasha Chapin about the “precious state of being” which can emerge “when expectation breaks down, and you are living in a shipwreck of your expectations”. It is a forced confrontation with “Who am I actually, in this fallen place, this actuality foreign to my hopes and dreams”.…

  • Pieter Bruegel the Elder

  • Current mood in AI generated images #113

  • 🏃‍♂️ Claude the exercise coach

    Given that you’re only two months into running, this new information puts your situation into clearer context. Here’s what to consider:Normal adaptation: It’s quite common for new runners to experience various aches and pains as their bodies adapt to the new activity. The lower abdominal soreness you’re feeling could very well be part of this…

  • Pandemic trauma as a driver of far-right radicalisation

    Excellent piece from The Manchester Mill: It wasn’t just about how much time people were spending online. Lewis now recalls that period as a moment when he could feel the presence of the state up close. “It was the first time in our lives we’d seen the government say: ‘You’re gonna do what we tell…

  • Minimalism as a philosophy of post-neoliberalism

    I tried watching this but found the people involved so unbearably irritating I could only get thirty minutes into it. It did make me wonder if lifestyle minimalism, which seemed modish amongst digital nomads and aspiring digital elites in the 2010s, could be seen as an early adaptation to post-neoliberalism. After two years of rapidly…

  • The libidinal economy of symbolisation

    The emphasis placed in Lacanian psychoanalysis on symbolisation isn’t an open-ended matter of putting everything into words. It rests on an account of libidinal economy in which some inarticulate matters are stumbling blocks for the subject, whereas others are not. This is how Bruce Fink describes it in the Lacanian subject: One of the faces…

  • I keep having dreams of things I need to do and waking up and not following through

    I keep having dreams of things I need to doAnd waking up and not following throughBut it feels like I haven’t slept at allWhen I wake to a silence and she’s facing the wallPosters of Dylan and of HemingwayAn antique compass for a sailor’s escapeShe says you just can’t live this wayAnd I close my…

  • GenAI beyond the bubble

    What will be left after the GenAI bubble bursts? Probably quite a lot given the accelerating capital investment which big tech firms are making in AI 👇 Over the last two years I’ve argued consistently that conflating large language models (as a technological development) with ‘Generative AI’ (as a hype cycle and market bubble) is…

  • Current mood in (not) AI generated images #112

  • Current mood in (not) AI generated images #111

  • Current mood in (not) AI generated images #110

  • We don’t have to be unequal, it doesn’t have to be unfair, poverty isn’t inevitable

    Reading Lynne Segal’s (super) Radical Happiness: Moments of Collective Joy reminded me of the closing lines from Jeremy Corbyn’s acceptance speech in 2015: I say thank you to everyone for all their support, friendship and comradeship during this election process. And I say thank you in advance to us all working together to achieve great…

  • Current mood in (not) AI generated images #109

  • The weird poetry which ensues when Claude tries to read my handwriting

    Tbf I can’t read it either 🤷‍♂️ Now do my pearls of wisdom to crystalise:What have they been doing for days?Via early concerts + sand wars?To what extent are my insights a result of my own hardNow as they wither, you observe?What soulless void/void would they call to create? Do mybeliefs void help me?Key ->…

  • Live the questions now

    From Rilke’s Letters to a Young Poet: Be patient toward all that is unsolved in your heart and try to love the questions themselves, like locked rooms and like books that are now written in a very foreign tongue. Do not now seek the answers, which cannot be given you because you would not be…

  • Current mood in AI generated images #108

    I want to unfold.- Rilke

  • Eudaimonic Bubbles as a normative rather than analytical concept

    I was initially extremely sceptical of Tony Lawson’s concept of eudaimonic bubbles, particularly in so far as that an intellectual community could be conceived of as taking on a quasi-bounded quality in the manner he’s suggesting: That answer I defend or explore involves the creation of wider-community-specific flourishing-facilitating contingently protected sub-communities that I refer to…

  • 📍Call for blog posts: the Intellectual Legacy of Margaret Archer

    Following our recent symposium we are inviting short blog posts (750-1500 words) reflecting on the intellectual legacy of Margaret Archer. These will be published on the Critical Realism Network blog. Here are some examples of themes these posts could address: We welcome submissions from scholars at all stages of their careers, including graduate students and…

  • 💔 The takeover, the sweeping insensitivity of this

    All those yearsThey were here firstOily marks appear on wallsWhere pleasure moments hung beforeThe takeoverThe sweeping insensitivity of this

  • Communicating with a computer through inner speech

  • Technosolutionism meets anomie

    This is not satire as far as I can tell. I’m sure it will flop (though it’s only $99 which will help) but it’s part of a broader trend towards technosolutions for contemporary anomie. This is the feature of the GenAI landscape which I think critics are massively underestimating. I’m convinced the technology will facilitate…

  • Five propositions about the social ontology of generative AI

    A summary from Claude of the arguments Milan Stürmer and I have made in recent conversations we’ve been recording Based on the contents of your meeting transcripts, here are five propositions about the social ontology of generative AI, attempting to replicate your own voices:

  • I’ve been using WordPress for 17 years

    That’s slightly terrifying 😬

  • Contemporary fascism mobilises passions which are a (fleeting) escape from the paralysis borne of a dying civilization

    From Richard Seymour’s Patreon yesterday: Specifically, we need to consider, in the context of relentless social comparison, steepening class inequality, a culture of extolment of winners and sadism toward losers, and of the increasingly toxic psychological consequences of failure, the persecutory and vengeful passions secreted by the social body. Rather than simply blaming disinformation, or…

  • The psyche is intrinsically constituted by its relation to infinity

    the psyche is intrinsically constituted by its relation to infinity. This infinity is that object of infinite desire which, even though it does not exist (it is a fantasy), nevertheless consists … From the moment that American capitalism implements the “American way of life” as a new libidinal economy through the psychopower of marketing, it…

  • A philosophical sketch of two models of generative AI: conversational agents and copilots

    In Generative AI for Academics I argue there’s an important distinction between conversational agents (which when used properly require thought + reflection) and templated systems (which by definition are intended to avoid thought). I realised when reading Ethan Mollick earlier that this could be more helpfully framed as conversational agents and co-pilots: You see these…

  • Conversational agents can be sponsors of literacy

    I just encountered this notion via Tusting et al’s Academics Writing and it immediately helped me clarify the sense in which Claude now shows up in my professional lifeworld: Professional writing practices may be acquired and sustained as much through engaging with “sponsors of literacy”, as through formal training or education. Brandt develops this idea,…

  • The triangulation of centrists emboldens the far-right

    Over the last few days I’ve been thinking back to this Richard Seymour piece about the strange connections between centrists and the far-right: Against all this, official liberalism has one move, which is to supplement its growth discourse by triangulating the far-right. Just as Biden had sought to neutralise Trump by appropriating parts of his border…

  • Will OpenAI go the way of WeWork?

    I think this comparison by Gary Marcus is overstated because OpenAI have a remarkable (if flawed) product whereas WeWork had commercial office space masquerading as a disruptive innovation. But the structural case he’s making here cannot be dismissed and it’s easy to imagine how as a slow unravelling could rapidly accelerate: I said it before,…

  • The original source of the claim that ChatGPT overuses the word ‘delve’

    The original source of the claim that ChatGPT overuses the word ‘delve’ seems to be AI Phrase Finder which makes the claim on the basis of “our dataset of 50,000 ChatGPT responses”. No more information is provided about this dataset. There’s also no information provided about what constitutes the ‘most common words’ within it. Presumably…

  • I want the joy of simple colours

    I want the joy of simple colours, street organs, ribbons, flags, not a joy that takes my breath away and throws me into space alone where no one else can breathe with me, not the joy that comes from a lonely drunkenness. Anaïs Nin quoted on pg 24 of Lynne Segal’s Radical Happiness Drunk in…

  • To evade AI detectors you literally just need to ask LLMs to avoid the statistical features which detectors rely on

    My understanding is that detectors rely on perplexity and burstiness as statistical features which are presented as prima facie signs of AI use. But Claude 3.5 is perfectly able to modify the perplexity and burstiness of texts in response to natural language requests to do so. Here are examples of low, moderation and high burstiness…

  • The privilege of escaping from the Twittering Machine

    I’ve been involved in training academics to use social media for a long time. It’s a field I stumbled into during a part-time PhD, as a committed blogger who became a Twitter enthusiast in 2010, leaving me with a sense this was a valuable shift in academic life. I imagined that social media would lead…

  • You can’t ban the English Defence League because it doesn’t exist but you can force Telegram to shut down organising groups

    I spent my train journey home last night thinking about exactly the point which Phil BC makes forcefully here: And the Daily Mail have reported that Yvette Cooper is considering banning the English Defence League. The first problem is the EDL doesn’t exist except as lazy short hand for far right mobilisation. She would be banning a…

  • Rest in peace Maggie

    I’m so glad we did this in Warwick, even if it meant it took ages to happen 😊

  • Against an instrumentalist understanding of AI: critical realism and conceptualising artificial intelligence

    This is a useful concept from Andrew Dryhurst in a recent paper in JCR. I’ve been prone to arguing for the same framing by talking about the need to historicise AI, in terms of a broader history of digitalisation then platformisation. I think Dryhurst’s framing here helps me account for how a particular framing of…

  • What need do conspiracy theories serve?

    This is a brilliant discussion from QAA with Jesselyn Cook about her new book The Quiet Damage: QAnon And The Destruction. The opening section of the interview captures the distinct thesis which QAA have developed over recent years concerning the power of a community focusing their anger on a singular enemy (the other’s other as…

  • What will be left after the GenAI bubble bursts?

    This is such an important question, initially from Cory Doctorow and then picked up by Alex Hanna. It is indisputable that it is a bubble which means that it will burst: Doctorow thinks that the residue of the bubble popping will be minimal–large models will no longer be cost-effective to train, but small open-source models…

  • A simple way to understand Margaret Archer’s morphogenetic approach

    I just heard Doug Porpora give a great explanation of Archer’s morphogenetic approach as an approach to thinking about social change. The problem is that, as he put it, people get bogged down in all the t’s which litter these diagrams: In contrast suggests Doug, rightly I think, the claim she is making is extremely…

  • Ten propositions about generative AI and the future of academic writing

    This is Claude’s summary of the arguments from the first 20k words of my new book on AI and writing, as well as the how to write blog post series I did earlier this summer.

  • And it’s sick that all these battles are what keeps me satisfied

    Now there’s gravel in our voicesGlass is shattered from the fightIn this tug of war, you’ll always winEven when I’m right’Cause you feed me fables from your headWith violent words and empty threatsAnd it’s sick that all these battlesAre what keeps me satisfied Try and touch me so I can scream at you not to…

  • If IBM is here to make your dreams come true, you can probably say the same thing about nightmares too

    A sky without a world a shore without a seaA sun without heatShining a light you can’t seeIf IBM is here to make your dreams come trueYou can probably say the sameThing about nightmares too

  • The AI video generators are superficially stunning but deeply useless

    I’ve been playing around with them occasionally and I just can’t get them to produce what I’m actually asking for. For example this is Luma’s response to the prompt “A deeply content capybara luxuriating in my suburban back garden with a fluffy black cat reclining on his belly“: Is this a capybara or is it…

  • Current mood in AI generated images (and music) #107

  • Stare into the void until the void stares back

  • Should I talk about LLMs, conversational agents, GAI or GenAI?

    I wrote a book in which I consistently used the acronym GAI to refer to generative artificial intelligence. But I’ve been struck that I’m literally the only person using this, leading me to revert to GenAI which I’m uncomfortable with because it feels horribly buzzy. But I have a pedantic objection to talking about LLMs…

  • Product placement as a commercial strategy for GenAI firms

    This is interesting from Which about the dangers of using ChatGPT for product recommendations: Interest in artificial intelligence has exploded and the technology is now finding its way into the products and apps we use the most. With all that computing power at work, can a robot now offer sound advice on the best products…

  • Creating rich multidimensional action spaces for GAI to navigate

    It’s widely understood that one of the most effective ways of working with conversational agents is asking them to assume roles. One of the most frequent problems new users have is that the responses are generic, usually because they’ve failed to fully explicate their expectations and they’ve not asked for a response in a particular…

  • The monster became self-conscious of its size and intoxicated by the belief in its own omnipotence

    What an eye witness description from Roger E. Money-Kyrle, cited in Barbara Ehrenreich’s (incredible) Blood Rites: Origins and History of the Passions of War, loc 290: The people seemed gradually to lose their individuality and become fused into a not very intelligent but immensely powerful monster, which was not quite sane and therefore capable of…

  • You assumin’ I’m a human, what I gotta do to get it through to you I’m superhuman?

    Uh, summa-lumma, dooma-lumma, you assumin’ I’m a humanWhat I gotta do to get it through to you I’m superhuman?Innovative and I’m made of rubber so that anythingYou say is ricochetin’ off of me, and it’ll glue to you andI’m devastating, more than ever demonstratingHow to give a motherfuckin’ audience a feeling like it’s levitatingNever fading,…

  • Some thoughts on machine translation in higher education

    I really enjoyed this discussion with Klaus Mundt and Michael Groves for the TELSIG Podcast. There’s a reading list attached in the YouTube comments. Here are some notes from the discussion which are my attempts to characterise the insights shared in the podcast, rather than offer my own analysis:

  • The political economy of hopelessness

    Richard Seymour on what happens when centrists triangulate against the far-right in the interests of electoral pragmatism. The part in bold is particularly bleak: Against all this, official liberalism has one move, which is to supplement its growth discourse by triangulating the far-right. Just as Biden had sought to neutralise Trump by appropriating parts of…

  • Current mood in (not) AI generated images #106

  • Some thoughts on the emerging platform economies of generative AI

    I’m in total agreement with Carlo Perrotta here that custom GPTs and AI agents constitute a familiar platform economy being cultivated by OpenAI: In all scenarios, from the lowest API access tier to the highest enterprise one, proprietary assets and infrastructure must be hired from OpenAI’s closed development environment according to a Software as a…

  • Avant-garde theorising is a reflection of the competitive individualism of the academy

    This is a great account in Kathleen Fitzpatrick’s Generous Thinking of how what I think of as avant-garde theorising, valorising conceptual and linguistic novelty as an end in itself, expresses the competitive individualism of the academy rather than being some sort of radical bulwark against it. From loc 471: The need for a different way…

  • Lacan on the trauma of birth

    From Anxiety: The Seminar of Jacques Lacan, Book X, pg 327: By emerging into this world where he must breathe, first and foremost he is literally choked, suffocated. This is what has been called trauma – there is no other – the trauma of birth, which is not separation from the mother but the inhalation,…

  • I see you fading, fading, but you saved me

    Oh, it haunts meIt’s just when your screen went blackBut I still feel you above meAnd you’ll always guide me backI see you fading, fadingBut you saved meAll night, all nightBut you break meAlright, alright(I’m here)

  • Will GenAI be seized upon as a short-term funding fix for UK higher education?

    This is a great account by Glen O’Hara of the failing model of financing UK higher education, as two short term fixes (stacking art & humanities courses, aggressive international recruitment) have been seized upon in uneven ways destructive to the sector as a whole: For a while, some vice chancellors could patch up losses by…

  • Current mood in AI generated images #105

  • We’re now up to 7 episodes of our podcast about generative AI in higher education

    (Annoyingly, none of the multiple ways of embedding Spotify seem to work on WordPress, so please click above to access)

  • Can you get Claude to explain why it refuses to insult you? An exploration of LLMs, guard rails and self-reference

    I’ve seen numerous people observe that Claude 3.5 is less venerable to ‘counter-scolding’ than previous versions of the model. If you get a scolding response (e.g. “I don’t feel comfortable continuing this role-play or providing that type of harsh, demeaning feedback, even in a simulated context“) a counter-scold often led Claude to back down (e.g.…

  • Join us on August 3rd to celebrate the intellectual legacy of Margaret Archer

    August 3rd, 10am-5pm at the University of Warwick Margaret Archer’s work has had a profound impact on social theory, challenging and reshaping our understanding of agency, structure, culture and their interplay in producing social change. Her contributions to the discipline have been wide-ranging, from critical interventions in conceptual debates to discussions about the nature of our…

  • Are special issues of journals becoming the norm rather than the exception? How will GenAI interact with this process?

    This is a really interesting analysis of the growth of special issues as a commercial strategy by journals, emerging in the grey zone between traditional and predatory publishing: Since 2000, this practice has been turned into a commercial strategy by new publishers such as Hindawi, MDPI and Frontiers to accelerate growth and generate revenue. In…

  • The stagnation of the UK economy and living standards

    Three charts from today’s Adam Tooze newsletter 👇 How do you fix this while leaving fundamental questions of wealth distribution untouched? I don’t just have an instinctive hostility to Starmer at that stage because, in a fit of post-traumatic naivety, I voted for him in the belief he would unite the party around a softer…

  • The risk to authorship when using GenAI for notetaking

    I’ve found a valuable use of conversational agents like Claude and ChatGPT in accelerating the notetaking process because it helps turn shortform notes into longform notes, as well as drawing out connections within them and suggesting potential implications. However I began to try this at a conference earlier and I rapidly realised how dangerous it…

  • Why digital elites are embracing Trump

    While I see the value in exploring the ideological infrastructure supporting the authoritarian turn amongst digital elites, I think Dave Karpf is right to argue that the material driver is a pushback by venture capitalists, whose interests are not identical with big tech, against increasing regulation of the sector: This, by the way, is the…

  • One discovers the light in darkness, that is what darkness is for

    From James Baldwin’s final essay in Nothing Personal: One discovers the light in darkness, that is what darkness is for; but everything in our lives depends on how we bear the light. It is necessary, while in darkness, to know that there is a light somewhere, to know that in oneself, waiting to be found,…

  • Current mood in (not) AI generated images #104

  • AI Policy and Higher Education Under a New UK Government

    In this episode of Generative Dialogues, Mark Carrigan and Helen Beetham discuss the implications of the recent UK election for AI policy and higher education. Key topics include: Throughout the discussion, Mark and Helen balance critical analysis of current trends with cautious optimism about potential positive developments in research funding and university support under the…

  • Generative AI and the Automated University

    Generative AI arrives into a university system that is already to a significant extent automated. Core activities are carried out through digital platforms, key aspects of knowledge work are datafied and algorithmically governed, and university communities are subjected to new digital forms of monitoring and control. Artificial intelligence also has a long history in the…

  • Generative AI and the Enjoyment of Academic Writing

    In this episode of Generative Dialogues, Mark Carrigan and Helen Beetham engage in a thought-provoking discussion about the intersection of writing, creativity, and generative AI. Drawing on their personal experiences and academic insights, they explore the nuanced ways AI is reshaping our relationship with writing and intellectual inquiry. Key topics covered: This wide-ranging dialogue offers…

  • Hatred, even of meanness, makes you ugly

    You who will emerge again from the floodIn which we have gone underThinkWhen you speak of our faultsOf the dark timesWhich you have escapedFor we went, changing countries more often than our shoesThrough the wars of the classes, despairingWhen there was injustice only, and no indignation.And yet we know: Hatred, even of meannessMakes you ugly.…

  • Current mood in (not) AI generated images #103

  • Pierpaolo Donati’s relational humanism

    I’m finding this extremely generative for thinking about GenAI in organisations. From Being Human in a Virtual Society loc 1500: Traditional humanism: The human person is a self-sufficient substance that is realized in society according to nature (the goods of relationship exist as a virtue of the people through which they pursue their perfection and…

  • You can now build and publish simple web apps using Claude

    The new artefacts function in Claude allows simple web apps to be built in natural language instructions. It’s still early days but it suggests a new horizon of zero code development is about to open up. Click below to check out a really simple use which was produced in well under a minute from two…

  • Nothing can ever be beautiful enough to wipe out the traces of imperfections and faults

    From Joel Dor’s Clinical Lacan pg 85: The hysteric is the harshest judge when it comes to the ascent to the ideal of perfection. Nothing can ever be beautiful enough to wipe out the traces of imperfections and faults. This despotic requirement inevitably causes symptoms to appear, of which the most striking is the hysteric’s…

  • I wander through the pet store asking, “What is my true nature?”

  • Two ways to use Claude as a meeting scribe

    I’ve found Anthropic’s Claude immensely useful in a couple of recent planning meetings. In a workshop we were putting post it notes on flip chart paper as part of a planning process. I then shared a photo of the flipcharts and asked it to provide a series of provocative questions to feed into our discussion:…

  • “Mummy, I done an album!”: the excremental character of creative production

    The child is asked to hold in. He is made to hold in too long, to start to introduce excrement into the realm of what belongs to his body, and he starts to make it a part of his body, which is considered, at least for a while, as something not to be lost. Then,…

  • Current mood in AI generated images #102

  • A relational theory of existential challenge

    From Being Human in a Virtual Society: A Relational Approach, by Pierpaolo Donati loc 656: When we do not know what to do with others and with the situations of life, or what relation to have with the contingent world around us, then we feel confused, weak, fragile, sad, and in crisis. Every existential situation…

  • Current mood in (not) AI generated images #101

  • Our conception of the lifeworld needs the mundane as much as it needs the dramatic

    This passage from Nikhil Krishnan’s A Terribly Serious Adventure: Philosophy at Oxford 1900-1960 (pg 213-214) captures something I’ve thought about a lot in relation to the reception of Margaret Archer’s work on reflexivity, particularly how its sensitivity to the utterly quotidian tends to be missed by many sociological critics: When Murdoch’s book on Sartre was published,…

  • The unacknowledged debt of Richard Rorty to the ethos of post-war Oxford philosophy

    From Nikhil Krishnan’s A Terribly Serious Adventure: Philosophy at Oxford 1900-1960: The big claims were about the imminence of a final dissolution: ancient knots would be cut, the old metaphysical doctrines hunted to extinction. Once the old detritus was cleared, then the revelation, ‘of a whole world of infinite subtlety and diversity with its own…

  • “I love you and that painful decaying love is the only thing between you and the shrieking nothingness you live the rest of the day”

    I finally saw Network this evening and I am obsessed 😍 everything about this film is amazing, including my favourite break up scene ever*: (Obviously I am not including the phrase ‘cockmanship’ in this judgement 🤭) I feel I must interject hereYou’re getting carried away, feeling sorry for yourselfWith these revisions and gaps in historySo,…

  • Some thoughts on the UK general election 2024

    And yet I can’t help but take an immense primitive pleasure in the impending death of the Conservatives in their current incarnation 🥂

  • The street art I saw and the street art which was really there

    If you’ve read my blog for a long time you will have seen this image I love, which I took in Berlin a few years before Covid. I found it bleakly resonant for all sorts of reasons I’m unwilling to write about directly on the blog, but it’s fair to say it’s an image which…

  • Claude 3 onwards is considerably less prone to hallucinate than other LLMs

    That at least is my growing experience. Not only does it invoke a threshold of knowledge, it resists inducements (e.g. ‘tell me more’) or scolds (e.g. “you’re supposed to be helpful, why won’t you answer the question?”) which were reliably effective at getting Claude 2 or ChatGPT 4 to cheerfully bullshit. Interestingly GPT 4 could…