Post #195: I Concede (AI Wins)
7 June 2026
Surely my type—the tradition-minded bookish set, with a lifetime’s investment in reading, language, and mental culture of a particular kind—cannot be expected to cheer the destruction of our own human capital. These rapidly depreciating treasures of yesteryear, now rapidly approaching junk-bond status, have been defining points of pride for us, labors of love, and deep sources of meaning. And now the omni-brain can perform, with a proficiency almost disconcerting, what it would take a scholar hours, days, or perhaps months and years to do, if it could be done the human way at all.
I am only too aware how much was fusty and snobbish about the old ways, at universities especially, and nowhere more so than when it comes to studying classical Greek, “that pride-producing language” (Thackeray, see no. 126). The veil was pulled back mercilessly for me by mean Mr Veblen, and by a last bitter bout with my Britishy demons two years ago (nos. 126, 173). Goodbye to all that, I suppose, and fare thee well, Britannia. And yet, along with the sweat and the tears, it was never all dust and rust, but some genuine lifeblood too, and a few glimpses of the soul here and there.
My defeat, then, can hardly be a lighthearted one, but it stared me unmistakably in the face when this morning I thought to ask ChatGPT about a famous passage in Aristotle’s Metaphysics (982b12-13) that had been giving me plenty of trouble the day before: “It was on account of wonder that human beings began to philosophize.” The machine delivered admirably, not only the Greek text and a translation, but also a little discussion of how wonder induces perplexity (aporia, see no. 124), complete with the most apposite references to Plato (“Philosophy begins in wonder,” Theaetetus 155d). Granted, the machine did not provide what is perhaps the human essence at issue: the inclination to ask in the first place (on account of the disposition to wonder) and a good sense of what to ask. Still, it was a bravura performance in Year 4 Anno Chatboti.
The rout that the bots are making could also have been illustrated, to wider appeal, with a casual demonstration that my girlfriend made over lunch the other day—first translating a few complex insurance documents from Thai into English, preserving all the graphics, within half a minute or so, and then showing me how one could now speak virtually anything into the machine and get it back, quite eloquently rendered, in practically any other common language, in a voice that no longer caricatures the human but captures it convincingly, and can be made to impersonate anyone to boot.
In other words, the language problem that has beset and bedeviled mankind since the days of Babel has been solved, and the translations that get churned out in a second are now so damn smooth and dependable that one rubs one’s eyes in wonder, remembering how uneven and erratic they used to be on Google Translate until very nearly yesterday. (To say nothing of earlier efforts at automating customer service: I recall with dread my encounters with algorithms that I could not get to understand even my most basic queries and concerns. Now, tease and taunt the machine as much as I like, it ends up unwinding my most complex and tangled questions with such precision and ease that it keeps boggling the mind not yet accustomed to witnessing one technological miracle after another.)
Even on the political side, as someone who has taught this fraught subject for many years (though I have mostly retired from it now), I must bow before how ably and reasonably, on the whole, the bots can summarize complex political issues. One might quibble, of course, but in a world that often looks as if it is fraying epistemological on all sides, I wonder whether we do not sorely need such a supra-human arbiter, even if it is far from a binding process (nor should it be). I already dimly anticipated something of the sort a year and a half ago (see no. 169 in January 2025), but the plot is thickening by the day. Which is not at all the same as a blind techno-optimism, but still a hopeful lining around the dark clouds gathering in our cultural skies, so far as I can discern them.
Not that my ability to look ahead amounts to much, or anything. As I keep reminding my students, nobody can predict the future; all we can do is project trends forward and combine them with more or less intelligent and imaginative guesses, to give the false impression, in hindsight, that someone was able to foresee things. But it is not so; nobody can know for sure, and things might always stall, miscarry, unravel, or go into reverse. That said, the tea leaves, though no magic ball, are telling a very different story at the moment, read them any way you like.
None of these miraculous developments are likely to help our ability to think for ourselves, alas, and that must be a serious concern for anyone with a bit of philosophical sensitivity. Such worries are nothing very new either: we have been prosthetic gods for some time, as Freud put it in his Civilization and Its Discontents (no. 26, no. 35), and growing artificial wings, while deeply impressive, points back all the way to Daedalus. Merely to draw out the implications of what can now be done, not by the crack troops but by the common foot soldiers, already exceeds altogether the modest powers of my imagination. What the world will look like if the pace is merely maintained, not stepped up—let’s say for another five or ten years, to be cautious—is already beyond fathoming, to say nothing of what might happen if there is further acceleration, or if the machines start working on themselves at their own speed, not ours, and we witness a bootstrap leap in technology beyond anything we have ever seen.
To think that barely three and a half years ago, in early 2023, I could still teach an introductory session—for a class on Rome, as irony would have it—and express my confidence that tests of the kind I would give, asking questions that required reflection, were hardly open to ready answer-by-Google, the abiding challenge of earlier years. I remember vividly how a particularly alert student (in an unusually able and enthusiastic group) asked me, not at all snidely or maliciously but with a kind of amused wonder, whether I had perhaps not heard of ChatGPT yet. And indeed, I was at the time still blissfully ignorant. But such bliss cannot last, as I keep telling students who like to toss me the dumbed-down version of Thomas Gray’s lines (“Where ignorance is bliss, ‘tis folly to be wise”), as if the point were to affirm that such is often the case, and wisdom to be slighted, not sought out. It may be so in an age where the wretched stickmen reign supreme over our cultural landscape (no. 83); but not in 1742, for all the shortcomings of the eighteenth century. So far as the life of the mind is concerned, we should not be too sure that we live on the brighter side of things.
But wait, there may be an angle I’ve been missing. I have pondered before how strange it is that while we strive to be as perfect as we can (some of us at least—per aspera ad astra, while the fuel lasts), it is not our triumphs but our imperfections that tend to endear us to others (no. 21). I wonder whether the same might not emerge from our dealings with the bots before long, as we discover that although we greatly value (and depend on) the machines’ formidable capacities, it is our flaws and limitations that define us as human beings and make us interesting to each other. Is it not becoming quite clear already that even the most interesting text out there will not come fully to life, and will likely fail to move us, unless we find a human author to go with it, a real human voice to give it resonance? And while machines can simulate such voices (and do it very well), they lack the essential ingredient: the authenticity, the scars and scratches, the halting and the stumbling, that bear witness to the lived experience behind the words. Take that flawed component away and we may still prize the sheer power of knowledge, and its stupendous fruits, already harvested or still at the seed-stage; but the distinctively human element will be conspicuous by its absence, and along with it, that which ultimately means something to us—at least as long as we remain recognizably human ourselves.
The bots may not yet have arrived at the stage where they make no more mistakes at all, but could anyone doubt that they are getting closer every day? (Glitches become grist for the mills of further progress: they will get rarer and rarer as every previous instance gets ironed out and precluded for the future, much as accidents have made driving and flying so much safer over time.) To test how far the machines had come in sorting out human messes, I got personal a few months and asked the bot whether a certain DP in Bangkok might an asshole. The answer was reassuring, though tempered by awareness on my end that humans have devoted much time and energy to teaching the machine contemporary manners and not offending anyone, on pain of reprogramming. (Kindly refrain from repeating my experiment: I suspect that the algorithm would pick up on it and start sensing fire where there appears to be smoke.)
More recently, I followed up by asking what “Beggar’s Buddhism” is supposed to mean. This time I got, first, a decent answer as to how one might understand the label in general terms, and then a reference to my book and blog, with the appraisal that it was advocating a life of simplicity in the midst of urban plenty, or something along those hackneyed lines—when in fact I’ve been advocating very little, presenting a gentleman’s creed (within my modest means), and not a peep about slumming it, but much said about the magic of money (nos. 24, 37, 47, 65, 68, 88, 99 etc.). Gotcha: for once I had caught the machine on clearly the wrong foot. But not for long. When I tried to replicate the same thing, just now, I got quite different and rather better answers. To wit (I summarize): that the term is informal and does not refer to any distinct school of Buddhism, but to a pragmatic, personal, everyday take on the practice as described by an earnest and inquisitive lay meditator, writer, and educator living in Bangkok:
It emphasizes practicing mindfulness and the Dhamma in everyday life (like doing laundry or dealing with daily pain [nice touch!]) rather than pursuing intense, monastic ideals. It represents the act of “begging” for insight [not quite, but a creative reading]—stumbling along the spiritual path with uncertainty and humility, rather than claiming to have attained ultimate enlightenment.
Brilliant writing this is not, even after a little stylistic clean-up work.* The represents is especially feeble, while to say that the stumbling and ragged meditator in question (see no. 189) does not claim to have attained ultimate enlightenment, while literally true, misses the mark so comically that it might pass for parody. Nonetheless, if the machine was able to catch up so adroitly with one obscure scribbler, lost in the great haystack of the web, then one can only marvel at what mere sand might keep learning to do, now that we have taught it to think.**
Ta panta rhei: all things are in flow, all things in flux—the beating heart of Heraclitus’s and the Buddha’s teaching alike (no. 134)—and therefore any reliance on the ever-changing swirl of Samsara must disappoint expectations before long and prove a sure recipe for suffering. Now you hear it, now you don’t, as I keep repeating in different ways (nos. 11, 12, 134): now the message falls on fertile ground, now on weeds that may become fertilizer one day (no. 163), now on sand or even stone, and then perhaps on receptive soil again. So be consoled, weary heart: it is already much to have a direction to steer in, or a goal to aim for, however short one’s arrows may keep falling. It is even something to have a mere choice of nightmares, as Marlow sighs in Heart of Darkness (no. 65). So choose your suffering carefully (no. 14), for even AI, in all its triumphant might, cannot relieve you of it.
*Presumably the bots could be taught to be much better stylists if they were trained on our best writers, rather than roaming the web as freely as they do. This begs the question of how to identify the best, of course, but it also points to a more fundamental limitation of the machines. We humans do not pick our words as a matter of logos alone, or of aesthetic sentiment, or of emotional proclivity, but through a complex process of assimilation and filtering that reflects an entire lifetime of experience, development, and reflection—of what influences came to us, at what points in our lives, and what we made of them. How we write is and must ever be a digest of what we’ve read, thought, and lived through; a machine can mimic the results, perhaps very convincingly, but unless it were to evolve sensitivities of its own, the voice must remain hollow, whether we are able to detect the false note or not. The Buddhists may remind us that a human writer’s self too lacks a core: but in saying so they do not mean to deny the authenticity of a voice, from moment to moment; they mean that there is no one abiding identity behind it—that the voice too changes all the time. Indeed one might suspect that an author who spoke with perfect consistently all the time would not be as interesting to us as someone who took unexpected turns now and then, and perhaps even contradicted himself occasionally.
**Perhaps this has become an online trope already, but I heard it for the first time from Jack Clark the other day, in a conversation with Rory Stewart and Matt Clifford on The Rest Is Politics, 2 June 2026.