Post #40: This Is the End

27 June 2023

This is the end, beautiful friend

The end of our elaborate plans

The end of everything that stands

—The Doors

Let me start, to put things in perspective, by pointing out some of the big ways in which I do not believe that the end is nigh, mine or the world's. The end in a planetary sense was never at issue; we speak of destroying it, but that is a mere conceit of human might, dressed up as fear. We have, at most, the power to make the planet uninhabitable for ourselves and a few kindred species, and I foresee nothing of the sort.

The world is not likely to end in a flash of nuclear fire or because we overused our air-conditioning, and while we may have horrible ordeals to pass through yet, if recovery was possible from the Thirty Years' War and the Black Plague, from our twentieth century world wars and genocides, from napalm raids in the morning and the killing fields of Pol Pot, then we are not looking at extinction any time soon, I don't think. As for my own part, the doctors and actuarial tables tell me to be prepared for a few more decades, possibly as much as another half-century, however keen I may or may not be on the prospect. Even the end of my blog is nowhere in sight, although I will probably cut off the print edition here and publish only the first forty posts for now.

What set off the scare was a lot more limited in scope. All it took was a message from a close friend who has, like me, devoted much of her life to the English language, who enjoys some credit as a writer and a journalist, and who told me that ChatGPT “writes such good sentences” that she found herself “tempted to let it do all her writing”! To hear such sentiments from a fellow teacher makes me shudder, given what they imply about how tempted students must be feeling who care mostly about getting by and looking good—rather than being good and doing right by the language, let alone the art of writing. That said, the world does not end in the classroom either.

When I make myself as optimistic as possible, I can imagine the bots of the near future raising general literacy levels considerably—not so much by encouraging the habit in humans, but at least by cleaning up verbal messes in their pedantic smartass manner, deploying their technical aptitude (which will keep increasing rapidly as they are fed more and more materials to learn from), and perhaps even developing a modicum of style and elegant diction here and there. Just how high the standard of literacy will be set by the programmers remains to be seen, but perhaps there will be a dial before long on which one can specify the desired level of eloquence, more or less on analogy with chess computers and their levels of proficiency.

For the better human writers, the Big Blue moment has not yet arrived; the machines are beginning to write more or less competently, but so far they do not (or at least should not) impress anyone with the great beauty, cogency, or liveliness of their prose. But the time may not be far off when even our Kasparovs of the quill and keyboard must concede defeat, since they can only learn their art and craft from a few thousand books in a lifetime, while the machines can ingest and digest them by the millions. It is as if one tried to hold one's own with chopsticks against light sabers, and it may be just a matter of time before even the inner citadel will fall. That the city of books has been at the mercy of the invading hordes for some time already, with the libraries on fire everywhere, will not come as news to the scattered diehards who are still, here and there, staking their life's labors on a hopeless last stand.

Going by the example of chess, for which the moment of defeat came and went twenty-five years ago, one might wonder whether there is really quite so much cause for concern. Chess absorbed the heavy blow to human pride in 1997, but survived, and even thrived (from what little I know) as an art and a sport. The worldwide hierarchy of players has become more defined, not less, and so far as I am aware, the human game has not become any less artful. There may even be more players than ever. But the crucial point is that in chess it must have proved possible to establish both a rigid norm and a policing system that precludes anyone from bringing the computer into a competitive game. Learning, training, and sparring are one thing, the rankings and the master’s grades quite another; if this line could not be rigorously drawn and defended to the satisfaction of the players, there would be no recognized human grandmasters anymore.

An example with very different implications is that of aviation, which also used to have an artful and competitive, even heroic side, in the days of its pioneers, but which has lost practically all its glamor to the autopilot. There it was the utilitarian dimension that did for the human factor in the cockpit, and with enormous practical benefits to mankind: by making flying so much safer, what was lost on the side of human skill was more than made up on the side of human life. For the truly hell-bent there remain the niche thrills of stunt flying and dog fighting.

Such utilitarian benefits may, to some extent, come to apply in the domain of commercial writing as well. Vast swamps of verbosity could presumably be handed over to the bots for draining—at a minimum, to spare us humans the drudgery of coming up with such dreary routine materials, and more hopefully, to help improve the prose where there is room for it, though much will depend on how the programming is done and what models the bots will be given to emulate. Garbage in, garbage out remains the motto of the day.

To teach the bots good literary taste, the instructors themselves would need to possess and understand it at least well enough to set the tone and direction—and somehow I do not get the impression that training master stylists is the point of the exercise. If the bots should find their way alone, proving capable, against all naysaying, of real polish and refinement acquired without their masters' guidance and against the currents of the times, then their success would mark a technical triumph of such magnitude that I would have to withdraw my reservations and bow before the human ingenuity that made possible such prodigies. Then I would find myself obliged to bow before the golden calf, but not until then.

Alas, the art of writing, allied as it is to art of thinking along human lines, is not (in contrast to chess) a contest with clear and largely uncontested rules, and I fear that it will be much harder to establish the kind of red lines that saved the game of kings as we know it. We may see some movement, before long, towards certified bot-free content, but with two crucial caveats. The first is that legitimate uses may prove quite hard to distinguish from the illegitimate. Even the spell-checking function is a primitive bot, after all, and competent grammar-checkers are so to an even greater degree. At a more principled level, if getting help from a human editor is legitimate, then why not from a bot-editor who makes his services freely available to all comers? If anything, would not the playing field be leveled and the contest made more honest, since it is probably the best writers, generally speaking, who have usually had access to the best live editors, compounding their advantages and making their talents appear even more formidable than they really are?

Second, more decisively and alarmingly still, where will the readers line up? How much do, how much will they care, given the ongoing erosion in prestige that serious reading and writing have been suffering for generations—what with the successive shock waves of radio, film, and television proving just as deleterious to print culture as the digital tsunami. How much longer will human readers have either the ability or even the interest to discern the difference between the human artist and the bot imitator? There are, after all, billions of contemporaries on our planet who are sincerely convinced that their everyday radio tunes are better music than Bach, Beethoven, or Brahms. The great deluge may not flood and drown absolutely everything, but how many islands will be left as refuges for those us who have made their commitment to a life of letters above all else, and how isolated and remote will our last sanctuaries be?

The art of sailing was not completely destroyed by the advent of steam ships; but the world of sailboats that Joseph Conrad still knew, loved, and mourned, did disappear and became the stuff of legend and museum exhibits, though here and there a few lone exceptions may defy history within the narrow confines of naval training ships. The utilitarian benefits of the change to steam and diesel may have been enormous, but that should not fool us into thinking that nothing of beauty has been lost, even in view of a few showy luxury schooners that may keep the tradition alive to some extent even among civilians.

Vinyl records have not disappeared, film photography survives in niches, and classical music is still holding its own in some corners, though kept afloat largely by car commercials and beset by Muzak on all sides. But these are the shrunken worlds of hobbyists, enthusiasts, and collectors of government subsidies and charitable donations. To what extent they are still major cultural forces is debatable; they are certainly not so in a mainstream sense.

So my world will not end, at least not for the foreseeable future; but it will keep contracting while what surrounds it will become more and more unrecognizable and alien to me. Then again, having never properly arrived in this blessed 21st century of ours, culturally at least, and having instead regressed towards the eighteenth and nineteenth centuries from the twentieth for several decades now, all this is nothing very new for me. In the end, it all comes down to self and its conceits, I suppose—only Dukkha arising—and before long it will be time to leave the noisy show behind, which may not be such a bad thing as it would have seemed to me ten, twenty, or thirty years ago.

PS: I made a point of not entering upon the dread domains of ChatGPT before, but today I decided to put the beast to the test by feeding it my opening paragraph and asking for improvements.

The result made me feel safe, for now, in that the revised version was so clumsy and inarticulate that it was more laughable than menacing; but the self-confidence with which the machine proffered its third-rate wares was anything but amusing or reassuring, and I while I was not tempted by its tasteless offerings, it made me wonder, as I said, nay worry desperately, about how much longer there will be readers who can be relied upon to tell the difference. That worse writing will often be found more impressive on account of certain fool's-gold contrivances is, after all, nothing new under the sun, but has ever been the stock-in-trade of the pseudo-sophisticated.

To put it bluntly, the bot writes like a know-it-all undergrad who thinks he is a genius because he has figured out how to plagiarize Wikipedia without getting caught, and who knows how to use a thesaurus badly. So they are developing some recognizable human traits, these little robo-creeps, only not the right ones. They may acquire more mature and well-rounded personalities with time, and they may come for my job too before long—but not yet, not quite.

“The future is coming, and you're not in it,“ Maverick gets told in the one quotable scene in an otherwise deplorable remake of a far from literary-grade original. “The end is inevitable; your type is headed for extinction!“

“Maybe so, sir,“ says Maverick, “but not today.“