TL;DRussia Weekend Roundup

11 February 2023: Intelligent artifice, plus texts and tunes

(NB: Before I get started, a quick apology is in order. I’m on holiday this coming week, and so there will be no TL;DRussia Weekend Roundup on 18 February.)

I have a confession to make: I spent way too much time in my formative years watching Star Trek. It was the perfect antidote to a childhood in American suburbia, combining mild peril with the promise of endless exploration. I used to pretend my school bus was a starship, the subdivisions along the route were star systems, and most of the kids at school were, well, Klingons. It’s a miracle I made it out alive.

But of all of the Star Trek tech that captured my imagination — the phasers for defense, the transporter beams for instant escape — it was the U.S.S. Enterprise’s onboard computer that always seemed most amazing. Never mind the unintelligible user interface: you could ask the thing a question — any question — and it would give you an answer, drawn from the sum total of galactic knowledge, definitive, and analytical. That kind of technology, it seemed to me, must be hundreds of years away.

I know you see where this is going. My Facebook feed exploded this week — for the umpteenth time — with academic fretting about the future of student essays. There are still a lot of questions that ChatGPT cannot answer. It does well with generic explanations of things, even very complex things, but tends to fall down when you ask more complicated questions about the literature on a given topic, for example. But let’s not kid ourselves: it will, in short order, be very nearly as good as that computer on Star Trek. And then what?

One of my academic friends faced this reality with a question no computer is ever likely to answer: does it matter? And perhaps he’s right. We teach students to write essays in part because it’s a good way to practice critical thinking, but also because there is no other way to get a detailed answer to a complicated question. Or, rather, there was no other way. If we can get that now from ChatGPT, then maybe essay writing ought to go the way of subsistence agriculture?

I’m not quite ready to go that far — for one thing, the power implications of relying on a human-designed algorithm to come up with more or less standardized answers to complicated questions are severe — but but there is, I think, one kernel of truth to this idea. If we’re interested in critical thought, the real skill isn’t learning to answer questions: it’s learning to ask them.

What I’m thinking about

While we’re on the subject of existential crises, I was thrown into a little bit of one this week, courtesy of Vladimir Putin’s propagandist extraordinaire, Margarita Simonyan:

Francis Scarr @francis_scarr

I don’t know whether to laugh or cry RT boss Simonyan responds to Russians complaining that they don’t understand the invasion’s real aims She claims Putin deliberately chose the vague aims of ‘demilitarisation’ and ‘denazification’ so Russia had flexibility as things unfolded

Regular readers of this newsletter — or anyone else who’s been unlucky enough to hear me pontificate at just about any point in the past 12 months — know that I’ve been arguing for ages that Putin’s rhetoric around the war is purposefully vague. The obfuscation serves multiple purposes, but two in particular stand out: sowing uncertainty among his opponents about his aims, and thus impeding strategy; and giving himself wiggle room to sell to his domestic constituents whatever result he is able to achieve.

It’s not, on the face of it, a terribly controversial hypothesis, but it nonetheless runs counter to a lot of the conventional wisdom about this war — in particular, the assumptions that Putin cannot stop until he achieves the maximalist aim of complete control over Ukraine, and that Putin cannot accept incomplete control of the four Ukrainian oblasts he claims to have annexed. If you believe that Putin is operating on the basis of exclusively maximalist aims and immutable red lines, that gives you one set of possible vectors for the future of the war. If you believe, as I do, that Putin’s rhetoric is purposefully ambiguous, then the implication is that Putin has much more flexibility — and thus we need to prepare for a considerably wider set of eventualities.

I was secure in this analysis right up until Simonyan confirmed it. Simonyan’s job, after all, is very clearly not to tell the truth. The established modus operandi of Russian propaganda is to confuse — to flood target discourses with so many different narratives that no consensus could possibly emerge. So if Simonyan says I was right (in so many words), should I assume that I was, in fact, wrong?

The question, of course, is meant in jest — but then, as the Russian saying has it, in every joke there’s a bit of truth. (Or, as some would put it, in every joke there’s a bit of joke.) The reality, of course, is that Simonyan doesn’t know Putin’s mind any better than I do. It’s not her job to know what he’s really thinking, and there’s no reason for him to clue her in. What she said, then, is what she thinks it would be useful for her audiences to believe. It may or may not also be what she believes.

While it is analytically useful to recognize the degree of artifice in all of this, there’s one big drawback. Understanding that he may not believe what he says may allow us to avoid basing our strategies on false certainties. Indeed, confronting uncertainty head on is useful: it helps us ask better questions. But it also gives Putin the opportunity to multiply the range of eventualities for which we will need to prepare. Analytically, maybe we can handle that. The war will not be won by analysis alone, however. Analysis will need to inform allocations of money, materiel and manpower, and that materiality inevitably means that there is a limited number of eventualities for which we can, in any real sense, prepare.

One way to deal with that problem is to return to mind-reading. After all, Putin does presumably believe something, and it would be very useful to know what it is. Despite the fact that it puts me in agreement with Simonyan, however, I continue to believe that we cannot know Putin’s true intentions and should not try to guess. The task, then, is to get better not at answering questions, but at asking them: What are the choices Putin faces? How are those choices structured? What are the structural limitations on his options, and the tradeoffs involved? What do we know about how those variables tend to interact? By asking better questions, we can narrow our field of view and sharpen our focus.

Or we could just ask ChatGPT.

Or maybe not.

What I’m reading

Sticking with the theme of mind-reading, Joseph Cirincione had an excellent piece on Thursday — not, in fact, based on mind-reading at all — in the Daily Beast on Putin and the logic of nuclear escalation. The reason why Putin hasn’t gone nuclear, Cirincione writes, is not because we haven’t crossed any red lines in supporting Ukraine, but because , in his words, “Putin is losing slowly”:

There is no moment in this war where Putin was faced with a decision to go nuclear or go home. Ukrainian forces are advancing, but they do so by tens of kilometers, not hundreds. Russian forces are being ground down, but not routed. There could be a sudden collapse, but a slow retreat back to Russia’s borders seems more likely.

In this, Cirincione echoes the analysis published in September by Ukrainian nuclear arms expert Polina Sinovets, who similarly argued that the greatest risk of nuclear escalation would come if Putin were to face a clear “now or never” moment. The problem for Western decision-makers, however, is that “now or never” is very much a subjective judgment. A year into the war, we have a clearer understanding of what kinds of situations are not likely to move Putin’s finger towards the button, and that’s useful. But we still don’t know much about what, at the end of the day, might trigger a nuclear response — and it would be irresponsible not to acknowledge that residual uncertainty.

Other interesting texts that caught my eye this week included:

• An excellent investigation in Wired by Masha Borak, on the development and deployment of facial recognition technology by Russia’s growing police state. Apart from providing a useful history of the growth of coercive power in Russia, and a sobering window into an emerging dystopia, the piece helps us understand how the foundations for Putin’s decision to war were laid. To be clear, neither the author nor I are saying that Putin knew nearly a decade ago, when Russia began investing in surveillance technology in earnest, that he would launch this war. But the repressive apparatus he so methodically built made this war possible, and it would be foolish to ignore that fact.

• Picking up on the discussion of the Russian economy from last week’s newsletter, I came across a detailed discussion by four leading Russian economists — Oleg Vyugin, Evsei Gurvich, Oleg Itskhoki and Andrei Yakovlev — in Re:Russia on Russia’s economic future. The question, they write, is not whether the Russian economy will degrade, but the speed with which it will do so.

• The speed of degradation of Russia’s cultural scene, by contrast, is not in doubt: it is proceeding with astonishing speed. Following on the news that Irina Lebedeva, the director of Moscow’s New Tretyakov Gallery, had been dismissed and replaced by the daughter of a FSB general, New Lines Magazine has a solid and depressing — if entirely unsurprising — report on the onslaught that the state and its lackeys have inflicted on all forms of genuine cultural expression.

What I’m listening to

As the final inking of this week’s roundup finds me at Dulles, awaiting a flight to Heathrow, I bring you the tune in my head. I’ve mentioned Laura Marling in the newsletter before — as one of a very small number of artists who has never, in my estimation, written a bad song — but never actually shared her music here. I first heard this song in a hole-in-the-wall club in Finsbury Park, before I had ever listened to any of her albums. I’ve been hooked ever since.