13 Comments
User's avatar
Barry Eisler's avatar

I’m not surprised these two experts ignored the part where you wrote, “Is this likely to happen within, say, the next year or two? No. But when you’re talking about a war between superpowers that could go nuclear, the likelihood doesn’t have be very high to justify a concerted effort to lower it.”

I’ve had the same reaction whenever I ask someone to rate the chances of the Ukraine war escalating to extermination if the west continues its current approach. Assuming the chances aren’t zero, are they fifty percent? Five percent? One percent? Not once has anyone answered the question. Instead it provokes duckspeak: Ukrainian agency, nuclear blackmail, CHiMA (Chamberlain Hitler Munich Appeasement).

Oh, and also “Putin would never, he is rational.” Exactly the misleading, whistling-past-the-graveyard response you’ve encountered here. And also a response pristinely free of even the notion of human accidents and folly, let alone the actual history.

Catherine Wilson's avatar

I agree about rational actorism, which MAD is unfortunately based on. A Russian TV presenter back in 2022 when Putin put his nuclear arsenal on high alert asked why we needed a world at all if [imperial] Russia was not in it.

Stuffysays's avatar

All very Dr Strangelove. And how scary is that?

John Pitts's avatar

Did you realize as you wrote it that asking politicians to exhibit “poise, wisdom, and judicious restraint” is on par with asking Trump to turn the other cheek.

Ted Tillich-Farris's avatar

Humans are certainly not rational actors as rationality is not even a quality of human activity.

The idea of humans as rational actors is philosophical and practical nonsense. https://tedtillichfarris.substack.com/p/why-the-question-of-human-rationality

What me worry?

scott's avatar

In your nonzero days, didn't you argue for rational actorism and against the existence of rogue states? I guess it feels different when we're the rogue state.

Robert Wright's avatar

I still think we should act rationally toward rogue states and that the rationality of 'rogue states' is underrated. (Iran has been a much more rational actor than America lately.) But the degree of, and consistency of, rationality being assumed by people who aren't worried about AI's geopolitically destabilizing effects is not something I want to bet the world on.

JinATX's avatar

I was going to make a similar comment. This post seems like an evolution in the very worthy cognitive empathy project, which seems to me to assume actors are rational. But there are demonstrable exceptions.

Robert Wright's avatar

I've never said or meant to imply that the cognitive empathy project assumes rationality. But it's true that adroit cognitive empathy sometimes shows people to be more rational, given their goals, constraints, etc., than a surface reading might have suggested.

JinATX's avatar

After posting my comment, I thought more about what you've said about cognitive empathy and realized that you haven't said it assumes rationality. I've made that inference, which is arguably a reasonable one, but maybe that's just me.

I look forward to your book. The more I think about it, the more I see how complicated the topic is. I'm interested to see how you work through the various nuances.

Jamie Freestone's avatar

I like rational actorism (obviously rife in nuclear weapons discourse too). But I can also see that according to a certain definition of rationality, even the "irrational" actors (history's dictators, Trump, etc.) are simply psychopaths with no social emotions and a steeper risk profile than the average. I.e., of course Hitler was willing to risk all of Germany for *his* plans — he's rationally *self*-interested, albeit with a high tolerance for risk and no guilt or compassion so he doesn't care about the downside to others. Maybe the issue here is that non-psychopaths are habitually bad at modelling psychopaths.

sean's avatar

Thank God for Trump. Wherever would we look to express our exasperation with illogical, childish narcissistic behavior without him. The same behavior that has prompted endless cries of “We’ve got him now!” from Trump opponents over the last 10 years. That, like musings about AIs

over-creating paper clips due to poorly styled human directions, are a bit overwrought A poorly sandboxed AI wandered off to complete its task while its creators were on a coffee break and then we’re off to contemplating destruction.

Here’s a reassuring thought-there is no money in nuclear destruction.

That is not a complete guarantee that mistakes are impossible.

But as long as Andreeson, Musk et al control the board with their outsized egos and sociopathic bent toward self-absorption and riches I’ll trust them to avoid. blowing us up.

Which is meant to be reassuring as they would never let Trump near the button.

mary-lou's avatar

not sure if it's reassuring to realise that Andreeson, Musk et al 'control the board. it resembles the very unstable idea that 'my enemy's enemy is my friend', so that's really getting worrisome.