Links for 2024-07-23
Quote
If history teaches anything it is that there can be no peace without equilibrium and no justice without restraint. But I believed equally that no nation could face or even define its choices without a moral compass that set a course through the ambiguities of reality and thus made sacrifices meaningful. The willingness to walk this fine line marks the difference between the academic’s—or any outsider’s—perception of morality and that of the statesman. The outsider thinks in terms of absolutes; for him right and wrong are defined in their conception. The political leader does not have this luxury. He rarely can reach his goal except in stages; any partial step is inherently morally imperfect and yet morality cannot be approximated without it. The philosopher’s test is the reasoning behind his maxims; the statesman’s test is not only the exaltation of his goals but the catastrophe he averts. Mankind will never know what it was spared because of risks avoided or because of actions taken that averted awful consequences—if only because once thwarted the consequences can never be proved. The dialogue between the academic and the statesman is therefore always likely to be inconclusive. Without philosophy, policy will have no standards; but without the willingness to peer into darkness and risk some faltering steps without certainty, humanity would never know peace.
-Henry Kissinger, White House Years (1979) (emphasis mine)
Links
Conspicuous Cognition’s Are people too flawed, ignorant and tribal for open societies? and the follow up The marketplace of misleading ideas - There are two classes of reaction one can have when another writer publishes an essay on the topic that you’ve been meditating about and trying to write about for years: disappointment or joy. Conspicuous Cognition’s two-part essay (soon to be three) is for me a cause for joy: since I read Walter Lippmann’s Public Opinion back in 2010, I’ve wanted to write something about it. These two essays sum up everything I wanted to say about the book and the issues it discusses better than I have managed. Now, when I want to refer to the ideas, I can just link to them, which is a tremendous time saver, and lets me get onto thinking the next thought.
Russia 1985-1999: TraumaZone - Virtuoso documentarian Adam Curtis, long time spelunker in the BBCs vast film archives, emerged into daylight in 2022 with this stark 7-part chronicle of post-Cold War Russia. The subtitle is “What it felt like to live through the collapse of Communism and Democracy” and that is an exact description of what it is about. Foregoing his distinctive narrative voice, Curtis tells the story is entirely through B-roll footage shot by BBC camera crews across the length and breadth of Russia, with title cards to give context to what is happening on screen. It's a brilliant piece of work, in my opinion one of the very finest films Curtis has made.
A word of warning: it is a profoundly disturbing and unsettling series. Whatever one thinks of Russians, their leaders, and the way their country acts in the world, if you have the least bit of compassion you will find it hard to watch ordinary people being hammered by forces beyond their control as their society disintegrates around them.
While I do recommend watching it, I do not recommend bingeing it.Ribbonfarm Studio’s Massed Muddler Intelligence - a compelling vision of a methodology for scaling AI applications from chatbots to effective agents via multiagent workflows. I’ve been following the academic research on this for some time, and doing lots of experiments with frontier model APIs. The essay is an insightful piece on how to stick one’s head up above the clouds of our contemporary AI debates where most people are focused on “one model to rule them all” and blind to the potential of networks of agents working together. Los Alamos was no single person, nor was Bell Labs, and there is no reason to expect that the transformations from AI will come from some single oracle we all consult.
80,000 Hours Carl Shulman: The economy and national security after AGI (Part 1) - interesting interview with Carl Shulman on the theoretical limits of computation given terrestrial and then solar system resources. The scenario Shulman describes, where in a matter of decades we have billions of human-level or greater intelligences available to us for all manner of tasks, is extreme. The mistake critics of the existential risk-focused school of AI Safety research make is to focus on the improbabilities of the extreme scenario. Shulman provides strong arguments against the common dismissals of the extreme economic growth he forecasts, and his reasoning is worth following through.
The point of paying attention to arguments like Shulman’s is not the extremity of the scenario he paints, but that there is a distribution of scenarios from here to the extreme scenario. Like climate change, the extreme scenario can very well not come to pass, but that doesn’t mean things are not going to change a lot and one should be prepared, at least on an individual level, for those potential changes.Clark Ashton Smith’s The Door to Saturn - Of the weird tales authors of the 1920s and 30s, H.P. Lovecraft attracts almost all the attention with his Cthulhu Mythos cycle. Robert E. Howard follows behind with his Conan stories. Almost forgotten is the astoundingly talented Californian poet-fantasist Clark Ashton Smith, mutual friend of them both. For my money, Lovecraft composed a few stories superior to Smith’s1 but Smith was the more consistently excellent writer. His works are full of unmatched atmosphere, poetic prose, and sardonic humour.
Specifically The Call of Cthulhu and At the Mountains of Madness.