Why do you think, "Standard decision theory says that basic values should be constant, and not change when you learn new things about the world."?
Decision theory was my concentration in graduate school (albeit circa 1990) and I don't recall this. For any given decision problem, you had to assume that you could apply the same value function to all outcomes. But there was no restriction that different decision problems couldn't use different value functions. As a practical applied matter, the usual case was actually different value functions across problems.
Now, if the value function changes across problems, the decisions in those problem may not be consistent. But if you come to believe that your previous value function was in error, that's a feature, not a bug.
I clearly recall covering this topic, but not that your assertion is categorically "true". If your point is well established, you shouldn't have any trouble providing some explict, highly-cited references.
What I suspect is that your use of "says" and "should" are different from the way I typically interpret them.
If you could clarify precisely what you mean, it would help in evaluating the extent to which decision theory should be seen as supporting your hypotheses about value changes.
There should be a caveat to this - since if you learn about powerful agents whose values conflict with your own, you may decide to compromise with that agent on values - instead of duking it out with them - which could be costly. For more examples of this sort of thing, see this section: "While it is true that most rational systems will act to preserve their utility functions, there are at least three situations in which they will try to change them" - https://selfawaresystems.com/wp-content/uploads/2008/01/ai_drives_final.pdf
Western elites are coming back around. The rise in religious exploration points to this. It reminds me of Spengler - the soul of the grandfather returns in the grandson.
Modern world changes fast in technology and politics, not just culture. How would you get cultural drift back on track? Maybe create lots of little societies, and track how they flourish?
Many culture changes I would categorize as "questioning received wisdom."
It began with the scientific revolution. Science provided incontrovertible proof of the value of questioning received wisdom. Copernicus questioned the centrality of our position in the cosmos, and later thinkers took this meta-idea to question the centrality of certain races, or aristocratic classes, or genders, or religious deities, or sexual orientations, and so on.
Like every great idea, people will ride that horse as far as it will go. Are we close to an end? I doubt it because our culture still holds onto a lot of incorrect ideas.
Though I'm not sure I follow what is meant by the second modern mistake: "not realizing that our increased cultural scales gave more data to help innovation of stuff that can vary within cultures, like tech and business practices, but less data to promote innovation of shared cultural features like values and norms"
I think we are Romantic in the way people used to be Medieval--- the tension between science and subjectivity, which is at the heart of what you describe, arises in C18th. Modernism is just one part of that.
"reasoners and innovators gained status" being the biggest thing that started it all.
Winner societies being copied values-wise is odd; african, muslim, south american, chinese, indian values have changed less than the western societies that are and have been 'winning' have changed.
Atheism and 'rationality' in the populous are not based on much real capability for thought, rather atheism became a cultural norm in a lot of the western world and 'rationality' holds up like wet sand when abandoning the normed tracks
Why do you think, "Standard decision theory says that basic values should be constant, and not change when you learn new things about the world."?
Decision theory was my concentration in graduate school (albeit circa 1990) and I don't recall this. For any given decision problem, you had to assume that you could apply the same value function to all outcomes. But there was no restriction that different decision problems couldn't use different value functions. As a practical applied matter, the usual case was actually different value functions across problems.
Now, if the value function changes across problems, the decisions in those problem may not be consistent. But if you come to believe that your previous value function was in error, that's a feature, not a bug.
Seems like your training didn't consider the same agent making a set of different decisions across time.
I clearly recall covering this topic, but not that your assertion is categorically "true". If your point is well established, you shouldn't have any trouble providing some explict, highly-cited references.
What I suspect is that your use of "says" and "should" are different from the way I typically interpret them.
If you could clarify precisely what you mean, it would help in evaluating the extent to which decision theory should be seen as supporting your hypotheses about value changes.
There should be a caveat to this - since if you learn about powerful agents whose values conflict with your own, you may decide to compromise with that agent on values - instead of duking it out with them - which could be costly. For more examples of this sort of thing, see this section: "While it is true that most rational systems will act to preserve their utility functions, there are at least three situations in which they will try to change them" - https://selfawaresystems.com/wp-content/uploads/2008/01/ai_drives_final.pdf
You could have linked to this on laziness and excessive discounting of the future https://www.overcomingbias.com/p/culture-drift-predicts-decadence
Link added.
Western elites are coming back around. The rise in religious exploration points to this. It reminds me of Spengler - the soul of the grandfather returns in the grandson.
Modern world changes fast in technology and politics, not just culture. How would you get cultural drift back on track? Maybe create lots of little societies, and track how they flourish?
Many culture changes I would categorize as "questioning received wisdom."
It began with the scientific revolution. Science provided incontrovertible proof of the value of questioning received wisdom. Copernicus questioned the centrality of our position in the cosmos, and later thinkers took this meta-idea to question the centrality of certain races, or aristocratic classes, or genders, or religious deities, or sexual orientations, and so on.
Like every great idea, people will ride that horse as far as it will go. Are we close to an end? I doubt it because our culture still holds onto a lot of incorrect ideas.
Very intriguing. This makes a lot of sense.
Though I'm not sure I follow what is meant by the second modern mistake: "not realizing that our increased cultural scales gave more data to help innovation of stuff that can vary within cultures, like tech and business practices, but less data to promote innovation of shared cultural features like values and norms"
I don't know what you mean by "modernist" but I think Romanticism is a much more significant moment for this break
What term would you use for the entire modern era of the last few centuries?
I think we are Romantic in the way people used to be Medieval--- the tension between science and subjectivity, which is at the heart of what you describe, arises in C18th. Modernism is just one part of that.
Most of my readers would be confused by that terminology.
Maybe so, but the intellectual roots of what you are talking about are not "modernist" and to say so introduces a confusion of its own, imo
It looks a lot like you failed to overcome your (well, you and your poll base's) biases here
Care to elaborate?
"reasoners and innovators gained status" being the biggest thing that started it all.
Winner societies being copied values-wise is odd; african, muslim, south american, chinese, indian values have changed less than the western societies that are and have been 'winning' have changed.
Atheism and 'rationality' in the populous are not based on much real capability for thought, rather atheism became a cultural norm in a lot of the western world and 'rationality' holds up like wet sand when abandoning the normed tracks
What are the 6 llms you used
DeepSeek ChatGPT Claude Gemini mistral Meta.ai