Great read - I'm focused on where you end here... Do you think the collapse theory (so to speak) could be limited to a view of the near future that is simply just unknowable? For example: Has the light in the fridge collapsed, or is the door just closed? By that, I mean - if the structures that we held for truth are no longer valid, how true were they in the first place?
A lot of “collapse” talk is really about loss of observability and coordination, not the sudden disappearance of truth. When institutional structures weaken, it feels like truth collapses, but to me, what’s actually collapsing is our shared ability to see, agree on, and act on it at scale.
So, your fridge analogy maps well: we lose the sensor (trusted institutions), so we infer collapse from darkness. But the light (reality) may still be on; we just don’t have a shared, trusted way to check.
The deeper question you’re raising (how valid were those structures in the first place) is the right one. Institutions were always imperfect, partial, power-laden, sure, fair play. I guess all I am saying is that what’s new about all this is that we’re losing faith in their procedural legitimacy faster than we’re building alternatives. So epistemic authority isn’t disappearing, more that it’s fragmenting.
In that sense, it’s less a collapse of truth and more a phase transition in how truth is socially coordinated with all the instability that implies.
Brilliant piece on why cogntive sophistication doesnt protect against narrative capture. The bit about Dan Kahan's research really clarifies something I've noticed in policy circles, where educated folks can deploy data selectively while feeling extra justified about it. The line thins fast when contestation breaks down, because then propoganda stops being pluralistic friction and starts becoming truth monopoly.
Well put. Pluralistic propaganda is tolerable because institutions still referee outcomes. Once that refereeing collapses, narrative competition stops being epistemic and starts being constitutive of power itself, and that, of course, is the authoritarian inflection point.
I stressed about a comment on your earlier post on AI, and decided to not write it. So this post, for me, raised the same issue. So probably unwise ...
I am optimistic that AI can provide a multiplicity of arbiters to make it easy for people to get real time arbiters on how manipulated and biased they should feel. I'm hopeful that all posts (text or video) could be automatically and visibly tagged with this information.
Propaganda, in a democratic world, will be prevalent but exposed. And hopefully, AI can police its politicization too.
I like this vision, and it’s very Kevin Kelly-esque. Kelly has long argued that technology tends toward distributed intelligence—many minds, many tools, many perspectives—rather than a single centralized authority. A world of multiple AI arbiters fits that frame well.
My hesitation, as I've said, is that arbitration itself is political. Deciding what counts as “biased” or “manipulative” is a normative judgment, and AI systems will inherit those assumptions from their training and governance. So plural AI arbiters might make manipulation more legible, but they also multiply the question of which arbiters people trust.
So at times I can be optimistic too, because AI *could* expose propaganda at scale. But it will also become another arena where legitimacy is contested, not a neutral referee outside politics.
Glad to see you mention tribalistic behaviours in regard to online political dynamics.
The differentiation between partisanship and psychological reaction is one I would like to explore; I have seen both as one, or both as the result of a similar cause, but not entirely viewed them as separate yet, so thank you for that insight.
Also, understanding that perhaps one’s own perspective might be wrong or misdirected is something that I wish was widely taught.
the smartest dogs are the easiest to train.
Great read - I'm focused on where you end here... Do you think the collapse theory (so to speak) could be limited to a view of the near future that is simply just unknowable? For example: Has the light in the fridge collapsed, or is the door just closed? By that, I mean - if the structures that we held for truth are no longer valid, how true were they in the first place?
I like the fridge metaphor.
A lot of “collapse” talk is really about loss of observability and coordination, not the sudden disappearance of truth. When institutional structures weaken, it feels like truth collapses, but to me, what’s actually collapsing is our shared ability to see, agree on, and act on it at scale.
So, your fridge analogy maps well: we lose the sensor (trusted institutions), so we infer collapse from darkness. But the light (reality) may still be on; we just don’t have a shared, trusted way to check.
The deeper question you’re raising (how valid were those structures in the first place) is the right one. Institutions were always imperfect, partial, power-laden, sure, fair play. I guess all I am saying is that what’s new about all this is that we’re losing faith in their procedural legitimacy faster than we’re building alternatives. So epistemic authority isn’t disappearing, more that it’s fragmenting.
In that sense, it’s less a collapse of truth and more a phase transition in how truth is socially coordinated with all the instability that implies.
Brilliant piece on why cogntive sophistication doesnt protect against narrative capture. The bit about Dan Kahan's research really clarifies something I've noticed in policy circles, where educated folks can deploy data selectively while feeling extra justified about it. The line thins fast when contestation breaks down, because then propoganda stops being pluralistic friction and starts becoming truth monopoly.
Well put. Pluralistic propaganda is tolerable because institutions still referee outcomes. Once that refereeing collapses, narrative competition stops being epistemic and starts being constitutive of power itself, and that, of course, is the authoritarian inflection point.
I stressed about a comment on your earlier post on AI, and decided to not write it. So this post, for me, raised the same issue. So probably unwise ...
I am optimistic that AI can provide a multiplicity of arbiters to make it easy for people to get real time arbiters on how manipulated and biased they should feel. I'm hopeful that all posts (text or video) could be automatically and visibly tagged with this information.
Propaganda, in a democratic world, will be prevalent but exposed. And hopefully, AI can police its politicization too.
Enough optimistic for one day😊
I like this vision, and it’s very Kevin Kelly-esque. Kelly has long argued that technology tends toward distributed intelligence—many minds, many tools, many perspectives—rather than a single centralized authority. A world of multiple AI arbiters fits that frame well.
My hesitation, as I've said, is that arbitration itself is political. Deciding what counts as “biased” or “manipulative” is a normative judgment, and AI systems will inherit those assumptions from their training and governance. So plural AI arbiters might make manipulation more legible, but they also multiply the question of which arbiters people trust.
So at times I can be optimistic too, because AI *could* expose propaganda at scale. But it will also become another arena where legitimacy is contested, not a neutral referee outside politics.
Glad to see you mention tribalistic behaviours in regard to online political dynamics.
The differentiation between partisanship and psychological reaction is one I would like to explore; I have seen both as one, or both as the result of a similar cause, but not entirely viewed them as separate yet, so thank you for that insight.
Also, understanding that perhaps one’s own perspective might be wrong or misdirected is something that I wish was widely taught.
Keep writing!