What it captures, in my view, is that the bottleneck is no longer knowledge production but trust allocation.
Once AI begins compounding the research layer, the decisive question becomes not who can generate claims, but who controls the systems through which claims are filtered, ranked, and believed.
That is why this is not just an epistemics story.
It is a governance story.
Authority is moving upstream into infrastructure.
The model, the interface, the audit layer, the incentive structure, these start mattering more than the institution that publicly certifies the result.
So trust lock-in is doing more work here than it first appears.
The first dominant epistemic systems may become proto-institutions long before they are recognized as such.
Yes, please examine all links and make sure they work as intended. The article is decent right now, but it would be great if the links to the interesting “design sketches” pages worked.
What it captures, in my view, is that the bottleneck is no longer knowledge production but trust allocation.
Once AI begins compounding the research layer, the decisive question becomes not who can generate claims, but who controls the systems through which claims are filtered, ranked, and believed.
That is why this is not just an epistemics story.
It is a governance story.
Authority is moving upstream into infrastructure.
The model, the interface, the audit layer, the incentive structure, these start mattering more than the institution that publicly certifies the result.
So trust lock-in is doing more work here than it first appears.
The first dominant epistemic systems may become proto-institutions long before they are recognized as such.
They will not simply assist judgment.
They will pre-structure it.
The links in this post are mostly broken.
The links are now fixed; thanks for flagging this!
Yes, please examine all links and make sure they work as intended. The article is decent right now, but it would be great if the links to the interesting “design sketches” pages worked.