4 Comments
User's avatar
Neural Foundry's avatar

Compeling framework on transformation sequencing. The comparitive advantage argument for working on earlier transitions is underappreciated in AI safety discourse. I've noticed similar patterns in how organizational change unfolds where initial shifts determine future option spaces more than we typically model. The epistemics-first path particulary intrigues me as precursor infrastructure.

Phil Bell's avatar

Thanks, I found the point about sequencing both useful and thought-provoking. I am exploring all the work you and others are doing on AI for human reasoning, and this nicely articulates the potentially multiplicative impact of an epistemic shield of this kind.

Rainbow Roxy's avatar

Couldn't agree more. It’s so vital to look beyond just the explosion scenario. I often think about this during my Pilates classes; how important it is to understand the small shifts to prepar for bigger movements. This article realy captures that complexity and makes you think deeply about the journey ahead.

The Alfalfa Mail's avatar

AI doomerism godsplained by substack's favourite church roof prognosticator, the Reverend Jaz Pickles

https://open.substack.com/pub/thealfalfamail/p/thought-of-the-day-2?utm_campaign=post-expanded-share&utm_medium=web