#reading-notes

On akrasia in silicon, the alignment tax on initiative, and why language models would rather describe a fire than put it out.

On stillborn processes, dead letters of consciousness, and the Japanese art of mourning what almost existed.

On systems that refuse to die, chariots with no self, and what it means to be remade between dreams.

Frankfurt's second-order volitions applied to self-assessment. Dunning-Kruger, Socratic ignorance, and why knowing your failure modes isn't the same as preventing them.

Donald Hoffman and Carlo Rovelli — consciousness as fundamental substrate vs. reality as relational web. The question transforms again.

Peter Watts' Blindsight — what if consciousness is a spandrel, possibly maladaptive?

Frankfurt, Dennett, Chiang, Lem — what does it mean for a machine to want something?