AIXI is a dualistic agent that can’t work as an embedded agent… right? I couldn’t find a solid formal proof of this claim, so I investigated it myself (with Marcus Hutter). It turns out there are some surprising positive and negative results to be derived as easy corollaries of the paper “Universal Prediction of Selected Bits.” Interestingly, further technical advances in algorithmic information theory could substantially strengthen our results—I would welcome collaborations with strong theoretical computer scientists, (deep familiarity with agent foundations not required).
This work was supported by the Long-Term Future Fund and presented at the CMU agent foundations conference in 2025.
Formalizing Embeddedness Failures in Universal Artificial Intelligence
Link post
AIXI is a dualistic agent that can’t work as an embedded agent… right? I couldn’t find a solid formal proof of this claim, so I investigated it myself (with Marcus Hutter). It turns out there are some surprising positive and negative results to be derived as easy corollaries of the paper “Universal Prediction of Selected Bits.” Interestingly, further technical advances in algorithmic information theory could substantially strengthen our results—I would welcome collaborations with strong theoretical computer scientists, (deep familiarity with agent foundations not required).
This work was supported by the Long-Term Future Fund and presented at the CMU agent foundations conference in 2025.