Describes a scaled down AIXI agent that uses Monte Carlo Tree Search.

Gives a parameter-free universal similarity distance, the normalized compression distance (NCD), shows that it approximates optimality and explains how it can be used to extract a hierarchy of clusters from arbitrary domains with unknown features.

Universal prediction, i.e. Bayesian framework + Solomonoff prior + model class of all computable sequences, solves/avoids foundational problems of inductive inference, but has to be compromised in practice.

An algorithm M is described that solves any well-defined problem p as quickly as the fastest algorithm computing a solution to p, save for a factor of 5 and low-order additive terms.

The AIXI model is the most intelligent unbiased agent possible (though uncomputable). A modified version, AIXItl, is still effectively more intelligent than any other time t and space l bounded agent.