"Bayes is a very limited special case of the replicator equation"

Traits shared by all human cultures, without known exceptions.

A computer-based tournament that tries to answer the question: Which type of learning is most effective in a complex, changing world?

On incremental vs. paradigm change both in science and in evolution.

"What I believe is happening now is that true teme machines are arriving -- that is, machines that copy and produce variations and then select. That's what you need for an evolutionary process [independent of human beings]; that's natural selection."

Complex systems on the boundary between order and chaos may be those best able to adapt by mutation and selection.

Suggests to replace the current model of connecting software components (protocol adherence, chaotic failure modes) by pattern recognition (graceful error-tolerance).

A quantitative theory that treats evolution as a form of computational learning. Doing away with the distinction between learned and inherent (evolved) knowledge, cognitive systems can be viewed as pure learning systems.

A self-replicating machine in a cellular automata environment, designed in the 1940s without the use of a computer.

Construction of a chemical system capable of replication and evolution, fed only by small molecule nutrients, is now conceivable.

Argues that it is unlikely that complexity has any useful value as applied to "real" objects or systems and that even relativising it to an observer has problems. Proposes that complexity can usefully be applied only to constructions within a language.

AutoCore is a virtual machine for general purpose evolutionary computation based on recent advances in digital artificial life pioneered by the academic community for the study of theoretical evolutionary biology and ecology.

"There are really two big promising directions in machine learning/AI today: evolutionary computation and brain reverse-engineering. Some readers might be curious as to why I'm working on evolutionary computation and not neuroscience-based approaches."

How evolution tunes the content and frequency of genetic variation to enhance its evolvability.

"I'm only interested in ~human equivalent general AI, and only in co-evolution from a reasonable seed pool in a superrealtime virtual environment heavily skewed towards problem-solution as fitness function as a design principle."

Darwin is an automated theorem prover for first order clausal logic and the first implementation of the Model Evolution Calculus.

An upper bound on the speed of evolution is derived. The bound concerns the amount of genetic information which is expressed in observable ways in various aspects of the phenotype. Typically it can't increase faster than a few bits per generation.

Für den Giessener Verhaltensbiologen Eckart Voland ist „Fortschritt“ der Evolution wesensfremd, nur wir Menschen hätten ihn strategisch als Idee konstruiert, um uns für die „Tretmühle des Lebens“ zu motivieren.

"What a strange thing a human must be - an intelligence inscribed entirely by evolution, a mind that wasn't constructed by another mind. We are one of the oddest and most awkward sights in the universe."