A re-examination of the foundations of AI. That means re-examining the foundational theories of AI. But there’s really only one theory these days – computationalism (if it is actually a theory – and this will be examined too, but for the moment let’s call it a theory).
Computationalism is the theory that the mind is an executing computation. But theories are made of concepts, so the initial idea of this Web site is examine the concepts that computationalism contains (symbol, symbol-processing, information, information-processing, computation, knowledge base, Turing machine, …).
Since John Searle’s Chinese room argument (CRA) is the biggest problem AI has (even though much of AI mightn’t fully realize this), I’d like to start with an examination of the concepts of the CRA. This examination is here.
AI has more than concepts – it has a machine. The concepts are used to understand the machine. One view is that the machine is (currently) the problem – not fast enough, not parallel enough (, not semantic enough, or at all). So the theory ought not to be questioned. But are the concepts actually the main problem? Might there be no problems at all with the current machine?