This lecture, given at the University of Skövde on October 19th, 2006, is an extended version of one I gave in Laval in May ("In defense of transparent computationalism"). The main additions are examples of how the transparent reading of computationalism can save it from some standard anti-computationalist arguments (Gödelian, externalist, dynamical, Chinese room), and mention of the work of Bill Bigge at Sussex as an illustration of how Strong AI might be possible, even if computationalism is false.
I botched an example in the talk, but rectified matters during discussion. The question I meant to ask was "Is the nth sitting-down person's answer to this question not "yes"?", where the only permitted responses are "yes" and not answering. As a standing up person, I can answer this question correctly for all n, while no sitting-down person can (they must not answer when considering their own case).