@ErikJonker My problem is with "believe" (and the misuse of the word "algorithm").
Are there problems in any field of AI that can't be effectively computed on standard computers, but it's at least plausible that effective quantum algorithms can be constructed for them (as in Shor vs. factorization)?