Turing machines have a well-defined input, and output if they halt. So no, they are absolutely not useless, they are just "single-shot" models of computation.

Turing machines have a well-defined input, and output if they halt. So no, they are absolutely not useless, they are just "single-shot" models of computation. Certain software fit that model very nicely (e.g. compilers), others less so.