Quantum annealers (like DWave) can solve optimization problems efficiently. But they can’t run general quantum algorithms such as Shor’s algorithm, that requires a general quantum computer (like what IBM and Google are doing).
can possibly solve some (but without any guarantees on which or if any actually exist)
did it reach the point where it can do any useful work faster than standard computer? Last time I read about it it was something stupid like "It can simulate its own noise"