PhDs are over-valued/hyped for ML, especially as it is currently the practice of brute forcing undergraduate maths via CS fundamentals.
ML is a creative field of algorithm composition, and the hardest part is the actual programming. These are not the typical skills of a PhD student.
The current pace, and variegate provenance, of ML research suggests that a PhD student is going to be, at best, a scholar of the web wasting too much time on academics.
At least one of two things will happen in ML: there will be a singular advance in theory that opens up a world of practical applications (like QM), xnor the field of ML will be open to endless synthesis (like programming and music).
All that said, no doubt a PhD will pay for itself...
ML is a creative field of algorithm composition, and the hardest part is the actual programming. These are not the typical skills of a PhD student.
The current pace, and variegate provenance, of ML research suggests that a PhD student is going to be, at best, a scholar of the web wasting too much time on academics.
At least one of two things will happen in ML: there will be a singular advance in theory that opens up a world of practical applications (like QM), xnor the field of ML will be open to endless synthesis (like programming and music).
All that said, no doubt a PhD will pay for itself...