jhm
Well-Known Member
I am someone who got a PhD and then did 4 years of postdoctoral work in statistics. I also have more than 20 years as a practicing statistician in business. In my experience, deep learning does not happen that much as an undergraduate. Even as a graduate student, it is not until you do original research for a thesis or later that the real deep learning happens. I highly value the 4 years I spent as a post-doc, as that has given me a distinct edge within a business context. It has given me much more flexibility, creativity and independence in how I approach problem. To say that experience with highly specialized research is not valuable fails to understand just what one gains in the process. In my experience, it has paid dividends through my whole career.In physics postgraduates are typically already highly specialized and good PhD students are spending most of their time working on a particular research topic.
The "deep learning" is mostly done by that time and if they are good they never stop learning for the rest of their careers, and I think Elon is good.
In that sense I think leaving the PhD track was a net benefit for him, because it broadened his experience and his perspective.
Perhaps Elon could say the same for experience he got building Zip2. But I do think if he had done doctoral level research, he'd be making very good use of that experience today. My point is not to discredit Musk, who is quite extraordinary in many ways, but simply to affirm that there is lasting value in doctoral level research.