A personal data point: I have a PhD and looking back at it from 10+ years later I would say that it was a good thing for me because: A) I enjoyed it (most of the time) and B) I learned a lot.
As far as formally completing a degree (i.e. getting a printed piece of paper that says Ph.D. next to my name) I find it practically irrelevant for either getting hired in the industry or being useful once hired. My 2c.
I know this is addressed at sama, but as someone doing AI without a PhD:
1. Not at all. You need to develop a consistent work schedule that will consist of researching papers, applying the knowledge, and writing your own. There is a lot of theory and practice. Very exciting time to be in AI.
What you need is time and the interest in doing it. Feel free to email me if you have questions (not a sales pitch, just being friendly).
It's arguable that AI as a research discipline is in its childhood, thus, there are still plenty of low-hanging fruit to be plucked by intelligent, dedicated people.
This is similar to the state of physics in the 18th century. Significant discoveries were made by gentlemen researchers.
Today it is very difficult to advance the boundaries of older, more well-established scientific disciplines like physics or biology without a high level of training--if not a doctorate specifically, its equivalent in education and training. AI will get there eventually as well.
IMO software in general is in its infancy. We still build important computer programs the same way gunsmiths made guns in the 18th century--by hand, using intuition and art as much as science.
How much advantage (if any) you estimate a person holding this degree will have over other that doesn't? In short, although you said it's not necessary, do you think it's worthy?
IMO PhD is useful as a proxy for people who are intellectually curious and are able to dig into a problem until they reach the bottom of the rabbit hole.
I personally think these same skills can be found outside of PhDs.
Here are my questions:
1. Do you think people need to get a PhD degree to become useful research scientist in AI industry?
2. Do you think people need to get a PhD degree to become Member of Technical Staff (Machine Learning) in Open AI?
Thank you.