I am sharing this post with my grad students! This is one reason why we tend to value research experience more than stats (GPA + GRE) in our program’s admissions. Plenty of students can do the work but not all have the non-academic qualities to go the distance.
Mental resilience yes. I wonder if there's an argument to be made that AI is more resilient than the average doctoral student though. It doesn't really care about Reviewer 2....
Claus, I disagree with your characterization of "fearlessness" as a characteristic of successful PhDs. I do think "fearlessness" (in so far as one is fearless in learning adjacent things) is a good trait, but much like everything too much of a good thing can be bad.
I'm under the impression that the job market for PhDs is limited, and even before accepting a tenure track job one must complete at least one (often times many) low paid post docs. I wouldn't be surprised if the attrition from PhD to full professor is very high. As such, one should prepare one's self for a life outside of academia as best as possible, and that might mean saying "no" to acquiring some set of skills in favor of another (or, just finishing the PhD sooner than later). The "normal computational scientist" in your example might then actually be making the better decision, and might be "successful" in some other equally valid sense.
Much like a good estimator, some regularization for extreme perspectives can be useful. As an example, it would be beneficial to learn R as opposed to sticking with SPSS, but I disagree that spending "a thousand hours of tedious wet-lab experiments" (which is about 125 days working 8 hours a day, or nigh 7 months working only Monday to Friday -- a practice I know few wet lab scientists employ. This not to mention all the other stuff a PhD should be doing, so really the time line is longer than I've written here.) should be valorized, at least without more context.
I'm curious if you share this opinion or if you might encourage a PhD student in some computational discipline to learn wet lab techniques.
You're completely misunderstanding what I'm saying here. The kind of PhD student that professors think about when they're saying "this person is fearless" wouldn't care whether their professor thinks a particular project is a good idea or not. They would do it anyways. In fact, often their professors tell them they're too ambitious. I've seen computational students do exactly what I described in my post, and when they did so everybody in the lab (including the advisor) thought they were crazy, and yet they're now tenured professors and some of the most respected scholars in their field. I use plural because I've seen it more than once.
For somebody who just wants to get a degree as quickly as possible so they can move on and take an industry job I would certainly not recommend that they take on a massively ambitious project. But for some people this is exactly what they want and need. And also, they're typically graduating on time. Students that take 8 years to graduate have other issues, including likely an abusive advisor. All my students graduate in about 5 years, 5.5 if they're slow.
I think you very clearly identify the key qualities needed but form a psychological perspective you overindex on specific sources that might inhibit that. For example you write (clearly identifying a key quality): "Yes people in PhD programs are motivated by knowledge acquisition and learning, and yet even they don’t really want to learn things completely outside of their comfort zone.". But you tie it to fearlessness, and fear. Its not clear to me fear is the only, or even the main source of failure here. Its as much about locus of control (a believe in which things you can, and cannot control) or about the time horizon across which people weight options e.g. "I could do an okay computational thing now and I'd be guaranteed a return, the alternate is doing a stellar wet lab thing with questionable but potentially far higher payoff in the distant future". Undferlying what you describe as fearlessness is (IMO) a secure personal identity detached from becomign an academic, whicha allow peopel to take risks by not letting potrntial immidiate failures, or delays in any payoff impact their identity. You can spot patterns like this in degress where the PhD is a requirement for a further professional degree (clincial psychology), students can become very focused on the immidiate, an adequate/good thesis is essential as a delayed these means they miss placement for the obligatory clinical year or are placed in an unfavourable sport(literally across the country). Once policy ties the thesis to real life outcomes, or even people's prefered future careers/identities you see defensive choices (which arent fearfull, just optimal given policy). I may have drawn this deep into psychology but I feel being clear helps us shape the (academic/policy) environment for others to be what you describe as "fearless".
I am sharing this post with my grad students! This is one reason why we tend to value research experience more than stats (GPA + GRE) in our program’s admissions. Plenty of students can do the work but not all have the non-academic qualities to go the distance.
Yes, it's difficult to assess who has the right personality to succeed. Good grades or test scores don't mean much.
Mental resilience yes. I wonder if there's an argument to be made that AI is more resilient than the average doctoral student though. It doesn't really care about Reviewer 2....
That's correct. It also doesn't care about the scientific problem or not making up stuff.
The question is how do you build an AI that deeply cares about the things that matter but that can completely ignore the things that don't.
Claus, I disagree with your characterization of "fearlessness" as a characteristic of successful PhDs. I do think "fearlessness" (in so far as one is fearless in learning adjacent things) is a good trait, but much like everything too much of a good thing can be bad.
I'm under the impression that the job market for PhDs is limited, and even before accepting a tenure track job one must complete at least one (often times many) low paid post docs. I wouldn't be surprised if the attrition from PhD to full professor is very high. As such, one should prepare one's self for a life outside of academia as best as possible, and that might mean saying "no" to acquiring some set of skills in favor of another (or, just finishing the PhD sooner than later). The "normal computational scientist" in your example might then actually be making the better decision, and might be "successful" in some other equally valid sense.
Much like a good estimator, some regularization for extreme perspectives can be useful. As an example, it would be beneficial to learn R as opposed to sticking with SPSS, but I disagree that spending "a thousand hours of tedious wet-lab experiments" (which is about 125 days working 8 hours a day, or nigh 7 months working only Monday to Friday -- a practice I know few wet lab scientists employ. This not to mention all the other stuff a PhD should be doing, so really the time line is longer than I've written here.) should be valorized, at least without more context.
I'm curious if you share this opinion or if you might encourage a PhD student in some computational discipline to learn wet lab techniques.
You're completely misunderstanding what I'm saying here. The kind of PhD student that professors think about when they're saying "this person is fearless" wouldn't care whether their professor thinks a particular project is a good idea or not. They would do it anyways. In fact, often their professors tell them they're too ambitious. I've seen computational students do exactly what I described in my post, and when they did so everybody in the lab (including the advisor) thought they were crazy, and yet they're now tenured professors and some of the most respected scholars in their field. I use plural because I've seen it more than once.
For somebody who just wants to get a degree as quickly as possible so they can move on and take an industry job I would certainly not recommend that they take on a massively ambitious project. But for some people this is exactly what they want and need. And also, they're typically graduating on time. Students that take 8 years to graduate have other issues, including likely an abusive advisor. All my students graduate in about 5 years, 5.5 if they're slow.
I think you very clearly identify the key qualities needed but form a psychological perspective you overindex on specific sources that might inhibit that. For example you write (clearly identifying a key quality): "Yes people in PhD programs are motivated by knowledge acquisition and learning, and yet even they don’t really want to learn things completely outside of their comfort zone.". But you tie it to fearlessness, and fear. Its not clear to me fear is the only, or even the main source of failure here. Its as much about locus of control (a believe in which things you can, and cannot control) or about the time horizon across which people weight options e.g. "I could do an okay computational thing now and I'd be guaranteed a return, the alternate is doing a stellar wet lab thing with questionable but potentially far higher payoff in the distant future". Undferlying what you describe as fearlessness is (IMO) a secure personal identity detached from becomign an academic, whicha allow peopel to take risks by not letting potrntial immidiate failures, or delays in any payoff impact their identity. You can spot patterns like this in degress where the PhD is a requirement for a further professional degree (clincial psychology), students can become very focused on the immidiate, an adequate/good thesis is essential as a delayed these means they miss placement for the obligatory clinical year or are placed in an unfavourable sport(literally across the country). Once policy ties the thesis to real life outcomes, or even people's prefered future careers/identities you see defensive choices (which arent fearfull, just optimal given policy). I may have drawn this deep into psychology but I feel being clear helps us shape the (academic/policy) environment for others to be what you describe as "fearless".