Leaving for the industry. That's where they're all going. Grants are only a small part of the problem. I'm a math phys PhD candidate at a top school, all my cohort could get academic jobs if they want. Doesn't matter, well over 50% of them leave. Even ignoring the fact that academia pays like a tenth of what we'd make in the private sector, we also have to deal with a stupid amount of teaching duties, inane bureaucratic hoop-jumping, administrative bloat and grift, and yes, the whole grant nonsense. The fact is that if a PhD in physics left for quantitative finance right this minute, they'd be treated with infinitely more respect than anything academia can give them, and that's ignoring the financial aspect of it.
As for why science and tech is getting slower, I don't know if the general thesis is true (how are you even measuring levels of tech here?) but certainly a lot of frontier science hinges on two big, connected issues: (1) the technical machinery needed to make substantial progress in many fields now itself takes years to master, which not many academics are willing to do; and (2) the level of abstraction required for the most frontier of frontier work is getting so challenging that the totality of what Einstein through Feynman knew about math and physics is now considered basic, and the kind of black magic being done here would be challenging to even the most talented theoretician.
On the other hand, academia is also getting wider: as soon as machine learning became possible (there was a hardware barrier in the late 90s that prevented the earliest ML papers from being implemented), we suddenly saw a lot of new, low-hanging fruit to pick up, which is still the case in ML. Just look at the example of diffusion models. Their equivalent in statistics dates back to maybe even the 70s and 80s, never mind their equivalents in math and physics. When did they get implemented in ML? Half a decade ago?
I don't know how to predict the pace of science, I don't even have a grasp on its current pace, other than that yes, scientific journalism is so stupid that I can hardly blame the public for thinking that nothing important has happened. But contra what some people like Hossenfelder might suggest, I don't think physics is in a rut. Maybe empirical particle physics. We aren't picking up anything that's as monolithic in public consciousness like Einstein's relativity, but we have plenty of math and physics work today that are every bit as intellectually and practically dense. Of course nothing has the same oomph, but well, nothing has the same oomph as Caveman Grug discovering how to count, and we don't say that scientific progress has been declining since Grug.
p
Leaving for the industry. That's where they're all going. Grants are only a small part of the problem. I'm a math phys PhD candidate at a top school, all my cohort could get academic jobs if they want. Doesn't matter, well over 50% of them leave. Even ignoring the fact that academia pays like a tenth of what we'd make in the private sector, we also have to deal with a stupid amount of teaching duties, inane bureaucratic hoop-jumping, administrative bloat and grift, and yes, the whole grant nonsense. The fact is that if a PhD in physics left for quantitative finance right this minute, they'd be treated with infinitely more respect than anything academia can give them, and that's ignoring the financial aspect of it.
As for why science and tech is getting slower, I don't know if the general thesis is true (how are you even measuring levels of tech here?) but certainly a lot of frontier science hinges on two big, connected issues: (1) the technical machinery needed to make substantial progress in many fields now itself takes years to master, which not many academics are willing to do; and (2) the level of abstraction required for the most frontier of frontier work is getting so challenging that the totality of what Einstein through Feynman knew about math and physics is now considered basic, and the kind of black magic being done here would be challenging to even the most talented theoretician.
On the other hand, academia is also getting wider: as soon as machine learning became possible (there was a hardware barrier in the late 90s that prevented the earliest ML papers from being implemented), we suddenly saw a lot of new, low-hanging fruit to pick up, which is still the case in ML. Just look at the example of diffusion models. Their equivalent in statistics dates back to maybe even the 70s and 80s, never mind their equivalents in math and physics. When did they get implemented in ML? Half a decade ago?
I don't know how to predict the pace of science, I don't even have a grasp on its current pace, other than that yes, scientific journalism is so stupid that I can hardly blame the public for thinking that nothing important has happened. But contra what some people like Hossenfelder might suggest, I don't think physics is in a rut. Maybe empirical particle physics. We aren't picking up anything that's as monolithic in public consciousness like Einstein's relativity, but we have plenty of math and physics work today that are every bit as intellectually and practically dense. Of course nothing has the same oomph, but well, nothing has the same oomph as Caveman Grug discovering how to count, and we don't say that scientific progress has been declining since Grug.
More options
Context Copy link