Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Physicists taking jobs as Silicon Valley software engineers

wired.com

Post reply on HN
Print view View original

141–150 of 178 posts

Re: Physicists taking jobs as Silicon Valley software engineers

#141

Post by niels_olson »

FlorianRappl wrote: post #78
ejrowley wrote: post #31

I studied physics, with intention of becoming a developer (which I did). I didn't have much appetite for a CS course and felt the physics would give me a broader set of skills.

I followed a similar path and can only agree on the attitude. Most of the relevant CS knowledge is just that: knowledge. Studying physics is usually less about knowledge and more about general problem solving skills.

I'd go beyond "general problem-solving skills". Physics gives you high quality math training (albeit somewhat truncated with respect to number theory), coupled with demanding experimental methods, and a deep, principled understanding of the foundations of the universe. Nothing is beyond your reach. Any other quantitative or logical discipline is just laying there, waiting for you.

Re: Physicists taking jobs as Silicon Valley software engineers

#142

Post by fnbr »

sjg007 wrote: post #104
fnbr wrote: post #42

Earlier quoted context omitted.

I did a master's in econometrics and that was my experience. In the last year of my undergrad, I took a course in machine learning that seemed to have way more intellectual rigor than any of the material covered in grad school. At one point, I mentioned to the professor that I was concerned that the model he had presented was overfitting, and he had no understanding of what the term meant. I think that economics stud…

With deep learning overfitting has been thrown out of the window..

Obviously overfitting is always a concern, but deep learning has (when done correctly) found ways to mitigate concerns about overfitting.

Dropout, for instance, or k-fold cross validation.

Re: Physicists taking jobs as Silicon Valley software engineers

#143

Post by rpcope1 »

koolba wrote: post #51
physcab wrote: post #22

Earlier quoted context omitted.

I was a Physics undergrad+ Physics / Materials Science PhD and I ended up just quitting 4 years into my PhD. You can do it too. You don't owe your advisor anything and you've definitely put in your indentured servitude. I ended up getting a job as a data analyst / scientist and never looked back. You can work on interesting problems, use the same experimental design knowledge and methods, and get paid $100k more than…

Four years? Is it common for a Physics PhD to take that long? I thought most fields are closer to two years.

For many in experimental tracks, when I was in graduate school for a physics PhD, 4 years seemed like the shortest you could get out in. There were a lot of people that started immediately at 22 and were almost 30 by the time they were ready to graduate.

Re: Physicists taking jobs as Silicon Valley software engineers

#144

Post by throwaway61AR »

amai wrote: post #125

While being a student a physics professor told me: "Study physics now. Because later when you have a job, you might get paid to learn using some software tool or a new programming language. But no company will ever send you into a quantum mechanics lecture. " He was right.

No he wasn't. At the google X labs we had a quantum mechanics lecture from a guest speaker, literally only a couple of months back.

Re: Physicists taking jobs as Silicon Valley software engineers

#145

Post by eanzenberg »

throwaway61AR wrote: post #144
amai wrote: post #125

While being a student a physics professor told me: "Study physics now. Because later when you have a job, you might get paid to learn using some software tool or a new programming language. But no company will ever send you into a quantum mechanics lecture. " He was right.

No he wasn't. At the google X labs we had a quantum mechanics lecture from a guest speaker, literally only a couple of months back.

1 lecture at arguably undergrad level vs. 30 lectures (2 semesters) of grad level qm

Re: Physicists taking jobs as Silicon Valley software engineers

#146

Post by Koshkin »

The effort one has to put into getting a physics degree seems to me comparable to that of becoming a physician (a medical doctor), and it pains me to see all that effort wasted. Some argue that it is the acquired skill in solving problems that matters, but, having seen physicists and mathematicians (and even "computer scientists") by training trying to solve software engineering problems, I have to disagree. Likewise, from what I have also observed, problem solving skills of a good software engineer may not be by themselves all that useful when trying to tackle problems in physics and mathematics.

Re: Physicists taking jobs as Silicon Valley software engineers

#147

Post by throwaway61AR »

montecarl wrote: post #135
deepnotderp wrote: post #47

Nitpick: newton's method is intractable in Deep learning since it's computationally intractable to compute the hessian matrix

Newton's method might be, but quasi-Newton methods like L-BFGS[1] are usable if you can calculate the gradient of your problem and have enough memory to store several previous coordinates and gradients (so maybe 10x the memory requirement of gradient descent). [1] https://en.wikipedia.org/wiki/Limited-memory_BFGS

Yep this was used by Jeff Dean et al. in one of Google Brain's earlier experiments https://research.google.com/archive/large_deep_networks_nips...

Re: Physicists taking jobs as Silicon Valley software engineers

#148

Post by Koshkin »

ythn wrote: post #14
circlefavshape wrote: post #12

Earlier quoted context omitted.

/me raises hand Degree in physics, working software for 20 years and still feel like an impostor ... not that I remember any physics at this stage

Impostor syndrome seems bogus but only because I've never experienced it and therefore can't understand it. Why do you feel like an impostor? The world is analog, not digital. There's no rite of passage that will suddenly switch you from being an "impostor" to being "legitimate". I've met EE majors who turned out to be better software developers than CS majors and vice versa.

> The world is analog, not digital.

While true as far as our current mathematical models are concerned, there are good reasons to expect that deep down the world's fabric is discrete. (Whether that counts as being "digital" is probably just a matter of language.)

Re: Physicists taking jobs as Silicon Valley software engineers

#149

Post by analog31 »

NumberSix wrote: post #62
crazypyro wrote: post #52

Earlier quoted context omitted.

For anyone curious, here's a one year phd employment survey for physics: https://www.aip.org/sites/default/files/statistics/employmen... 803 responses, 297 in potentially full-time positions. 62% (of all physics phds employed at potentially full-time positions) in non-physics field, break down by industry: Engineering - 20% Computer software - 14% Business or Finance - 11% Other sciences - 8% Education - 2% Medical s…

The use of the phrase "potentially permanent positions" in the AIP report for industry positions, especially in the computer industry, is highly misleading. Academia, including some government labs and institutes, has tenured positions and other positions with guarantees of job security until retirement. The vast majority of industry employees are "at will" and can be laid off at any time for any reason or no reason.…

I think "potentially permanent" is just a historical oddity. They used the same term 25 years ago when I was in school, and there was still an expectation of people going into tenure track position.

Re: Physicists taking jobs as Silicon Valley software engineers

#150

Post by andrepd »

NumberSix wrote: post #43

This migration of physicists to the Silicon Valley and computer industry is not new. It has been true at least since the first big physics employment bubble crashed in the late 1960's, early 1970's. The post-Sputnik boom in physics degrees and grad students produced a huge surplus of physicists by the late 1960's. Dennis Ritchie started at Bell Labs in 1967. Back in the 60's, 70's, early 80's a fair number of physici…

>The heyday of FORTRAN in physics was a long time ago.

Ahh, how I wish that were true...

Post reply on HN