Math / physics / cs phd worth it?

Deloitte diracdelta
Aug 6, 2018 20 Comments

Been working in AI product development for past 3 years. Need more growth and want to be involved in the cutting edge of research, or working at a place like DeepMind.

I love math and physics (majored in math, read dozens of physics textbooks for fun) but I’ve always heard that a phd in those subjects don’t pay, even from top schools/advisors. I was hoping to hear some counterexamples with a dose of reality.

comments

Want to comment? LOG IN or SIGN UP
TOP 20 Comments
  • Microsoft / Eng Tier 1
    PhDs are something you do out of personal interest, prestige, and for the good of humanity.

    Don’t do a PhD for the money. You’ll be disappointed.
    Aug 6, 2018 2
    • Deloitte diracdelta
      OP
      I like that , I should think more in terms of impact over money
      Aug 6, 2018
    • Microsoft Yiwb71
      Me too. But are you sure a PhD will give you the best opportunity to make an impact?
      Aug 6, 2018
  • Agilent Technologies / Design 031689
    Don't do it unless you need it. The PhD process is grueling and will almost certainly break you down as part of the process. I did it out of necessity to work in my field but if I didn't have to do it, I wouldn't have. The nice thing is that I get to be called "Doctor ___" every once in a while but it's mostly just used as an insult.
    Aug 6, 2018 2
    • Deloitte diracdelta
      OP
      Oh wow that’s a different perspective from everyone else who replied. Good to know.
      Aug 6, 2018
    • Agilent Technologies / Design 031689
      Most people will give you some bullshit response about "pursuit for knowledge sake" but whatever, some career fields need it. In my field, we're ultimately just really clever technicians
      Aug 6, 2018
  • Lyft 2hia5b
    Not worth. some of the most common deep learning algorithms were *rejected* as PhD thesis coz some asshat wasn't feeling it. Academia is not science driven.
    Aug 6, 2018 3
    • Northrop Grumman / Eng gaussend
      Don't you think a lot if those algos are kinda emipirically derived with some loose theoretical backing?
      Aug 6, 2018
    • Agilent Technologies / Design 031689
      To most people who analyze actual experimentally derived data, the most common deep learning algorithms ARE garbage. If you take a skilled scientist, they could come up with a better conclusion out of pure intuition.
      Aug 6, 2018
    • Lyft 2hia5b
      Dude stop writing, my deep learning algorithm already predicted your comment
      Aug 7, 2018
  • Oracle / Eng Ndhhhs
    Earning a Phd is an achievement. Just like earning 1M is another kind of achievement. Dont do it for money. Do it for what you want to achieve.
    Aug 6, 2018 0
  • Microsoft DOyu71
    Well, the rank and file dont get payed too well compared to tech but have a look at your favorite state's employee salary database and you'll find a stack of professors from large schools making $1M or more. That doesnt mean that you will, and as an assistant professor you won't be making much, but if that's your passion, go for it.
    Aug 6, 2018 0
  • Autodesk / Eng
    mCyn20

    Autodesk Eng

    PRE
    Autodesk
    mCyn20more
    They dont pay well now, but i would think they will soon have to
    Aug 6, 2018 4
    • Deloitte diracdelta
      OP
      Why would that be?
      Aug 6, 2018
    • Autodesk / Eng
      mCyn20

      Autodesk Eng

      PRE
      Autodesk
      mCyn20more
      The amount of workforce that is needed in specialized fields with an understanding of coding is going to be very much in demand because of the technologies we would like to explore in the coming years
      Aug 6, 2018
    • Deloitte diracdelta
      OP
      Would there be a huge demand for this much education though? It seems like it would be more for engineers with experience. You would only need one phd on a team I would guess
      Aug 6, 2018
    • Autodesk / Eng
      mCyn20

      Autodesk Eng

      PRE
      Autodesk
      mCyn20more
      Unfortunately one phd on a team is not going to be sufficient in the future
      Aug 6, 2018
  • Intel BobbySwan
    Had a Physics PhD friend enter Goog as L4. Your CS PhD would have to include some non-trivial papers in the field of AI to get a chance at FAIR or DeepMind.
    Aug 6, 2018 1
    • Deloitte diracdelta
      OP
      That makes sense, though they have AI residencies that help
      Aug 6, 2018
  • Northrop Grumman / Eng gaussend
    If you want to do research, you have to get it Imo. If you're interested in analysis and theory more so than practical stuff, it's one of the only ways to get paid (poorly) to do so for 4 to 7 years.

    If you want to chase money and are happy coding, then probably just stay the course.
    Aug 6, 2018 0