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

Don't Become a Scientist (1999)

yangxiao.cs.ua.edu

Post reply on HN
Print view View original

81–90 of 119 posts

Re: Don't Become a Scientist (1999)

#81

Post by laurencerowe »

jillesvangurp wrote: post #78

I did my Ph. D. between 1998 and 2003. I did not regret doing that but after one year as a post doc I left the academic world for one simple reason: I realized the career paths were limited to two options I did not like: option 1) try to shoot for a tenure position somewhere via a long series of post doc positions. Expect to be mostly busy managing postdocs and researchers and lobbying for funding by the time you get…

I’ve never done a PhD (though have worked as a software developer in a research lab) but I feel I’ve had the most freedom to push the state of the art in the tech I work on in open source while funding myself through consulting.

My fear of going the startup route is that it ends up more like being a PI and having to chase funding and so much of the success lies less in the tech and more in the execution of how you sell it.

So go for funding and get the money to hire a team to do the fun stuff or do the fun stuff yourself on the side…

Re: Don't Become a Scientist (1999)

#82

Post by TrackerFF »

Looking back, we only had two guys in our class that went for Ph.D - both ended up doing a Ph.D in ML, coming from electrical engineering. This was back when ImageNet changed the scene. Both were extremely sharp. One of them ended up leaving for industry after 5 years postdoc, the other publishes like a machine, had stints at some really prestigious schools (Stanford, CMU) - but as far as I can find, still postdoc 10 years after getting his Ph.D

So while this is only N=2 sample size / anecdotal, it would seem to me that even in red hot fields like AI/ML you'd be better off finding some good company / private research group that matches your interest, and go for that. Probably going to get paid 5 times more, and not have to think about research grants and tenure track.

Re: Don't Become a Scientist (1999)

#83

Post by LarsDu88 »

dr_dshiv wrote: post #36
LarsDu88 wrote: post #20

I can retrospectively look back at folks in my (non computer science) PhD cohort, and say that this bearish outlook is only true for folks who stay in science too long or pursue actual academic careers. The vast majority of folks I know flipped into more lucrative careers in industry... some of which require PhDs or Masters degrees. Furthermore, it really depends on your track. You can wind up in a lab studying somet…

> Sadly the most well funded scientific research right now is actually being bankrolled by techbro oligarchs for tax deduction purposes What’s an example?

Chan Zuckerberg Institute The Arc Institute (bankrolled by one of the Stripe guys) Deepmind (think alphafold) Heck, even Deepseek

Re: Don't Become a Scientist (1999)

#84

Post by cryptography »

Here's what nobody talks about: the author was RIGHT about the structural problems.

But completely wrong about the solution.

The PhD glut? Real. The postdoc treadmill? Absolutely real. The funding crisis? Still here.

But here's what changed:

The same skills that make you survive a PhD—deep research, systems thinking, hypothesis testing, data analysis—became the EXACT skills the market desperately needs.

2025 reality: - AI companies hiring PhDs at $300K+ base - Biotech startups led by former academics - Data science roles requiring scientific rigor - Deep tech ventures solving real problems

The trap wasn't the PhD. The trap was assuming the ONLY path was tenure-track academia.

The researchers who thrived? They took their training and built different careers: → Industry R&D leadership → Technical founding teams → Quantitative roles in finance → Policy and strategy positions → Scientific consulting

The irony: that essay discouraged a generation from science right before scientific thinking became the most valuable skill set in the economy.

The lesson isn't "don't get a PhD."

It's "don't limit yourself to one narrow definition of what a scientist does."

The best training for solving hard problems is still solving hard problems.

You just get to choose which ones.

Re: Don't Become a Scientist (1999)

#85

Post by bemmu »

jillesvangurp wrote: post #78

I did my Ph. D. between 1998 and 2003. I did not regret doing that but after one year as a post doc I left the academic world for one simple reason: I realized the career paths were limited to two options I did not like: option 1) try to shoot for a tenure position somewhere via a long series of post doc positions. Expect to be mostly busy managing postdocs and researchers and lobbying for funding by the time you get…

Sorry for possibly an ignorant question as I don't know much about research.

Do you think if you manage to have an exit from a startup, you might be able to return to research (at a university) but then be your own funder, so that you could then research whatever you want?

Re: Don't Become a Scientist (1999)

#86

Post by somenameforme »

zdragnar wrote: post #65
somenameforme wrote: post #56

Earlier quoted context omitted.

You need to keep in mind context. He lived in a time when the overwhelming majority of society was self employed and there was no formalized, let alone compulsory, educational system whatsoever. Looking up the exact history there, the first compulsory education began in 1852 (Jefferson died in 1826 at the age of 83), where children 8 to 14 were required to spend at least 3 months a year in 'schooling', with at least…

Even into the 1950's and early 60's, my dad went to a one-room school, probably until he was around 14 years old. No running water or air conditioning, the job of the first student to arrive in colder months was to start a fire in the stove to heat up the room. Had he been born a few years earlier, it would have been unlikely for him to even graduate. 1940 was the first year that the graduation rate hit 50%.

Absolutely, although by then electricity, radio, urbanization, and other such things had already radically reshaped the overall character of society to be something much closer to today than of Jefferson's time.

Jefferson, in modern parlance, would probably be a 'pragmatic libertarian.' He envisioned independent self-reliant people, and in fact (like many of the Founding Fathers) was somewhat opposed to 'economically dependent' people, including wage laborers, voting - for fear that their vote could be coerced too easily, and that they might otherwise be irresponsible. That's where things like property ownership came from as a voting requirement.

And a major part of self reliance is an education that is both broad and fundamental which is where the 'pragmatic' part comes in, as I think fundamental libertarianism would view education as exclusively a thing of the private market, whereas Jefferson supported broad and public education precisely as part of this formula to independence.

Re: Don't Become a Scientist (1999)

#87

Post by b3kart »

azalemeth wrote: post #79
cozzyd wrote: post #24

Scientist here. It's certainly not always an easy path, but if it's what you want, it's what you want. It's not something you fall into though...

Also a scientist here. The part of this article I disagree with is "Science is a profession, not a religious vocation, and does not justify an oath of poverty or celibacy". I really think it is the closest thing I have to a religion. It alters how I see the world, defines the way I think about how to approach problems, and consumes every waking moment of my life. Of course, I've had many failed relationships either d…

Sounds like you’re describing how things are, while the author is describing how things are ought to be. It’s fine to be obsessed and sacrifice the rest of your life for a job if that’s your choice. If that’s the only option to succeed, however, you severely limit the pool of people who are willing to make that sacrifice, and (I believe) slow down the inflow of ideas and creativity that push science forward.

Re: Don't Become a Scientist (1999)

#88

Post by Kudos »

cryptography wrote: post #84

Here's what nobody talks about: the author was RIGHT about the structural problems. But completely wrong about the solution. The PhD glut? Real. The postdoc treadmill? Absolutely real. The funding crisis? Still here. But here's what changed: The same skills that make you survive a PhD—deep research, systems thinking, hypothesis testing, data analysis—became the EXACT skills the market desperately needs. 2025 reality:…

I've been on LinkedIn a bunch lately while I'm looking for work. The cadence and "But here's what changed:" are extremely LinkedIn coded. I hope this is just an LLM and people aren't actually starting to talk this way.

Re: Don't Become a Scientist (1999)

#89

Post by dominicrose »

cryptography wrote: post #84

Here's what nobody talks about: the author was RIGHT about the structural problems. But completely wrong about the solution. The PhD glut? Real. The postdoc treadmill? Absolutely real. The funding crisis? Still here. But here's what changed: The same skills that make you survive a PhD—deep research, systems thinking, hypothesis testing, data analysis—became the EXACT skills the market desperately needs. 2025 reality:…

Quantity matters, how many scientifics work for an AI company versus how many are unemployed?

Re: Don't Become a Scientist (1999)

#90

Post by xnorswap »

Kudos wrote: post #88
cryptography wrote: post #84

Here's what nobody talks about: the author was RIGHT about the structural problems. But completely wrong about the solution. The PhD glut? Real. The postdoc treadmill? Absolutely real. The funding crisis? Still here. But here's what changed: The same skills that make you survive a PhD—deep research, systems thinking, hypothesis testing, data analysis—became the EXACT skills the market desperately needs. 2025 reality:…

I've been on LinkedIn a bunch lately while I'm looking for work. The cadence and "But here's what changed:" are extremely LinkedIn coded. I hope this is just an LLM and people aren't actually starting to talk this way.

Yeah, the only difference between that post and a full linkedIn post is the AI generated illustration.
Post reply on HN