Hacker News Reader

Comment by btrettel | Hacker News Reader

The US already produces too many scientists measured by people with PhDs. As someone who finished their PhD during the pandemic and faced a bad job market for research jobs, I feel that without increasing the federal science budget, the current generation of US PhD graduates will have to compete with more experienced Russian scientists entering the US job market. From what I understand exactly that happened during the early 90s and many new PhD graduates found themselves without a job related to their degree.

I don't recommend to anyone to get a PhD these days.

View on HN

It always amazes me that when the upper class is bothered by the more competitive job market the "immigrants are stealing our jobs" is fine and tolerated idea. When it's the lower class its no one is taking your job, they are creating jobs. Similar discussions when people talk about IT and Indian visa holders.

I didn't say they were "stealing" my job, just that new graduates won't be able to compete well against more experienced folks. A PhD is a bad deal even without more immigration.

Edit: I don't think your comment is fair. My recommended solution was not to cut immigration or anything like that.

Here is Eric Weinstein on the subject.

https://www.ineteconomics.org/perspectives/blog/how-why-gove...

My personal view is more similar to yours. We simply dont fund science almost at all. The economy is simply dysfunctional.

We do need a big ARPA like investment in research. Right now we depend on defense spending but we are simply not spending enough on research.

MIT has an endowment of $20M per professor.

We spend plenty on research, just inequitably, and with bizarre pyramid schemes.

I will be very worried for the future of civilization when these hundred-year institutions that were supposed to be at least partly self-funding acquiesce to the popular demand to bankrupt themselves over not even a generation. MIT could spend that much… for 8 years. Then what?

I think you're missing the big picture.

MIT's best research was 1960-2000. At the end of that period, the endowment was around 1/10th of what it is now. That's a very healthy level of funding. $2M per professor is enough that it doesn't need to worry about money, but that there aren't obscene levels flowing into odd pockets in odd ways.

Once an institution is overcapitalized like that, dynamics change. There's a ton of corruption in 2022-era MIT (and peer institutions), conflicts-of-interest, and the insane level of competition leads to a lot of unethical behavior. There's so much money swishing around that it's tempting to skim off the top. Combine that with the traditional MIT rule-breaking ethic, and you've got trouble.

I would love for MIT to spend 90% of its money, and be back to 2000-era endowment. I would even more love for MIT to expand it's faculty and student bodies 10-fold, and give more people an opportunity to be there.

What I hate is MIT now being brand-obsessed, money-focused, hyper-competitive, and valuing high-impact and high-profit research over high-integrity research.

Quite an uncharitable comment, and untrue in aggregate even if this user had been saying what you mischaracterized his comment to say.

The only meaningful buyer of research is the federal government, and the amount of research it buys is fixed by Congress. It's not responsive to "market conditions" in the same way as e.g. software, where software engineers create a booming sector that drives demand for more software engineers.

There's a Chinese American mathematician who had the exact thing happen to him: https://en.wikipedia.org/wiki/Yitang_Zhang

Dude obtained his PhD in 1991 and had some bad blood with his advisor. Coupled with the influx of Soviet talent and some stubborness with staying in academia, he had to deliver food and make sandwiches for almost a decade before finally landing a lecturer job and eventually academic breakthrough more than a decade after that.

I use to think this way. I do not anymore. A PhD is independent research training which is by far the most important skill for new knowledge and hence new business and new technology creation. We need way more PhDs than we have if we are to survive and thrive as a species.

I agree that a PhD is training, though I will disagree that it's independent research training as few PhD graduates can do truly independent research in the sense of not needing to collaborate with others. Academia emphasizes collaboration far too much in my view.

I don't think a PhD is necessary to learn how to do research, however, and I think it's actually a bad way to learn how to do research. The system is quite abusive in my view and pays poorly. I think it's likely someone could do better by taking a well-paid job and going on long research sabbaticals every once in a while.

A while back I recall someone on HN comparing a PhD to an apprenticeship, which is not consistent with my experience. My PhD advisor did not provide a lot of value I couldn't have got on my own.

And yet everyone says they can’t find enough talent.

It can’t be both. PhDs can do more things this just become a professor, which does have a fixed number of spots.

> everyone says

a vigorous and unending renegotiation between management, the agents of management, and people who "do it", in action.. advanced negotiating tactic - broadly give public opinion in mainstream media about "shortages" of whatever it is that you want more control over.. next, read carefully for market change and new entrants.. rewards, feast on newbs and do-gooders with special offers and limited term deals.. repeat..

I need specifc talent. I don't care how great you are a physics, pharmacy, mechanical engineering, or any of dozens of other PhDs that you might have, I'm hiring people who program and your degree isn't the right one.

Don't be too quick to write someone off for their degree. The best Python programmer I know has a PhD in civil engineering. And I've certainly got the impression that people write me off as a programmer far too soon because my PhD is in mechanical engineering.

Obviously a PhD is not a substitute for actual programming skills.

But what is so specific about your needs that the candidate absolutely must have a CS degree -- and for which it simply will not suffice, as is nearly universally standard throughout the industry, to have "equivalent experience"?

I didn't say I'm hiring CS degrees. I just said a non CS degree isn't anything. There are lots of great programmers with many different backgrounds. However if you don't have a CS degree you need to prove you can program. Even those with a CS degree need to prove they can program, but odds are a bit better they can.

I just said a non CS degree isn't anything.

A strange view of the world (even the world programming), to say the least.

Some PhDs are great programmers. Some read half of how to learn YYY language in 24 hours and think they are god's gift to programming.

In general I find people without a CD degree don't get why clean code is important. Since I'm on a project with tens of millions of lines of code (big, but I'm well aware that other projects are much bigger) clean code is important.

With you basically, but I phrase it as: "Without a CS degree or equivalent experience."

The good news is there are more employment opportunities than just programming for these folks.

Meh ... PhD are super-specialized, which is great if that's what you are looking for, but may not be particularly good if not.

For example, a person with a PhD in history will probably be a good writer and researcher, but may need a couple of years training and experience before they're an amazing journalist, for example, and may only be OK at PR, even after a couple of years. If you want PR 'talent' then the history PhDs would not be very helpful.

A bio PhD would be amazing at many things, but would probably suck as a nurse.

So it's not just the raw talent, but also training and inclination.

This depends.

I know people doing/with PhDs who are great theoreticians and good programmers but I also know people who (will) have all the creds in the world but simply know fuck all about actually making useful software/anything. Even in an industrial research context they'd be less useful than you'd think because they wouldn't be able to translate it into useful business applications.

>PhD are super-specialized

Yes they have specific expertise in some field. But also they know how to do intendent research. This is the critical skill.

Why do people need to change professions?

My wife working in research at the moment and wouldn't recommend that either (as she works her thousandth weekend in a row for low pay getting stressed about output to find the next research position)

From what I remember of the academic job market in the 1990s, this was a real issue. You had all these top level researchers filling positions that would have been taken by newly minted PhDs. It resulted in a spurt of quality work being produced at the destination institutions, but I was even then concerned about the effect on the younger generations both in the West and the ex-USSR.

There's too many phd's for the number of available academic jobs, but that has nothing to do with broader hiring...

I'm thinking of non-academic jobs too. There are a lot of research positions requiring a PhD in my area. There's still a massive surplus of PhDs in my area despite this.

Get good. Artificial scarcity of PhDs (or really anything) is not the answer. We want meritocracies right?

I think the quality of my research is considerably above average. Unfortunately the people making hiring decisions mostly look at things like number of publications and years of experience and rarely the actual publications themselves.

Their loss? Do you want to work with organizations that hire that way? What is your goal?

It is their loss, and I don't want to work with organizations that hire like that. But that's how hiring is done in academia. A lot of the non-academic research labs work similarly. (Classified research does not have strict publication requirements in my experience, though.)

Having a job doing valuable research would be nice. I'm not too picky about the topic, but it seems that most employers view the capabilities of a candidate very narrowly. So I'm probably stuck with things like what I did during my PhD.

PhDs aren't fungible.

Reply on news.ycombinator.com