[godelski]

> Don't forget that this is actually money laundering.

For anyone questioning this line, let's remember a few things

- Graduate students (in this setting) are typically funded, so the cost of their credits (often higher than undergrads) are determined by the university and such costs are a major factor of what is taken out of the grant the professor gets their portion (sometimes after the uni's cut!). The rest then goes to the student's salary and hopefully some left over for new lab equipment.

- Grant money must all be used and cannot be put aside for future investments. It is better to buy shitty lab equipment because you don't have enough for good equipment and can't invest any excess (even if by being spartan elsewhere). If you don't use the money in the allotted timeframe you're considered to have improperly managed the funding.

- A funded graduate student is considered 49% employee and 51% student.

- Graduate students in year 3+ (median 5 years for PhD) are not taking courses and doing full time research and likely being a TA at the same time. (Tuition costs do not change)

- A successful graduate student sees their advisor less and less as they dive into their niche area of research where the advisor no longer has any level of expertise. (This is what's supposed to happen)

- When a graduate student stops taking classes they still pay for credits and at the same rate (albeit through funding, which they are often writing for at this point. But prof gets the award).

- Universities pay students and professors to publish papers and judge success by publication in venues

- Students and professors "judge" works submitted to venues by other students and professors for no pay (i.e. on university time)

- Venues take copyright ownership over works they deem valuable and put it behind a paywall

- Universities pay for access to venues where their researchers published in and where their researchers performed volunteer service for.

- Promotions are given to those who's name is on the most works, regardless of position or contribution to that work.

Think about it this way, what if we framed this as a job? Your job considers you a junior part time employee for the first 5 years and if you don't complete all 5 years every other job will treat you as a junior part timer. Your first two years 50% of your time is spent doing training, 50% of your time is spent teaching the interns (who pay, but who spend 100% of their day training), and whatever time you have left is spent performing research. You're told you're a part time employee because 51% of your time is training. After two years you finish training but get no change in pay (maybe +$100/mo), nor graduate to a full time employee. By year 4 your manager never shows up except few months your manager comes around telling you that you need to make sure to make a deadline and they need to read your report first. They demand it is in their hands a week early so they can review it. 3am the night before the deadline they ask for major rewrites, this is the first you've heard of any problems. 10 minutes past the deadline you're still getting requests to "modify the graphic" with instructions like "a little to the left" or "I don't like the colors" and the iterative process can only be performed by back and forth submissions with random delays as your manager won't touch the source code. Every few months your manager stops by to check on progress and ask you to write a report that needs to be written by tomorrow. They'll slap their name at the top and if successful they advance their career. Your reward is via proxy. After 5 years, you write a large report about what you did the last 5 years filled with stuff you've mostly done over the last 18 months and pretend that you had a plan all along. If they approve, they usually do (but will ask for changes), you can go be a manager if you're lucky or get a full time position. Or if you go the post-doc route, 75% employee.

Idk, this sum it up pretty well? Anyone want to add anything?

[jacquesm]

This is really painful to read.

That whole system seems to be so ripe for disruption.

[godelski]

Well just know you're not alone. I hope you got out without killing your passions.

Fwiw, I intend to lead by example. I love researching. I have a long term internship where I even do research (unfortunately not closely tied to my PhD work lol). But since I read math books and research as a hobby, I intend to simply do what I call for (in other comments) and just post to GitHub + Openreview + Arxiv and call it a fucking day. I hope to get others to join me in this paradigm shift. We all fucking rely on arxiv anyways and I'm pretty sure more of us find works via twitter/google scholar/semantic scholar/word of mouth more than we find works via journal/conference listings (twitter post of "just got accepted" counts as former, not latter).

I'm not so sure we need "disruption" as much as we need to just cut off the fucking leeches. The problem was turning school into a business. Thinking that profits align with education of students. But we have no strong evidence that higher ranked schools produce higher quality students, but rather only better connected ones.

Idk, maybe the private sector can disrupt it. But they'd have to perform a pretty similar feat, though there is a monetary benefit. Because the world is disillusioned that Stanford students are substantially better than Boston College students, you can pay the BC student less. In fact, many places do, but the issue is Stanford has a huge fucking media arm so we don't hear about that. They can also stop using number of papers as criteria but rather quality of papers (i.e. use domain experts to hire domain experts. Novel idea, I know...)

I'm just shooting in the dark here. I'd actually like to hear other peoples suggestions. Even if we're just spitballing at this point (I don't think anyone has strong solutions yet, that's okay), we just need to get the ball rolling at this point instead of talking about what a ball's relationship to an apple or the sour more rounder apples that are orange.

[jacquesm]

I got lucky: I never went in. My family more or less imploded in the middle of my highschool track and I went to work instead and that put me on a faster road to a lot of interaction with the computers of the day than school would have given me and that led to an interesting career. If that hadn't happened I may well have ended up in academia and I somehow feel I dodged a bullet there because my ideas of what university was like at the time seem to have very much been informed by pink glasses and meeting the occasional very interesting person who was part of the academic world.

[godelski]

Don't get me wrong, there is a lot I like about academia. Honestly, there hasn't been any other point of time in my life that I've been able to dedicate so much of my time to learning and researching. Even as I'm in a long term internship, that freedom is slipping away. Academia is supposed to be about protecting that freedom to explore and learn, but simply too much bullshit took over. Bureaucrats love metrics regardless of the value of those metrics. Maybe I wouldn't feel as disenfranchised if I wasn't in the fast moving world of ML with where peer review is like playing a slot machine except bigger schools and big labs get access to slot machines with higher payout rates (I see no quality difference between works from different institutions, rather the arxiv wave primes reviewers or language/proprietary {models,datasets} also prime reviewers).

I actually want people to feel disenfranchised at this point though. Because if there's anything I've learned, it's that we don't fix things before they break or even when they are noticeably broken. Rather we fix things when they're so broken that they're unusable, and typically only fix to minimal usability. Which is such a waste of resources. Maintenance is far cheaper.

[ModernMech]

>- Graduate students in year 3+ (median 5 years for PhD) are not taking courses and doing full time research and likely being a TA at the same time. (Tuition costs do not change)

Typically they will only take 1 credit after entering candidacy though, down from 9, so overall tuition drops significantly.

> It is better to buy shitty lab equipment because you don't have enough for good equipment and can't invest any excess

Sounds like you should have asked for a more appropriate amount of money in the grant.

> If you don't use the money in the allotted timeframe you're considered to have improperly managed the funding.

Grants can often be extended and funding can be supplemented.

> Students and professors "judge" works submitted to venues by other students and professors for no pay

Sure, but we are compensated by other academics reviewing our papers for no financial compensation.

> Venues take copyright ownership over works they deem valuable and put it behind a paywall

They take copyright over the submitted manuscript. I maintain copyright over preprints. Anyway, the point of writing the research is to distribute it, not to own copyright over it, so I don't see the problem. You can choose the venue, and not all venues take copyright over the submitted manuscript. If copyright is important, then choose one of those venues.

[Fomite]

A bunch of this is just...wrong.

"- Graduate students (in this setting) are typically funded, so the cost of their credits (often higher than undergrads) are determined by the university and such costs are a major factor of what is taken out of the grant the professor gets their portion (sometimes after the uni's cut!). The rest then goes to the student's salary and hopefully some left over for new lab equipment."

There are a number of ways to cover students - TAships, university level scholarships, and grant funding. For grant funding, the cost of the credits is something we can budget for, is in my university markedly lower than an undergraduates, and is budgeted for. This is portraying "We had to budget for someone working" in a weirdly salacious light.

"- Grant money must all be used and cannot be put aside for future investments. It is better to buy shitty lab equipment because you don't have enough for good equipment and can't invest any excess (even if by being spartan elsewhere). If you don't use the money in the allotted timeframe you're considered to have improperly managed the funding."

It's not better to buy shitty lab equipment - while grants don't like funding large capital purchases that will cross projects (except for the grants for this), the equipment doesn't vaporize. The cluster nodes and servers I bought for my first project are still running, and indeed go in applications for new grants as equipments I have, in a section often titled "Facilities and Equipment".

As for not spending it out in time, there's what's called a "No Cost Extension", which is "Hey, we didn't spend the money in time, can we have a bit more time?". The NSF grants the first one of these automatically, and one grant I'm on is on it's third (a program office has been very understanding about the difficulties of conducting research in hospitals during a pandemic).

I've never had pushback from a program for getting an NCE unless it was genuinely something where we messed up spending somehow.

"- A funded graduate student is considered 49% employee and 51% student."

Nope, they're 100% students. This is both good and bad for them, but it's true. They're just expected to spend half - or less - of their time in classes, and the rest on research.

"- Graduate students in year 3+ (median 5 years for PhD) are not taking courses and doing full time research and likely being a TA at the same time. (Tuition costs do not change)"

Every university I have been at has had a mechanism for a massive cut in tuition once a graduate student has passed their preliminary exams and become a candidate. It's a big enough one that literally my first instruction to my students is "File your ADB waiver please."

"- A successful graduate student sees their advisor less and less as they dive into their niche area of research where the advisor no longer has any level of expertise. (This is what's supposed to happen)"

When my students are "on approach" they see me and their committee more and more. They're just expected to drive those meetings more as well.

"- When a graduate student stops taking classes they still pay for credits and at the same rate (albeit through funding, which they are often writing for at this point. But prof gets the award)."

Again, this is simply incorrect.

"- Universities pay students and professors to publish papers and judge success by publication in venues"

Perhaps in the most abstract sense, in that scholarship is a metric by which I was judged for tenure and promotion, and without that, I don't have a job and thus am not paid. But there has never been a "paper bounty" or something like that for any position I've been in.

Venues do matter, and some places are cutthroat about it, but other places aren't. In my department for example, publications that are in respectable journals appropriate for your discipline will carry you all the way to full professor.

"- Students and professors "judge" works submitted to venues by other students and professors for no pay (i.e. on university time)"

I consider this part of my service obligation, and indeed when filling out annual reports and the like, list reviewerships and editorial positions.

The rest of your stuff on publication is actually refreshingly correct.

"- Promotions are given to those who's name is on the most works, regardless of position or contribution to that work."

I sit on my college's tenure and promotion committee. This just actively isn't true. We look at the difference between solo and co-authored papers, where a particular individual is on a paper and the balance between first, last and middle authorships (I'm in a field that doesn't alphabetize). We also consider whether someone is expected to be there, or is anticipated to contribute a lot to work that others will end up being the lead for (as a modeler, this is occasionally the position I'm in).

Then there's positions that give considerably more weight to teaching or service.