Since the academic culture is "publish or perish", those who were able to game the system became successful much faster (and with much ease) than truly honest hardworking researchers. Now you will see papers published in OA journals with 7-10 authors each coming from different countries. There are so so so many dubious fake data papers. Even some are outright so stupid that they break the law of thermodynamics, yet the journal editors don't care when it's reported to them.

Yeah the only ones that matter are math but here's the thing. Chatgpt cN connect other research papers for ideas and information

Thanks for sharing your experience

Hello in the future i want to be a quantum Physicist i am still in high school tho i just wanted to ask how am i supposed to make contributions to science if now one cares about papers how are you supposed to share your breakthroughs? And are there still places that give you time to publish papers that actually improves are knowledge and where academics have more freedom to work on projects or something else?

Thank you for keeping this discussion rolling! In my experience these issues are not homogeniously distributed. It depends on the country, the institution and also the field. As a rule of thumb I would say that the more money is involved, the more competitive and toxic the research environment gets. What I personally find interesting is that the big publications (in high impact journals) in my field usually don't really help me to progress with my own research. I often rely on "smaller" publications which are more dedicated toward the experimental details and openly discuss the remaining problems. This alone shows me that metrics are not the best way to evaluate a researcher. Sir Paul Nurse (a Nobel Laureate) once used a military analogy by saying that the researchers at the forefront might push a field to a new direction but often leave a mess behind which need to be sorted out by other researchers (repeating experiments, double checking, using other mehtodology etc.). We tend to overvalue the frontier scientists (most awards and public fame is directed to those) but we don't really appreciate the footwork by the majority of scientists who sort things out. Therefore, many scientists chase these metrics even after they are already set and got their tenure - making it insanely difficult for younger people to get a piece of the cake as well. The only way out is when scientists themselves - meaning those who already have some power - break this cycle and go back to what science actually is about. I am glad that more and more people in the community at least acknowledge this problem and openly resist this absurdity.

Excellent!

Low hanging fruit has been picked. Now only those with great reach, beyond even some of the most intelligent people you may know, can bring something of value I think this is a good thing

Bro’s trying to promote his paper!

Ideally, Academia’s real product is knowledge. New knowledge regarding a current debate (papers). Systematic and didactic to enhance people’s skills (classes). solution aimed and technology applied (patents). Socially applied (extensions etc.), science dissemination outside of specialized fields (science communication). Such a wide range of responsibilities. This fixation on papers is detrimental to the academic duty as a whole.

I suggest try to go to big and reputable institutions and labs. That will help you a lot in terms of your career growth. One thing I have realized is that networking is far much important than your publications and CV in the academia. That's why I am saying try to go to reputable places and labs if you want to continue in academia. Your career is actually not equal to your papers if you know the right people in academia. Papers are of course important, but based on my almost 10 years of experience in academia, networking can excel you in your career more than your papers.

Academia is very unfair. Impact that a paper makes depends a lot on the reputation of the authors. If some big guys publish a paper, it is certainly going to be cited no matter what is in there. I had an experience with that. I published a paper. It was being cited, but not too much. Four months later there appeared another paper with the same results (frankly, I would say that it was by far sloppier and more incomplete than mine) by some guys from top universities. They acknowledged my paper saying something along the lines that "once our work was finished paper [x] appeared with which our paper has some overlap". Of course, it was published. Already this I find quite inappropriate. According to normal standards it should have never been published. I can understand that it can be frustrating if someone scoops you and I think that if the work is almost completed once the other paper with the same results appears, researchers should be still allowed to publish what they have. But not 4 months later! It took me 4 months to do it from scratch by myself and they were 3! I also think that the way they acknowledged my paper is completely inappropriate. But this is not yet the end of the story. Their paper received by far more citations than mine. They got invited to many conferences to speak about it, while I was not invited to any single one. The funny thing is that I even benefited from this! My paper started receiving more citations than before the paper of these other guys appeared! People cited their paper and saw them acknowledging me and then cited me as well. Still, they were being cited by far more than me. Here how it worked. Significant part of citations was very vague, so it was really hard to argue about them. For example, their paper was cited as "computations of this type are very difficult". Of course, in a vague way like that you can cite hundreds of papers and people usually choose to cite just a few papers of famous guys. This is one factor. But there were also papers citing them for a concrete result, which was, in fact, obtained by me. I wrote a couple of letters suggesting to look into my paper, where these results were obtained earlier. Responses were roughly like that. About a third replied to me that they will add a citation to my paper and they actually did. Some replied that they will add a citation and they did not. Some people did not even bother to reply. There were even some people answering something along the lines "if you want us to be so generous with citations, please, cite these 10 papers back". So, working in academia can indeed feel unfair. And it is all about citations.

Roughly one million papers are published every year in academic journals. Obviously 95% of them are useless and forgotten as soon as published. Evaluating by number of papers is absurd, it is the value of what is written which is important. We should think in terms of novelty and inforation, not number of papers.

PhD student here, I know for a fact I won't get into the mess academia is after I'm done with the PhD The whole system is absolute trash and I don't have the determination in me to fight to change it

Great video. This video has garnered 240k views, which is more than the total citations an academic could receive even if they published 200 papers, each cited 1,000 times. Tremendous impact, high value.

Humans ruin everything with bullshit.

Papers are bulls…, only PHD and above should write any kind of papers

Great and useful content.

In our department, quality trumps quantity. First of all in mine and most labs in our department writing a paper is usually a collaborative process. As a result, the first author is still able to spending the bulk of their effort at the bench collecting data that will hopefully be the focus of their next publication. I experienced a similar process as both a graduate student and postdoc. I would not tolerate postdocs that focus on padding their CV by splitting up a papers worth of data between 3, 4 or more manuscripts. I have been involved in at least 15 searches for new faculty. An applicant with 20 insignificant published paper will be dropped on the first pass. It is painful to have to read 3 or 4 papers in which the majority of the data presented is identical with 1 or 2 new experimental results tacked on to make each article novel. We tend to hire new faculty that have a history of doing research that is cutting edge and publishes interesting and thoughtful papers on a ‘consistent’ basis. When they come up for tenure the rule is rule is they need only 5 papers, most hopefully published or ready to be submitted. Some faculty, when reviewing tenure packages will review the CV, but only read the 5 papers submitted in the tenure package.

You actually believe in the "lived experience" scam? This nonsense was included in application procedures several decades ago to be a check on political correctness, not to prove competence. To be 'insulted' about being asked about your product and its effect on the science rather than being asked for proof of ideological purity is an indication that you are brainwashed. You may be correct in saying that the quality of academia has fallen, but it has fallen BECAUSE of the kind of crap you seem to prefer.

The answer is quitting academia. Really guys, it's not worth it.

This is why I stopped thinking about a future in academia. It seems as though my professors and seniors are just invested in publishing papers. Not their value.

Hi, Andrey - sounds like you have figured out the game in academia. I think it is a tragedy for science and human civilization, but it is what it is. I bailed out of academic mathematics years ago after similar realization that mainly worthless effort was rewarded, and that very poorly. I ended up in business, published a few things related to transformer diagnostics, and ironically am getting almost as many citations on a 1988 category theory paper (which got none for several years) as on my least unsuccessful recent transformer paper. Best wishes to you.

That is why I completed all the projects that were assigned to me in a relatively short amount of time by sacrificing weekends, family time, etc. Now, I am free! I dedicate my time on coming up with problems that actually matter and potentially have more impact than my previous works. P.S. I will leave academia once I am done with my PhD.

I also left academia after being thoroughly disappointed with the system. As a Ph.D. I realized that no one actually reads papers/theses because they're so hard to understand. So I wrote my Ph.D. thesis as a textbook: starting from master-course level, building up with examples, exercises, background appendices, and so on, up to the level of my own published papers. My students loved it. My thesis committee didn't: it was rejected. I had to write my thesis as paper-style: condensed and incomprehensible to anyone but a veteran in the field. What followed was a year of evening-time rewriting (without getting paid) merging all my papers into a very dense thesis. This was in the end accepted and I got my Ph.D. degree. Now, six years later, I still get regular emails from random people thanking me for my free introductory textbook thesis. It helped them a ton to get familiar with the field. I never got a single email about any of my papers or my rewritten Ph.D. thesis. Conclusion: no one needs papers. (Sure, with a very few exceptions.) We generally need clear and high-quality educational materials. But the system only forces us to write papers and rejects anything else. The system is broken. The above video also shows really well how the system is broken. Sadly the solutions provided are "How to survive and thrive within this broken system" and not "How to actually fix this system, so science can start making an impact and improving the world again."

Wow, i like how you perform this piece. I love rachmaninoff

Thanks for the advice Andrey. It was quite helpful to know others are going through the same (8 months delay here, and I'm not in the UK).

As a programmer, this reminds me of the stupid old attempt of measuring the output of programmers by the number of lines of code they write. The result is exactly like in academia: The codebase/literature gets swamped with low quality garbage, which only makes things harder with time. The next programmer or researcher than has to navigate this ocean of junk. So good development teams eventually realised that readability and code quality are the most important. Your should write your code in the length that is best to make it readable to others. If a function can be written in 10 well readable lines, then that's the optimum. Not 20 lines filled with unnecessary fluff, and not 5 extremely complicated ones. If the academic system cannot tolerate when a researcher wants to prioritise depth and quality over quantity, then things are clearly going wrong.

People should do the division and think, publishing 12 papers a year means you demonstrably spent 1 month on each. Publishing a paper per year doesn't prove 12 months of effort, but there is time for it. Regarding impact & benefit to humanity-- I think taking your papers results to youtube is a possible solution. Access to papers is so limited that getting info into the heads of people who need to know is going to be hard if researchers stick to traditional channels. That and the whole genre of youtube channels that discuss recent papers.

What happened to quality over quantity? It seems to plague everything from tech to healthcare.

Your point is important but I heard from my supervisors and older colleagues that you should forget about your paper the moment you completed and sent to a journal. You can occasionally return to it for revising, reviewing or citing but you should largely let it have its own life. Otherwise one can't stop thinking about regret and disappointment. Imagine Archimedes complaining about the impact of his paper called Stomachion on his contemporaries. He probably never did and how pointless would that be. Science transcends cultures, generations and people. Keep working my friends!

Everything you said is spot on. It’s a broken system. Leave my friend. All the hard work you’ve put in over the last 10-15 yrs deserves to be rewarded. You’re in a unique position because you are an engineer, so the transition to industry should be fairly simple. It might not seem like you can find something to be passionate about in industry, but I think that is one of the biased perspectives that academia forces on its members!!!

Maybe your paper was shit

Thank you for sharing your experience with the ATAS application. It was very helpful!

Well done Andrew. I cheated and played a simplified version transposed to C minor! I used to have a recording of Rachmaninov playing this, awesome!

There's a striking difference between working for a technology company such as Google, where the outside world knows little of what has been tried and failed and only hears about what has been brought to market, and scientists in the universities, who are obliged, both to keep their own positions and to please their grant providers, to keep publishing the minutiae of what they have been up to. It's not unlike medical research, where scientists funded by, say, a cancer research charity are have the same pressure to keep publishing, whereas those working for drug companies are bound by commercial secrecy. Instead governments are increasingly giving drug regulators powers to snoop into the companies research, on animals, healthy humans and patients, due to their history of concealing problems until they have made enough money from a new drug. Non-industry bodies talk of registering planned clinical trials so that we all may see which have been published, kept secret or abandoned. People outside that company also need to know what did NOT work, to avoid harming more animals or humans by repeating similar work. That's the reverse of academics' competitive pressure to publish: companies have to be dragged unwillingly to reveal what they are doing. Maybe we, the tax-paying public who provide the subsidies and tax breaks that these high-tech companies enjoy and pay for their huge public sector contracts, also need to demand transparency. If we had known what Fujitsu, the favoured son of the British computer industry, had been doing with their Post Office contract, a lot of official malfeasance and personal misery could have been avoided. Mrs Thatcher elevated commercial confidentiality to its current position in public life, pari passu with her privatisations. It's time to replace it with glasnost.

I was disappointed to learn that face recognition by machines is a growth area for research. That gives teeth to the paper tigers of the surveillance state. Eighty-odd years ago, physicists working on the Manhattan Project knew that they were working to defeat an alliance of the three most evil fascist regimes. Today, optimising ways to drill down from mass observation to monitoring individuals enhances the power of the Chinese Communist Party and those who seek to emulate them. That's the "trahison des clercs" of our time.

Wellcome to neoliberal late western capitalist academia

As an ex-social scientist, I was confronted with pretty much the same problem. There are a few towering figures, who dominate the field, while 90% of new researchers struggle to be recongnized. Positions are very coveted and there is very little chance to get tenure. Nobody actually has the time to really read anymore. There is a truely globalized labor market, turning so many people into nomads that family life becomes virtually impossible. Here's my solution: I left. These days, I work as a Civil Servant, read widely for pleasure only - and write essays and fiction (the second novel is now almost finisehd, yeahi!). One thing for you guys/gals in the STEM-arena, though, I really don't understand: Aren't there a ton of private companies yearning for your knowledge and expertise, offering you high paying, interesting positions? Really, someone with an advanced physics degree and expierence in modelling/coding/data science -- you'd surely earn >$200.000 p.a. and have your own little team in some industry, doing interessting, impactful stuff, after 2-3y of work max. So, why not leave for something much better?

I work in industrial R&D, and I read a lot of papers but don't publish many. My two cents is that the vast majority of academic papers in my field are almost useless. When the incentive is to just publish, the task becomes to generate results that are *plausible*, and *novel*. The quality of the experiments themselves can often be quite poor. Irreproducible or cherry-picked results are common. There are many papers exploring ideas that are really quite hopeless, and known to be so by the authors, but as long as the "result" is new, and riding the right hype train, it is easy enough to get published. If your deliverable is a simply a *story*, and not a physical product, or demonstrable & reproducible improvement upon the state of the art, the result is going to be a lot of bullshit.

Hearing this, I'm glad I'm not in research. I'd get depressed really quick.

Adam is very important, everyone who works with DL knows that without Adam it won't work so well or it will take ages to train, making it just impractical. It's a paper with many leaks (and a wrong proof) but its impact is unquestionable. However RMSProb was quite good as Adam but it has not even cared about a pubblication lol (it is quoted from a blogpost and a coursera video).

Exactly this

Maybe this calls for a switch in funding methods? You'd think with so much effort and work going into research, that it would be best for that work to be put towards what people need most. Maybe something like crowdfunding or building some business that links researchers with industry and the public, so that researchers can research on demand or more easily get employment under a company that requires the _research_ itself, rather than the papers. I dunno I'm not an academic myself, but as such I honestly doubt that papers for paper's sake do any good to anyone. If researcher's careers depended more on the value of their research rather than how many papers they published, then maybe research work would be better spent and more people (both the researchers themselves and humanity in general) will stand to benefit from actual meaningful scientific discoveries, rather than indirectly paying for glorified academic word bloat (through government spending aka taxes, and through increased costs of tuition for universities that fund academic paper-grinding)

Are you speaking about Jalal K ?

All is not lost. A hundred years from now someone will read your paper and have a good laugh.

I just love the metaphor at the end between videos on KZbin and papers. They have the same goal to be watched and shared !

Minkowski addition to studying the capabilities of Virtual Power Plants is business as usual today, by the way, people who are not in the field still have problems catching this concept.

really good work ! Congrats

Totally right, science is fucked. I guess it's a big issue of visibility or finding the right paper. And knowing the quality of the paper. I remember my diploma thesis and the search for prior art, lots of papers look awesome, but the results don't reproduce etc etc. So just the validation of other peoples work is a lot of work. And there is not that much time for it, if you work on your own stuff. No reward and incentives at all in science, to do the necessary scrutiny. And with the following inflation of papers, visiibility for the good ones becomes more of an issue. Have you tried to go to conferences and present your findings there? It worked at least for friends of mine to some extent. In the end it's not just writing the paper, but promoting it as well.

So quantity over quality and focus.