Papers and patents are becoming less disruptive over time

[mountainFrugal]

New Nature paper out today quantifying the decline of new disruptive ideas:
https://www.nature.com/articles/s41586-022-05543-x#Sec8
Article Summary: https://www.nature.com/articles/d41586-022-04577-5

From Discussion:

Overall, our results deepen understanding of the evolution of knowledge and may guide career planning and science policy. To promote disruptive science and technology, scholars may be encouraged to read widely and given time to keep up with the rapidly expanding knowledge frontier. Universities may forgo the focus on quantity, and more strongly reward research quality, and perhaps more fully subsidize year-long sabbaticals. Federal agencies may invest in the riskier and longer-term individual awards that support careers and not simply specific projects, giving scholars the gift of time needed to step outside the fray, inoculate themselves from the publish or perish culture, and produce truly consequential work. Understanding the decline in disruptive science and technology more fully permits a much-needed rethinking of strategies for organizing the production of science and technology in the future.

Sounds like the academic meta-crisis caused by misguided specialist/paper number over quality incentive structures.

[jacob]

I can't help but think that [any mature] academic research field is also subject to https://meaningness.com/geeks-mops-sociopaths ... eventually everything turns into sales and management.

Also consider that insofar knowledge is finite, the frequency of disruption would converge on zero. This is definitely the case in terms of the physical sciences in the developed world. (Productivity growth has a strong trend downward.) For example, there hasn't really been any "disruptions" in physics since the standard model in 1975.

[Sclass]

The system is broken.

I wrote a few papers. My name got pasted on some patents. It’s pretty embarrassing stuff. Didn’t advance human knowledge at all.

From what I can tell my academic research was a byproduct of grant gaming. It’s taken a few decades to come to this realization after dissecting my thesis advisor’s career. My thesis advisor was a big dog in our community. It was a tit for tat game between him and other professors. We were a paper mill. He and his graduates controlled the reviews. A complete sham to defraud the taxpayers of their money. No useful science came out of it. But boy did we publish and land big grants.

I later worked in industrial research. The patents were written by my coworkers to claim juicy patent bonuses on work that I don’t think was even patentable. But it was disclosed and ultimately filed. Some of my team made more money off patent bonuses than their base salary. It was a game and legal just pushed it through because they got their cut.

I never list them because to informed eyes they make me look very bad. Really stupid ideas that were undoubtedly already invented. Prior art was just ignored out of convenience.

So I kind of shrug my shoulders when breakthroughs are gauged using papers and patents.

@jacob one explanation I’ve heard about declining breakthroughs in fundamental science is that the human brain is no longer powerful enough to expand our current limits. It’s depressing and I’m not sure whether I agree or not yet. The current problems are a bit tougher than starting a fire.

[mountainFrugal]

New knowledge in a finite world of knowledge converging to zero makes sense. I had not thought about it like that.

Labs that have really good managerial PI's do tend to out compete others as far as "output" in grants written focusing on incremental ideas, papers published etc. I did see similar labs to what @Sclass describes during grad school/postdoc. In an extreme case within my field, all the postdocs and grad students coming out of a certain lab that had faculty positions would meet once a year as an "alumni" retreat to strategize about who was working on what, who was applying for what, etc. Although, I think their collective "alumni/mafia" contribution to the field was moving things forward quickly in a non BS way with innovative technique developments.

[jacob]

@mF - I've posted https://arxiv.org/pdf/astro-ph/0310368.pdf a couple of times. It came out when I was in grad school and later turned into a book. I'd say that as far as I know that paper describes astrophysics as it is (a relatively dead field). I'd be curious to know if your perspective on yourField is similar?

I think it's different in new/live fields where a) lots of things to discover; b) lots of money; c) rapid win-win competition due to demand outstripping supply of researchers. For example, solid state physics, molecular biology, or computer science.

On a fun note, it's a sign of field peaking when you get career advice from rock stars https://www.youtube.com/watch?v=8qrriKcwvlY ... people who took this advice basically graduated coincidentally with the end of the cold war.

[mountainFrugal]

I'd be curious to know if your perspective on yourField is similar?

My area of expertise has been growing rapidly. There has been no popularization through music videos by famous people that I know of. Is that when physics jumped the shark?

Interestingly, the paper you linked is very similar to how I view things. This may be because this is written as a critique of astrophysics, but is really a good critique of academia. For all the problems, because the field is growing there are new things being discovered all the time which reduces the "idea discovery" pressure, but does not help with the ever decreasing research funding pools and the larger numbers of researchers fighting for it.

Students and Postdocs
More or less the same. Students and postdocs are relatively cheap labor for field work, monotonous labwork etc. In an extreme case I saw in the department I was working in, PIs striking deals with foreign PIs to send over "students" as interns. The foreign institution would pay for the additional prestige of being on papers with US institutions by providing the lab labor. Even cheaper for the US based PIs to not have to pay anyone US wages to do the work that was being done in their lab.

In other cases, when the supervisor sees that things do not have the outcome he or she wants, the supervisor abandons the student.

Stories from more famous labs: "If you do not have a Cell, Nature, or Science paper, your work is not getting published out of the lab at all." This is essentially the opposite of science training. Other examples include pitting 3 postdocs against one another in a race to conclusive results for authorship on the paper. It is easy to see how this creates the most toxic of environments.

Made me chuckle because it is mostly true if the PIs even remember how to do actual lab science:

PhD students can produce ideas if allowed to produce them, even away from the track that was predetermined for them. In this regard, it would stand more to reason that the monotonous work should be in the hands of those with exhausted creativity, those aged, reputed experts who will produce nothing but copies of what they have always produced. However, the world of science does not stand to reason but that of power: the captain gives the orders to the sailor.

Another great zinger:

However, nature is difficult to understand, and truth may have nothing to do with the positions taken by some scientists who do not want to have disturbed the peaceful tranquillity of their lives.

My reason for posting the original paper was because it suggested different funding strategies for supporting entire (or partial) careers to focus on longer term projects and sabbaticals in other fields instead of grinding to squeak out yet another minor finding before a competing lab. The funding stakes are too high. Getting new funding is based on how many papers you have from your previous funding. And so it goes. I will continue to do some sort of research for the rest of my life whether I get paid or not. ERE offers an interesting side door to being able to do long term work on some idea without having to be affiliated with an institution. Publishing the ideas is another story. .

[AnalyticalEngine]

ERE offers an interesting side door to being able to do long term work on some idea without having to be affiliated with an institution. Publishing the ideas is another story. .

I met someone last week who was a psychology postdoc. Because the subfield he was in wasn't very hot, he was responsible for pulling in 100% of his own funding from grants. He had recently ran out of grant money and was basically looking at being "unemployed" where he was technically affiliated with an institution but wasn't getting paid.

With the internet being a disruptive force, I've wondered how viable it is to do your own research then publish it in a book or something rather than attach yourself to an institution and pray they throw you enough crumbs. Certainly the university brand helps and it's pretty hard to do research into nuclear physics on your own, but given all the horror stories I've heard about academia, this seems preferable if your field allows for it.

[jacob]

The keyword for doing work that is not paid for by the institution or mainly paid by a PI grant on a short-term basis is "visiting", e.g. "visiting assistant research professor".

I started out in nuclear physics before I gravitated towards astrophysics. It is certainly possible to do computational/theoretical nuclear physics out of your bedroom. The biggest moat would be journal access although that's somewhat alleviated by preprint servers (widely used in subatomic physics). Unlike mathematics, there are very few solo-authored papers in physics, so likely one would attach oneself to a professor at some institution who would become "last author" (aka, the guy who paid for it). This would likely also result in an edu address and the corresponding journal access; maybe even supercomputer/cluster access, which tends to be helpful but not necessary. I knew a dotcom [multi]millionaire who did something like that pursuing a second career as a researcher. The going rate to publish a physics paper 15 years ago was about $500.

[Ego]

Not a scientist, so grain of salt here but it seems to me society as a whole has grown more intolerant of disruption and has gotten good at using technology to marginalize potential threats to the status quo at earlier stages.

Understanding how to break or beat a system requires a certain knowledge of the system. Academia is set up to weed out (or drug) disruptors and reward the compliant.

Also, the world is flattening. Hierarchies to networks. Hierarchies topple conspicuously and spectacularly. Networks subsume less dramatically.

[jacob]

Not a scientist, so grain of salt here but it seems to me society as a whole has grown more intolerant of disruption and has gotten good at using technology to marginalize potential threats to the status quo at earlier stages.

I doubt that's deliberate. The water of the fish is "technique". Technique is essentially a context-free procedure, like a 10 step plan, and we're schooled in being flexible when it comes to technique. Breaking out of technique requires learning things outside the technique-world and that is not taught unless you're lucky enough to have a mentor.

As technology gets better and better, it can often replace technique.

Figure this as a self-reinforcing process that works until it doesn't. (Technique is blind to non-technique problems. The metacrisis is a consequence of going down the rabbit hole of technique.)

[ducknald_don]

Not a scientist, so grain of salt here but it seems to me society as a whole has grown more intolerant of disruption and has gotten good at using technology to marginalize potential threats to the status quo at earlier stages.

I wonder if aging populations will have more of a problem with this. It certainly seems that people become more conservative as they age and that political parties respond to this.

[suomalainen]

This seems relevant:

https://noahpinion.substack.com/p/the-third-magic

The gist being that rather than focus on specific ideas, think more about the paradigms that produced and/or nurtured those ideas. First was keeping ideas alive through recordation (oral tradition or writing) so that future generations could build on prior advances. The second was "science", pretty broadly defined. The third, coming wave, is AI's ability to squeeze out truths from mounds of data, even if the process looks like data -> black box -> "truth" (for some measure of the word). The author's thesis applied to the patent idea could be that in many branches of science, the lemon's been squeezed. Perhaps AI is the paradigm that finds unsqueezed pockets or even whole new fruit. I have my doubts about that last part, but an interesting framing of societal development nonetheless.

[mountainFrugal]

Perhaps AI is the paradigm that finds unsqueezed pockets or even whole new fruit. I have my doubts about that last part, but an interesting framing of societal development nonetheless.

At least within my field sophisticated statistical models are getting better at predicting overall system behavior for more and more interactions between previously experimentally isolated variables. Many of these models are not at the complete black box phase, but it does not take many more layers to make them so. Whole new fruit would be a challenge because usually we are asking AI to optimize for something that we already know about AND feeding in data that we already know might have some effect. Are humans smart enough to even know what questions to ask?

Cool read. Thanks for posting.

@jacob will check out that phrasing for finding affiliations. There are also groups like https://ronininstitute.org/about/ that allow you to have a place to apply for federal grants from. Institutional affiliation is not necessary to publish, but it does help in even getting the paper reviewed in the first place.

[suomalainen]

I can't help but think that [any mature] academic research field is also subject to https://meaningness.com/geeks-mops-sociopaths ... eventually everything turns into sales and management.

See also https://www.wired.com/story/tiktok-plat ... -doctorow/, which is a bit of a geeks-mops-sociopaths overlay onto social media platforms.