Aaron Clauset

Fig 2. Five average productivity curves (publications per year) are shown, one for each strata of university prestige, showing a rapid rise to an early peak (around expected tenure time) followed by a slow decline to a stable, prestige-specific value. Averages over researchers at all levels of institutional prestige follow similar productivity trajectories, in agreement with the conventional narrative, but at differing scales of output.

Our 2017 @pnas.org study of computer science shows how academic productivity varies over 20 years. There's definitely a pre-tenure surge, with productivity decaying slowly after, and then stabilizing (at a level we showed in a 2022 paper is determined by lab group size)
www.pnas.org/doi/full/10....

This is unfortunate:
www.nytimes.com/2025/07/28/u...
I wrote to the president of Harvard. I hope other faculty will speak their conscience, even if it means more struggle ahead.

Applications are open for SFI's 2026 Complexity Postdoctoral Fellowships

If you’ve recently earned a Ph.D. in any scientific field and want to pursue independent, transdisciplinary research, consider applying.

Deadline: October 1, 2025
Apply here: santafe.edu/sfifellowship

Out of curiosity, aimed at other faculty:

Does your university allow you to apply for Google or Amazon research awards, or are the legal terms from the funders found to be too onerous by university grants/legal/trademark?

Our team had an amazing week at @ic2s2.bsky.social in Norrköping Sweden and we will post pictures of our posters and talks soon - the big news is that we're so excited to host #IC2S2 in Burlington in 2026! youtu.be/p412S4GnPkc

As Universities Cede to Fascism, Faculty Must Refuse to Follow Once upon a time, universities sought to be spaces where critical thought was championed, diversity was celebrated, and academic freedom was valued. Since the Civil Rights Movement, many universiti…

In my latest blog post I lay out some things we as faculty can do to ensure that higher education remains an incubator that prepares future generations for a multiracial democracy as opposed to an incubator for Trump's fascist agenda. Check it out!
educationallinguist.wordpress.com/2025/07/24/a...

That Dominique Baker can be a real downer sometimes (the entire piece is very good and a useful overview if you want to get caught up)

I've had this headline rattling around in my brain since January and I finally wrote it. Been hearing too many people in the startup world buying into the neoractionary nonsense that maybe a little light fascism is good for silicon valley. It's not. It's very, very bad.

Happily, someone else looked into this and... it kind of works: bsky.app/profile/mela...

I did not! Thanks!

Oh! This approach is basically what I had imagined as one approach. So great that someone has already looked into evolving prompt structures, and that it kind of works.

That's a good point. Makes me think that meta-heuristics for prompt optimization might work... but exploring the prompt space deeply enough seems like an important requirement.

Yes, that's a great point. In my mind, it seems like using more structured approach to explore the prompt space (eg, genetic algorithms on tree-based representations of the prompt) would be quite cumbersome. Hence my wondering!

Folks who follow LLM stuff: has anyone used LLMs to "evolve" or otherwise optimize prompts for another LLM task? Like, using an LLM to search the space of prompts for ones that make another LLM score well on some specific task.

My inescapable conclusion has become that Israel is committing genocide against the Palestinian people. Having grown up in a Zionist home, lived the first half of my life in Israel, served in the I.D.F. as a soldier and officer and spent most of my career researching and writing on war crimes and the Holocaust, this was a painful conclusion to reach, and one that I resisted as long as I could. But I have been teaching classes on genocide for a quarter of a century. I can recognize one when I see one.

"My inescapable conclusion has become that Israel is committing genocide against the Palestinian people."

Written by an Ivy League professor who studies the Holocaust; who served in the IDF.

www.nytimes.com/2025/07/15/o...

Screenshot of the article that reads: It’s not just children. In addition to bolstering underfunded K-12 schools and protecting the rights of kids with disabilities, the department also manages the federal financial aid process for college students. “I am not certain whether or not students will be able to get financial aid next year,” Dominique Baker, a professor of education and public policy at the University of Delaware, told Vox.

Screenshot of the article that reads: While students and families are the most directly impacted by changes at the Education Department, those changes have also inspired a broader concern. Some of the programs terminated by DOGE are congressionally mandated, and if the president or Musk can simply stop them, “that means that Congress no longer actually functions,” Baker said. “There are no longer checks and balances for the executive branch,” Baker said. “A significant part of this goes beyond education and speaks to a constitutional crisis that shapes the future of our country.”

Here Vox's explainer for just what is going on at the Department of Education which includes me saying I don't have faith in financial aid disbursement next year and, oh yea, I think we're in a constitutional crisis

www.vox.com/policy/40233...

Abstract The ReproSci project retrospectively analyzed the reproducibility of 1006 claims from 400 papers published between 1959 and 2011 in the field of Drosophila immunity. This project attempts to provide a comprehensive assessment, 14 years later, of the replicability of nearly all publications across an entire scientific community in experimental life sciences. We found that 61% of claims were verified, while only 7% were directly challenged (not reproducible), a replicability rate higher than previous assessments. Notably, 24% of claims had never been independently tested and remain unchallenged. We performed experimental validations of a selection of 45 unchallenged claim, that revealed that a significant fraction (38/45) of them is in fact non-reproducible. We also found that high-impact journals and top-ranked institutions are more likely to publish challenged claims. In line with the reproducibility crisis narrative, the rates of both challenged and unchallenged claims increased over time, especially as the field gained popularity. We characterized the uneven distribution of irreproducibility among first and last authors. Surprisingly, irreproducibility rates were similar between PhD students and postdocs, and did not decrease with experience or publication count. However, group leaders, who had prior experience as first authors in another Drosophila immunity team, had lower irreproducibility rates, underscoring the importance of early-career training. Finally, authors with a more exploratory, short-term engagement with the field exhibited slightly higher rates of challenged claims and a markedly higher proportion of unchallenged ones. This systematic, field-wide retrospective study offers meaningful insights into the ongoing discussion on reproducibility in experimental life sciences

So what can we do in response? A key problem:

The transformation from long- to short-term funding leads researchers to flit from topic to topic, and move across fields chasing 'sexy' stories.

Science is done better when you leave someone to geek out on a single topic for their whole career.

5/n

A lovely graphic from The Allied Genetics Conference 2020 depicting the many model organisms that contribute to our understanding of science. https://genetics-gsa.org/drosophila-2025/professional-development/careers-at-tagc/

CONCLUSIONS:

To promote the best science being done, we need to look at the incentives of science publishing and funding. Long-term funding produces better science. Short-term and topical agendas encourages irreproducibility.

Funding model organisms generates science we can build on! 🪰🪱🦠🐠🐁🐸

6/6 🧵

What makes a country great in the 21st century is science, and it’s really insane to purposefully destroy its science enterprise.

Marcia McNutt (president of @nasonline.org) should resign.
NAS needs a leader who understands what it means to forcefully defend science's unique role in seeking and speaking truth.

A Nation's Story: “What to the Slave is the Fourth of July?” In July of 1852, Frederick Douglass delivered a speech titled “What to the Slave Is the Fourth of July?,” a call for the promise of liberty be applied equally to all Americans. Douglass’s speech empha...

nmaahc.si.edu/explore/stor...

I don’t think anyone is prepared for what they just did w/ ICE.

This is not a simple budget increase. It is an explosion - making ICE bigger than the FBI, US Bureau of Prisons, DEA,& others combined.

It is setting up to make what’s happening now look like child’s play. And people are disappearing.

The evil that men do lives after them

📣 We're on the lookout for a creative postdoc with strong computational skills!

Be the go-to person in the lab for building simple but powerful simulations that test wild ideas on biological rythems: from daily cycles of mussel groups at deep sea, to firefly flash synchronization!

More info below👇

STAGGERING: This new study of 133 countries is the first to estimate the impact of all USAID’s work. In 2 decades, it has saved *92M* lives. Current cuts, if not reversed, are forecast to cost up to *14M* lives thru 2030. www.thelancet.com/journals/lan...

"There is still time to make your voices heard. The reason they are moving so quickly is because the more the public learns about the bill, they more they oppose it. The hour is late, but not too late."

Yes, the question of “quality” is always lurking, even as it’s a tricky concept to operationalize. The second half of our story of productivity points a finger pretty clearly at academic labor, and we’re upfront that our results say little about the value of the papers, only the volume.

Slide titled 'network effects in scientific labor', arguing that your own productivity is a network effect, because it includes the productivity of your coauthors too

Slide showing a worked example of a collaboration network, and using a probabilistic model to marginalize the network effects of collaboration to estimate individual productivity (without network effects)

Slides showing results of applying the network effect model showing that collaborating with a highly productive or highly prominent coauthor in your early career has a lasting effect on your own productivity and prominence, suggesting that social capital (network effects) are a kind of semi-transferrable form of wealth

Slides describing a network analysis of homophily in collaboration networks, showing that diversity in collaborators is a learned tendency, and mid-career researchers who trained in highly diverse research groups go on to lead relatively more diverse groups themselves

Slides from my keynote lecture "What drives the productivity of scientific labor?" at the 2025 Oxford Summer School on Economic #Networks. Part 2 of 2: how does who you collaborate with shape your own productivity, and how are networks like social capital? aaronclauset.github.io/slides/Claus...

Title slide: "What drives the productivity of scientific labor?" a talk by Aaron Clauset at the Oxford Summer School on Economic Networks 2025

Summary slide arguing that prestige is a structural variable in the scientific ecosystem, and that it is differences in working environment (not pedigree) that explains why elite researchers dominate scientific discourse. Ends with a question: how exactly does environment do this?

Summary slide arguing that working environment (specifically, institutional prestige) explains differences in scientific productivity because there's just more available academic labor at elite places, and researchers use this labor to write more papers. Ends with a question: how much does it matter who you collaborate with? That leads into part 2 of the slides

Slides from my keynote lecture "What drives the productivity of scientific labor?" at the 2025 Oxford Summer School on Economic #Networks. Part 1 of 2: why and how do elite scientists dominate scientific discourse?
aaronclauset.github.io/slides/Claus...

Clearly I need to step up my shitposting. As a coping mechanism, if for no other reason