Lightnews — Scholar-powered news

Berna Devezer

@devezer.bsky.social

9.8K followers 930 following 27K posts

Metascientist @ uidaho. I work at the intersection of behavioral sciences, statistics, and philosophy. Love thinking and talking about science. Post lots of cat and food pics. Allergic to unsolicited advice.

Posts Media Videos Starter Packs

Pinned

Berna Devezer @devezer.bsky.social · Oct 6

Article fresh off the print (no paywall)! The tl;dr is in the quoted thread. We advance a measure of replication distance to capture differences btw an original study & its replications. We also discuss how statistics needs to go beyond ideal assumptions to serve science better.

#metasci #sts #stat

Berna Devezer @devezer.bsky.social · 50m

I like it. I believe it. I feel it.

Reposted by Berna Devezer

Andrea Stocco @andreastocco.com · 10h

A recent Nature paper has introduced Centaur, a LLAMA-based LLM that surpasses any cognitive models in predicting human behavior. Is this the ultimate unified theory of cognition? In this preprint, my colleagues and I expressed our doubts. Let us know what you think! arxiv.org/abs/2510.03311

Not Even Wrong: On the Limits of Prediction as Explanation in Cognitive Science

We offer a comment on the Centaur (Binz et al., 2025) transformer-based model of human behavior. In particular, Centaur was cast as a path towards unified theories of cognition. We offer a counter cla...

Reposted by Berna Devezer

Andrew Heiss @andrew.heiss.phd · 1h

I really really really liked The Oatmeal’s take on AI here theoatmeal.com/comics/ai_art

A cartoonist's review of AI art - The Oatmeal

This is a comic about AI art.

Berna Devezer @devezer.bsky.social · 15h

Cannot not respond to demand!!

Berna Devezer @devezer.bsky.social · 20h

Absolutely! I will be sending this around for sure.

Berna Devezer @devezer.bsky.social · 22h

Feta makes it extra creamy and tangy. And my Aegean palate prefers tangy over not-tangy every time.

Berna Devezer @devezer.bsky.social · 22h

Luckily we have a definition! 😬

Alan Richardson @alanrichardson.bsky.social · 23h

UBC did everyone a favour and defined excellence in our strategic plan.

Berna Devezer @devezer.bsky.social · 22h

🤩 Told'ya! (Seriously though, thanks for sharing the original recipe 🙏🏼)

Berna Devezer @devezer.bsky.social · 22h

😭😭😭

Berna Devezer @devezer.bsky.social · 23h

Maybe if they're measuring it on a social excellence scale...

Berna Devezer @devezer.bsky.social · 23h

LOL if he did it then...

Berna Devezer @devezer.bsky.social · 23h

😭 I haven't checked the COI on those papers but I guess I wouldn't be surprised.

Berna Devezer @devezer.bsky.social · 23h

LMAO that makes it much easier to achieve. Oh, you know, my research currently sucks but I always aspire to make it better. That makes me a pretty excellent researcher, thank you.

Berna Devezer @devezer.bsky.social · 23h

Exactly!

Berna Devezer @devezer.bsky.social · 23h

Operational excellence: TBD (Form does not work.)

Berna Devezer @devezer.bsky.social · 23h

Teaching excellence: Surprisingly, we don't know how to measure this either so you better get good student evaluations regardless of your topic, class sizes, difficulty of material, and biases directed at your identity. If you need to entertain the students for that, yup that works for us too.

Berna Devezer @devezer.bsky.social · 23h

Research excellence: We don't know how to measure research quality and we sure won't be reading any papers. So please make our jobs easier by providing some metrics that look good and a lot of publications, just to be on the safe side, and have demonstrable impact (by some terrible metric) and...

Berna Devezer @devezer.bsky.social · 23h

Which, you might be surprised, is still not there after 7 weeks of bickering on my part.

Reposted by Berna Devezer

Needhi Bhalla 💅🏽 @needhibhalla.bsky.social · 23h

www.nature.com/articles/pal...

Berna Devezer @devezer.bsky.social · 1d

As an aside, "Excellence" must be the single-most overused, empty, boilerplate word used in academic bureaucracy and it's typically a stand-in for "We don't know what we want so you should do everything".

Berna Devezer @devezer.bsky.social · 1d

Admin email from university: "Share your ideas: Help shape operational excellence at U of I"

Soliciting ideas for how to improve operational processes.

Except...

The form doesn't work—the form created by the Operational Excellence team.

Yeah.

Berna Devezer @devezer.bsky.social · 1d

They're committed to the bit.

Berna Devezer @devezer.bsky.social · 1d

Admin email from university: "Share your ideas: Help shape operational excellence at U of I"

Soliciting ideas for how to improve operational processes.

Except...

The form doesn't work—the form created by the Operational Excellence team.

Yeah.

Berna Devezer @devezer.bsky.social · 1d

“'It’s really on all of us to try to do our part in reducing emissions,' Feldman said."

Sounds very much like: People should stop their life-saving medications and sacrifice themselves so the rest of us can keep flying in planes, driving our cars, and use ChatGPT to create memes and write email.

CNN @cnn.com · 1d

The people who are most vulnerable to the hard-to-breathe air that comes with climate change may inadvertently be adding to the problem, new research finds. https://cnn.it/4pWQMoh

Berna Devezer @devezer.bsky.social · 1d

It is definitely a huge mistake to tout individual results as the last word on something instead of what it is, a possible, promising new venue for exploration, we're setting ourselves up for failure, and even, for many made-up crises.

Berna Devezer @devezer.bsky.social · 1d

I think that is a problem but it's not a problem of "replication". It is a problem of ignoring uncertainty in each study and treating their results as facts. No one study can make a scientific discovery singlehandedly. We need serious exploratory work, triangulation, and a programmatic approach.