Fiona Bradley

@fionabradley.bsky.social

Library person and PhD candidate (IR & PolSci) Many hats on various committees. I chair IFLA's Open Science & Scholarship Advisory Committee.

This is not the tone of podcast I usually listen to, but this one is worth a listen for the commentary on how people are getting way too caught up in pride and identity in their AI hot takes. A big reason why LinkedIn has become absolutely nauseating IMO.

Ed Zitron@edzitron.com · 2mo ago

Had @author-calnewport.bsky.social on Better Offline to talk about AI companies “doom-trolling” consumers using scary stories about conscious computers, and how we need to make fun of those leading the scare-mongering. podcasts.apple.com/us/podcast/b... Linktr.ee/betteroffline

I don't think the old paradigm is completely gone. There are still remnants, spaces that operate somewhat differently, but creator-centered, parasocial spaces that ride the waves of algorithmic recommendation seem to be dominating our attention, and (re)shaping the flows of information & influence.

If you are in your late thirties to mid-forties right now, there is a good chance that you have spent most of your life in a cycle of making some sort of home on the internet only to have it crumble beneath you like chalk and having to start over.

I used this as a framing example in a recent talk I gave about the "librarian's dividend" which is my probably-far-too-positive spin on this absolute garbage fire -- that at least maybe we're forced into more appreciation and value for the people and institutions who help us evaluate information.

The second, on June 25th, will be a talk about the sequel, of sorts, to my book Calling Bullshit. "Modern Day Oracles or Bullshit Machines?" is the new, free online course that Jevin West and I developed about how to think, learn, and thrive in a ChatGPT world.

Modern Day Oracles or Bullshit Machines: Teaching students to think, learn, and thrive in a ChatGPT world

Join Carl Bergstrom in a talk that helps teach students to think, learn and thrive in a ChatGPT world.In this session, Carl will provide an overview of the course Modern Day Oracles or Bullshit Machin...

education.unsw.edu.au

Australian friends! I'm going to be visiting Sydney in just over a week. I'll be at UNSW on June 23-25th and Macquarie on June 26-27th. I'd love to catch up with people in person, and also will be giving (at least) two talks at UNSW. The first is science-of-science modeling talk, on June 23:

UNSW School of BABS Special Seminar


Peer review meltdown
Carl T. Bergstrom, University of Washington

Host/Chair Prof Mark Tanaka

Monday 23 June 2025

1pm-2pm 

Mathews Theatre C

You’ve seen it yourself. Peer review is coming apart at the seams. Editors face a mighty struggle to recruit reviewers. Researchers are overwhelmed with review requests. Authors wait months or longer for low-quality reviews of their work. In this talk, I present a series of simple mathematical models to illustrate what is happening and why. (1) An elite journal relies on peer review to identify the top papers; knowing the quality of the peer review process, authors self-screen and send only their best work to this journal. But when the reward from publishing in the elite journal increases, submission volume increases. (2) When submission volume increases, review quality drops as the most qualified reviews are no longer available — but we prove that when review quality drops, submission volume necessarily increases as more authors try to sneak in undeservedly. This feedback process swamps the journals with submissions and erodes the quality of review. (3) We next consider what happens as elite journals proliferate and show that, paradoxically, as the number of elite journals increases, researchers self-screen more assiduously, but the review load continues to increase. To illustrate the consequences, we consider welfare measures for authors, reviewers, and readers. (4) Finally, we explore the way in which aggressive desk rejection policies can partially check this peer review meltdown.  

Ts&Cs of these sites are traditionally exceedingly complicated and one-sided (in favour of the platform). Traditional publishers also don't tend to look kindly on the inclusion of VoRs on these sites (which don't have CC BY licences). Please read what you sign! #academicsky #skybrarians #oa

Fiona Bradley@fionabradley.bsky.social · last yr.

We don't talk often enough about the problems of academic 'social networking' sites like Academia and ResearchGate for researchers and knowledge dissemination. Put your work in an institutional or disciplinary repository instead. They won't AI your work.

We don't talk often enough about the problems of academic 'social networking' sites like Academia and ResearchGate for researchers and knowledge dissemination. Put your work in an institutional or disciplinary repository instead. They won't AI your work.

Dr Dean Burnett@drdeanburnett.bsky.social · last yr.

This is my PhD thesis I did not ask for this I did not consent to this I did not approve of this I was not compensated for this I would not have advised this I do not like this And worst of all, the number of people who've read my thesis has still not increased.

Email saying "An AI created a podcast of your paper"

Said paper is my PhD dissertation, "The Role of The Hippocampus In Configural Learning"

As someone currently in the rewrite, craft, polish final stages of thesis writing, this. Also - Bender and Hanna's book 'The AI Con' is excellent and a great argument for the enduring value of longform writing. It's starting to feel like reading is becoming an act of resistance against AI.

Alex Hanna@alexhanna.bsky.social · last yr.

There's this fundamental mismatch in what AI people seem to think research is. As if you have a kernel of an idea and one needs the right re-combination of tokens to have some kind of insight. Research should be surprising, thought-provoking, (re-)discovering new literatures and subcommunities.

Congratulations Te Paea Paringatai (Waikato, Ngāti Porou) who will become the first IFLA (International Federation of Library Associations) President from Aotearoa New Zealand in 2027. Te Paea's role as President-elect begins in August 2025.Te Paea's election result was a landslide and is a

IFLA Elections 2025 Results: Governing Board

President-elect (one place) To serve August 2025-August 2027 as President-electTo serve August 2027-August 2029 as President The President-elect is an elected position to the Gove...

ifla.org

Reset is the show I heartily recommend most often to people who've never watched C-drama (and who are unlikely to sit through dozens of episodes of a historical palace drama or wuxia, though my goodness, Guardians of the Dafeng!)

Post nicht verfügbar.