This is a really excellent talk at #openrepos2025 by Ben Zhao about the limits of and misconceptions about LLMs, that is tailored for the #libraries #archives and #digitalpreservation crowd. He focuses in on how they work, and don't work -- and bridges that to a discussion of the impacts that the technology is having on the larger web (bots).

It gives me no measure of good vibes to know that a big crowd of people in my profession got to hear this for a keynote about AI.

https://www.youtube.com/watch?v=JicpcYwQe3w

We sadly missed this year's Open Repositories conference, which was organised in Chicago a couple of weeks ago.

But you can get some idea of the topics that were being discussed by browsing the presentation slides, which are now available in Zenodo:

https://zenodo.org/communities/openrepos/records?q=&f=subject%3AOR2025&l=list&p=1&s=20&sort=newest

#openrepos2025 #repositories

The depositar team were glad to attend the 20th International Conference on Open Repositories (#openrepos2025) earlier this month at the University of Chicago. Li Fan Wang presented "Automated Data Analytics: A Statistical Dashboard Built with GitLab CI/CD for a Data Repository Based on #CKAN" at Developer Track Session 2 on June 17. The presentation slideset, as well as the session proposal originally submitted to the conference, can be found at the depositar:

https://pid.depositar.io/ark:37281/k5w9w7145

Noch frisch zurück von den #OpenRepos2025 wünschen wir allen Teilnehmenden der #BiblioCon25 viel Spaß und verfolgen die Konferenz gespannt über Mastodon!

Kommende Woche findet dann nach diesem Konferenzblock wieder die #DSpace Open Hour statt: Dienstag, der 01. Juli, wie immer um 14 Uhr.

Die Zugangsdaten gibt es im DSpace-Slack oder hier via Private Message.

During our presentation at #OpenRepos2025 somebody asked about how do we handle saying no to a feature or even gauging the size and effort that it will take to implement a new feature.

Estimating in software is really hard, I even alluded to this XKCD comic about it: https://xkcd.com/1425/

The crux of my answer was: don't answer on your own, instead bring the question to the development team and let us review it. Some things that look easy are very hard to implement and the reverse it's true.

Ben Y. Zhao dropping real wisdom demystifying AI on the #OpenRepos2025 closing keynote about

“Moral of the story: You cannot build a reliable tool using broken parts.”

#ai

Great point by Sarah Barsness:

“Accessibility is a process. Technical standards change, user needs change, available tools change. Laws and policies serve as a starting point.”

https://www.conftool.net/or2025/index.php?page=browseSessions&ismobile=true&form_session=524

#OpenRepos2025

🎤 Live from #OpenRepos2025 in Chicago!

Julien Sicot and Jorge Rodrigues de Matos are presenting the recent evolution of EPFL’s institutional repository - Infoscience, including new automation features for data import.

Great to see our work shared and discussed with the international Open Science community!