Show newer

Pour @GrandjeanMartin, il faut solliciter les compétences humaines que ne peut pas imiter si l'on veut limiter la capacité des étudiants à l'utiliser dans le cadre académique.

heidi.news/articles/chatgpt-fa

Fond memories of driving a over the icy backroads of the Bavarian Prealps in the middle of the night. And compared to the VW 181, the Iltis was modern.

Show thread

First I thought, driving your open-top in winter is pretty hardcore. But The Thing doesn’t have a useful heating anyway, so I guess it doesn’t make a big difference.

Diesen Freitag um 11 Uhr treffen wir uns zu unserem Paper-Zirkel. Dieses Mal diskutieren wir mit @nilsreiter über Operationalisierung. Schau gern (virtuell) vorbei, auch wenn du das Paper noch nicht gelesen hast. Mehr dazu dhtheorien.hypotheses.org/1266 #DH #DigitalHumanities

I used to think that the Aurora reports “500 mA current required” because it uses a generic USB controller, and the value is hardcoded to the maximum. But I’m starting to think that it actually does draw a lot of power: I’ve now connected it to a USB C port, and it doesn’t seem to randomly disappear anymore…

This list of pitfalls for journalists to avoid in reporting on artificial intelligence is broadly applicable to emerging tech in general. Useful to anyone who wishes to spread insight instead marketing snake oil. aisnakeoil.substack.com/p/eigh
Many thanks to @Julia for bringing timely knowledge into my inbox:
themarkup.org/hello-world/2023 (And thanks to @craignewmark, for funding @themarkup)

Another reason why I will stop reviewing for the global conference: review responses. It’s humiliating to make authors beg reviewers for better grades. But since the abstracts remain unchanged, there is no reason to change my mind (unless I made an actual mistake, of course). If everything is meant to be negotiable, why have review criteria in the first place? Bonus: as the reviews are open, some authors will probably denounce reviewers that didn’t “correct” their grades.

Rendering #PDF documents:

#LaTeX still seems to be the best free option: It generates TOCs, indices, etc. It optimizes page breaks by moving figures. Etc.

– I convert Markdown to LaTeX to PDF (via Pandoc) and it works really well. To customize the results, you have to know *some* LaTeX but not much: I mostly web-search & paste.

– LaTeX engine I discovered today: tectonic-typesetting.github.io

– There are a few CSS-based engines but no free options that come close to LaTeX(?)

#Publishing

Reading a paper in English by a scholar trained primarily in France. It’s well written, without any obvious issues. When you look closer, though, you find quite some occurrences of these semi-false friends like affirm, implant, consolidate, suppress, valorize, etc.

I caught myself recently using “suppress” with the meaning of “supprimer” :-/

A RAM module (magnetic cores) of a German "Kienzle Fakturier-Maschine" (rb. 60s) from the Kollektion of horniger/Berlin:

#HardwarePreservation #Minicomputers

Wir laden ein: Am Freitag, den 3.2.22 findet wieder unser Paper-Zirkel statt. Dafür haben wir @nilsreiter mit seinem brandneuen Paper zur Operationalisierung in den #DigitalHumanities eingeladen. Mitdiskutieren kann jede*r (auch ohne das Paper zuvor gelesen zu haben): dhtheorien.hypotheses.org/1266 #DH

Show older
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.