Releasing Common Corpus the largest public domain dataset for training LLMs Common Corpus includes 500 billion words from a wide diversity of cultural heritage initiatives. is multilingual and the largest corpus to date in English, French, Dutch, Spanish, German and Italian.

KI meets Kolonialismus: 10.000 Akten des Reichskolonialamtes haben wir in einem KI-Projekt entziffert. Texte in dt. Kurrentschrift sind gut lesbar & durchsuchbar – wichtige Quellen zum Unrecht dt. Kolonialherrschaft. Nutzbar zunächst in Lichterfelde. ➡️ bundesarchiv.de/DE/Content/Pre…

2006 paper on cause of Alzheimer’s is the field's 4th most cited paper ever. But the author faked the data and it was later retracted. This wasted many years & $287+ million in funding as the field chased a dead end. Scientific fraud must be criminalized. chris-said.io/2024/06/17/the…

Lesenswerte Einordnung des #Compact- Verbots von NZZ. nzz.ch/feuilleton/ver…

OpenAI hat seine lange erwartete KI-Suchmaschine "SearchGPT" vorgestellt futurezone.at/digital-life/s…

Danke, Sportschau, wir übernehmen ab hier...

Unesco zeichnet Residenzensemble Schwerin als Welterbe aus 😇  spiegel.de/kultur/schweri…

We have been so inspired by the amazing athletes at the #ParisOlympics, we decided we needed to celebrate with Games of our own. 📚🥇🥈🥉🤓

Citations for cash: researchers have identified services where scholars can buy fake references in bulk go.nature.com/4csIaxW

Austausch zum Thema #Kolonialismus: Unsere Reichskolonialamt-Akten sind nun auch per #KI durchsuchbar. Mit Expertinnen & Experten haben wir diskutiert, wie sich das Interesse an Quellen steigern & ihre Einordnung verbessern lässt.@TahirDella Torsten Musial bundesarchiv.de/nachricht/roun…

Maybe Trump should stop messing with the Haitians Just saying 🤷🏼‍♂️

It’s January, which means another batch of copyrighted work is now public domain arstechnica.com/culture/2025/0…

Es ist nicht rechtsextrem, die Wahrheit zu sagen tagesanzeiger.ch/grooming-gang-…

Estimate the total costs of your academic publications using this tool built on OpenAlex data! #opendata

New study reveals that around 20,000 scientists may be publishing an "implausibly high" number of papers, raising concerns about academic integrity and the pressure to prioritize quantity over quality. #AcademicPublishing #ResearchIntegrity #sciencenews cen.acs.org/policy/publish…

The entire staff of the Institute of Museum and Library Services has been placed on administrative leave. npr.org/2025/03/31/nx-…

There is a new platform, Sci-Net (sci-net.xyz) where researchers can help each other by providing paywalled papers. The platform builds upon Sci-Hub meme token $scihub to reward active members.

AI-Reddit study leader gets warning as ethics committee moves to ‘stricter review process’ retractionwatch.com/2025/04/29/eth…

How generative artificial intelligence is changing work at a national laboratory anl.gov/article/how-ge…

Breaking the #Code: Multi-level Learning in the #Eurovision Song Contest #ESC2025 researchgate.net/publication/39…