https://fragdenstaat.de/artikel/klagen/2026/03/springer-verliert-gegen-fragdenstaat-vor-gericht/?pk_campaign=mastodon
| Where | #Berlin #Brussels #Munich |
LLMs are engines for “more of the same” - they homogenise language and ideas: culturally hazardous. This paper shows massive similarities across responses from 25 models
https://arxiv.org/abs/2510.22954
A great video outline: https://youtu.be/0PB09fsydZE?si=S0XwcRhBJleyEdpC

Language models (LMs) often struggle to generate diverse, human-like creative content, raising concerns about the long-term homogenization of human thought through repeated exposure to similar outputs. Yet scalable methods for evaluating LM output diversity remain limited, especially beyond narrow tasks such as random number or name generation, or beyond repeated sampling from a single model. We introduce Infinity-Chat, a large-scale dataset of 26K diverse, real-world, open-ended user queries that admit a wide range of plausible answers with no single ground truth. We introduce the first comprehensive taxonomy for characterizing the full spectrum of open-ended prompts posed to LMs, comprising 6 top-level categories (e.g., brainstorm & ideation) that further breaks down to 17 subcategories. Using Infinity-Chat, we present a large-scale study of mode collapse in LMs, revealing a pronounced Artificial Hivemind effect in open-ended generation of LMs, characterized by (1) intra-model repetition, where a single model consistently generates similar responses, and more so (2) inter-model homogeneity, where different models produce strikingly similar outputs. Infinity-Chat also includes 31,250 human annotations, across absolute ratings and pairwise preferences, with 25 independent human annotations per example. This enables studying collective and individual-specific human preferences in response to open-ended queries. Our findings show that LMs, reward models, and LM judges are less well calibrated to human ratings on model generations that elicit differing idiosyncratic annotator preferences, despite maintaining comparable overall quality. Overall, INFINITY-CHAT presents the first large-scale resource for systematically studying real-world open-ended queries to LMs, revealing critical insights to guide future research for mitigating long-term AI safety risks posed by the Artificial Hivemind.
RE: https://sueden.social/@kinghaunst/116193904891499184
Dieser Typ, der Kanzler sein will, hat von Leben doch überhaupt keine Ahnung.
Repeatedly getting Covid doesn’t build immunity, ‘it’s more like accumulating damage’
Leading vaccinologist Helen Petousis-Harris says there's not enough awareness of the impact of Covid-19 on overall health.
"What many people don't realise is that getting Covid repeatedly is not like building immunity, it's more like accumulating damage," she says.
Source: https://archive.md/Qsv13
Maybe it is time to revisit the 10-Point Plan to Cut Oil Use?
This could reduce global oil demand by 2.7 million barrels a day in 4 months.
Source: https://www.iea.org/reports/a-10-point-plan-to-cut-oil-use
Sky Glows over Paranal Observatory
Image Credit & Copyright: Julien Looten
Explanation: Are lasers from giant telescopes being used to defend the Earth? No. Lasers shot from telescopes are now commonly used to help increase the accuracy of astronomical observations. In some directions, Earth atmosphere-induced fluctuations in starlight can indicate how the air mass over a telescope is changing, but in other directions, no bright star exists. In these directions, astronomers can create an artificial star with a laser. Subsequent observations of the artificial laser guide star can reveal information so detailed about the changing blurring effects of the Earth's atmosphere that much of it can be removed by rapidly flexing a telescope's mirror. Such adaptive optics techniques allow high-resolution ground-based observations of real stars, planets, and nebulas. Pictured here, telescopes at Paranal Observatory in Chile study a colorful sky filled with green airglow and the Magellanic Clouds on the left, red airglow on the right, and the majestic central band of our Milky Way Galaxy arching across the center.
Klima im TV? Immer weniger. 📉
Wir haben ausgewertet, wie viel Klimaberichterstattung bei Das Erste und ZDF im Gesamtprogramm vorkommt. Das Ergebnis ist alarmierend – ein Thread. 🧵👇
#primetimefuersklima #wirfuerkv8 #medienklimakrise #klimavor8