Hab mich damals im tiefsten Winter aus dem Wochenbett mit meinem Neugeborenen ins Jobcenter geschleppt, weil mein per Post gesendeter Umschlag mit Unterlagen angeblich nicht angekommen ist, und mir die Streichung aller meiner Gelder angedroht wurde. Ausserdem wurde mir gesagt, dass ich eine Woche Zeit habe, um den Elterngeldbescheid einzureichen (zeige mir den Menschen, der eine Woche nach Geburt den Bescheid in den Händen hält...) und es gibt ja keine Möglichkeit, da direkt mit jemandem zu telefonieren, der deine Papiere bearbeitet. Es war so ein Psychoterror. Am Ende bekommt man einen Bescheid, auf dem ein Betrag steht, der ca 20€ kleiner ist als beim letzten Mal, hat keine Ahnung warum, und man wehrt sich nicht mehr, weil man Angst hat, dass dann erst mal wieder alles gestrichen wird. Derweil muss man etliche Rückbuchungsgebühren bezahlen, weil das Geld so spät kam, dass Einzugsverfahren geplatzt sind, und das zahlt man natürlich selber und nicht die Mitarbeiter in der Leistungsstelle.
Bärbel Bas: "Wer mitmacht, der hat überhaupt nichts zu befürchten."
Nichts außer Druck & Existenzangst:
28.10.2025 07:49 — 👍 1952 🔁 662 💬 45 📌 28
xkcd: Chess Position
xkcd.com/3082
29.04.2025 23:11 — 👍 0 🔁 0 💬 0 📌 0
xkcd: Tariffs
xkcd.com/3073/
11.04.2025 11:38 — 👍 1 🔁 0 💬 0 📌 0
Meta
Addressing bias in LLMs
It's well-known that all leading LLMs have had issues with bias-specifically, they historically have leaned left when it comes to debated political and social topics. This is due to the types of training data available on the internet.
Our goal is to remove bias from our Al models and to make sure that Llama can understand and articulate both sides of a contentious issue. As part of this work, we're continuing to make Llama more responsive so that it answers questions, can respond to a variety of different viewpoints without passing judgment, and doesn't favor some views over others.
We have made improvements on these efforts with this release—Llama 4 performs significantly better than Llama 3 and is comparable to Grok:
• Llama 4 refuses less on debated political and social topics overall (from 7% in Lama 3.3 to below 2%).
• Llama 4 is dramatically more balanced with which prompts it refuses to respond to (the proportion of unequal response refusals is now less than 1% on a set of debated topical questions).
• Our testing shows that Llama 4 responds with strong political lean at a rate comparable to Grok (and at half of the rate of Llama 3.3) on a contentious set of political or social topics. While we are making progress, we know we have more work to do and will continue to drive this rate further down.
We're proud of this progress to date and remain committed to our goal of eliminating overall bias in our models.
Meta introduced Llama 4 models and added this section near the very bottom of the announcement 😬
“[LLMs] historically have leaned left when it comes to debated political and social topics.”
ai.meta.com/blog/llama-4...
05.04.2025 22:08 — 👍 135 🔁 38 💬 5 📌 61
Interested in updates from researchers who are committed to advancing the methods of communication research and related disciplines? The Count has something for you... go.bsky.app/ByqiKgq
Feel free to share and comment, if you want to be added.
16.12.2024 08:44 — 👍 52 🔁 27 💬 30 📌 3
Sounds promising 🤞
02.12.2024 15:36 — 👍 0 🔁 0 💬 0 📌 0
If you, like me, don’t want to divide your attention between several apps and timelines, check out @openvibe@mastodon.social . Using it for a few days now and I love the combined Mastodon & Bluesky timeline (Threads & Nostr is also an option). Feels like 2009er Twitter again 🤓
18.11.2024 17:02 — 👍 0 🔁 0 💬 0 📌 0
My Mastodon profile says:
“if it is interesting, boost it. You are not on Twitter”.
📣 Same thing here 🦋
Share what’s interesting. A like is nice, but just a like for the author. No algorithmic amplification. 💥
16.11.2024 13:13 — 👍 52 🔁 21 💬 4 📌 1
Guardian will no longer post on Elon Musk’s X from its official accounts
Platform’s coverage of US election crystallised longstanding concerns about its content, says Guardian
There's been an influx of new followers & new people on this platform. Perhaps not a coincidence as staying in the other one is becoming less and less justifiable...
13.11.2024 14:32 — 👍 13 🔁 3 💬 0 📌 0
Which Best Describes Your Education? You Have:
22,531 Respondents Dem. Rep. Never attended college
15% 37% 62% Attended college but received no degree
26% 49% 50% Associate's degree (AA or AS)
16% 42% 55% Bachelor's degree (BA or BS)
24% 54% 44% An advanced degree after a bachelor's degree (such as JD, MA, MBA, MD, PhD)
19% 61% 37%
It's the education, stupid …
#election #Trump abcnews.go.com/Elections/ex...
06.11.2024 14:30 — 👍 2 🔁 1 💬 0 📌 0
Are u working on social media data in 🇩🇪 - the social media observatory smo.leibniz-hbi.de has your back - @flxvctr.bsky.social presents a really impressive overview about the great service, the project offers
26.04.2024 11:45 — 👍 4 🔁 4 💬 1 📌 0
Ich habe mich mit @johannasebauer.bsky.social über Krieg und Elmo unterhalten: leibniz-hbi.de/de/blog/naho...
29.01.2024 17:16 — 👍 0 🔁 0 💬 0 📌 0
Was die Berliner CDU mit "Verkehr für alle meint".
04.12.2023 08:59 — 👍 284 🔁 81 💬 3 📌 1
Happy that one of my favourite science channels "Kurzgesagt" wades deep into my research field (social media, networked public spheres, (myth of) filter bubbles, polarisation) (even 'cites' my PhD supervisor @snurb.bsky.social www.youtube.com/watch?v=fuFl... )
30.11.2023 15:58 — 👍 6 🔁 1 💬 0 📌 0
Das neue M&K-Themenheft beschäftigt sich mit #Telegram in der politischen Öffentlichkeit. Herausgegeben von mir, @janschmidt.bsky.social, Jan Rau, @flxvctr.bsky.social und Philipp Kessling. Als Open Access verfügbar.
www.nomos-elibrary.de/10.5771/1615...
22.11.2023 12:22 — 👍 10 🔁 8 💬 0 📌 0
(Das heißt nicht, dass das 1933 nicht anders war, aber die Gleichsetzung damaliger Millieus und Bildungsstände mit den heutigen führt hier zu einem Fehlschluss.)
10.11.2023 18:55 — 👍 1 🔁 0 💬 0 📌 0
Hier Hessen: www.sueddeutsche.de/politik/hess...
10.11.2023 13:16 — 👍 2 🔁 0 💬 1 📌 0
Hier noch die Bundesebene: de.statista.com/statistik/da...
10.11.2023 13:07 — 👍 1 🔁 0 💬 1 📌 0
Wenn wir uns die Meinungsführer ansehen, mag das stimmen. Die Empirie widerspricht aber, sobald es um die Wähler geht: www.tagesschau.de/wahl/archiv/...
10.11.2023 13:04 — 👍 4 🔁 1 💬 1 📌 0
Socioviz is a social media analytics platform powered by Social Network Analysis metrics. #SocialNetworkAnalysis #NetworkAnalysis #DataViz #ddj
www.socioviz.net
Climate activist, born 364ppm, based in Berlin
Philosopher from the slums. Latina. She/her.
PhD candidate in Exp Phil. Data Scientist in the making. Interested in Global Politics, Social Psychology, Technology, AI Policy. Working in #Polarization, #MotivatedSocialCognition #Misinformation #SocialMedia
Sociologist at @gesis.org working with EU-SILC in @fdzgml.bsky.social. Research interests: econometrics, survey methodology, sociology of wealth, family sociology, integration/migration, ethnic segregation. Typos galore🌻🏴🚩
Prof of Computer Science at USC
AI, social media, society, networks, data, and
HUMANS LABS http://www.emilio.ferrara.name
phd student @usc-isi | misinformation, networks, nlp
PhD candidate @QUT DMRC| AI-generated content, culture production, creative industries🤖🎨🪄| bubble milktea expert🧋
Information ecosystems & far-right political communication | IRex Tübingen |Weizenbaum Institute
Passionate poster 🦭
World-leading communication, media, and law research for a flourishing digital society. research.qut.edu.au/dmrc/
Academic @QUT; Researching various things, all related to antagonism/agonism/discourse online
🌐: https://www.essideh.com
Personality psych & causal inference @UniLeipzig. I like all things science, beer, & puns. Even better when combined! Part of http://the100.ci, http://openscience-leipzig.org
Journalistin, Juristin, Jongleurin mit den Bällen, die das Leben Dir so zuwirft.
Communication researcher at @uni-muenster.de
Opa sagt: Der Junge isst wie‘n Teufel.
Computational social science researcher. Assistant Prof at Rissho Univ, Japan. Ph.D. in Kyoto Univ.
Computational Social Science PhD candidate @QUT
Investigating the complex interplay between search engines, GenAI, and politics/news
Entrepreneur's mind. Athlete's body. Artist's soul ~ J.Clear
https://www.shir-weinbrand.com/
Canada Research Chair, Professor at Toronto Metropolitan University & co-director & researcher at @socialmedialab.ca | Social Media and Network Scholar | My opinions are my own
Co-Dir. SocialMediaLab.ca | M.S., J.D. | Developer of social media analysis tools for research in the public interest | PoliDashboard.org, Communalytic.org, DeepfakesTracker.org, ConflictMisinfo.org, NoteTracker.socialmedialab.ca, & KM.socialmedialab.ca
Professor of Data & Knowledge Engineering at Heinrich-Heine-University Düsseldorf. Head of Knowledge Technologies at GESIS - Leibniz Institute for the Social Sciences (Cologne). Views are my own.
http://stefandietze.net
Researching the Internet is serious business. Mostly.
(We're mainly over on AoIR.social, our organisation Mastodon instance, which is open to all members.)
https://aoir.org/
https://aoir.social/@aoir