Mal 'ne ganz pragmatische Frage:
Wenn sich @debian jetzt nicht klar vom Einsatz von #AI / #KI / #LLM distanziert, …

… puh. Wie geht Ihr damit um?

Wie hält es die @fsfe oder das GNU Projekt ?
Was ist mit der viel zitierten Freiheit, mit der Transparenz, Unabhängigkeit und der Mündigkeit (der Entwickler)?

Gentoo? Kenne ich noch nicht. Ist aber vielleicht eine Idee wert. 🤔 Aber auch da werde ich recht schnell z.B. auf diesen Linux-Kernel treffen. …

Every post/reply about AI you write should come with a link to a reminder that Corporate LLMs:
- Are for-profit plagiarism machines that have cannibalized the internet
- Come at a serious human and environmental cost
- Can't be trusted and create real IP problems, AND
- Can make you dumber if you stop thinking for yourself

#ai #noai

"Die Bilanz sendet ein Warnsignal

Die offenen Kundenforderungen stiegen innerhalb von sechs Monaten von 38,5 auf 63,1 Milliarden US-Dollar. Fünf direkte Kunden stehen für zusammen 70 Prozent dieser Summe. Solange diese Unternehmen zahlen, bleibt das beherrschbar. Gerät jedoch nur einer der Großkunden in Schwierigkeiten, könnte aus gebuchtem Umsatz ein handfestes Bilanzproblem werden."

https://www.wallstreet-online.de/nachricht/21301828-nachgehakt-nvidia-zerfaellt-kartenhaus

#ai #aibubble #aibubbleburst #nvidia

AI systems are stochastic parrots. They mix and match what they have learned by reading what we have produced over the decades.

Some people have told me that AI systems suggest solutions similar, or identical, to those I was already suggesting in some posts on my blogs back in 2020, 2021, 2022... before the AI era. And yet, every now and then, someone writes to me saying that one of my blog posts from back then is "clearly AI generated because AI suggests the same method".

This phenomenon is becoming fairly recurring, and I know at least two people who have stopped contributing publicly to Open Source and education because they got tired of being accused of producing slop.

If we want to stay away from all this, if we want to demonstrate that AI is harmful and not a solution, shouting at anything that "might look AI generated, and often isn't" is not how we solve the problem. Let's produce good content. Let's explain the long term cost of all this. Let's do useful things. Let's show what the human mind is capable of doing great things.

Because otherwise, we risk the opposite: only AI systems and vibe coders will remain, while everyone else, disgusted, walks away for good from this wonderful world of ours.

If we want a world with less AI slop, let’s not make it unbearable for humans to create.

#AI #Tech #IT

It amazes me and, if I'm being honest, frustrates me and angers the hell out of me, at how many people are okay with using AI. I mean, they just don't give a shit about any of the ethical issues with its use because it "makes their work easier" or they're "afraid of being left behind". If you use AI, you're telling me that you are okay with art being stolen, personally created content being stolen, energy prices surging for so many folks who are already struggling, clean water sources being used up, contributing to climate change (see the latest news in Nepal), making the rich even richer while we all struggle, and I could go on, and on, and on.

This is really stressing me out so fuck all of it.

#noai #fuckai #ai

Interesting: i gave claude and Qwen 3.5 9B the same problem to solve do a small home assistant automation. Qwen was running on my 5.5 year old Laptop (8 core Ryzen, 32gig Ram no useable GPU(at least i think so)) claude was done in 20sec, my Laptop needed 240sec but the endresult was nearly identical and correct. Also the annotations were nearly identical. Okay it is slow but better use my green power here than using claude which site in some coal powered AWS etc datacenter
#ai #janai #qwen3.5

#noai #ai #slop #antislop #hellokitty #art #comic #llm

Alt...

never slop hello kitto

Der Artikel "Who checks what AI can do?" ist am 20. August 2026 im Wissenschaftsmagazin Science erschienen (Vol. 393, Issue 6813) - und wurde ins Deutsche übersetzt.
https://www.science.org/doi/10.1126/science.ael2161 #AI

Überlegung zur Kennzeichnung von AI-Web-Inhalten

Dirk hat mit „Handgedacht“ mal eine Überlegung in den Raum geworfen, dass in Zukunft lieber Non-AI-Inhalte gekennzeichnet werden, als AI-Inhalte. Thomas hat das direkt mal aufgegriffen und /handgedacht, also eine Slash-Page, in den Raum geworfen.

zum Blogpost…

#blogpost #blog #Blogs #AI #Kennzeichnung #MetaTag #SlashPages

🧬 How can conflicting membrane #protein data be used reliably for research & #AI? MetaMP harmonises 4,000+ protein structures from major databases, making discrepancies traceable and supporting benchmarking & #ML. #RKI #Bioinformatics

Read more
🔗 https://spj.science.org/doi/10.34133/csbj.0165

There is something deeply satisfying about true data sovereignty.

No rate limits, no third-party APIs snooping on prompts, and zero cloud lock-in. Just raw self-hosted inference pulling 154 tok/s with an absurd 0.11s TTFT on local silicon.

Running Qwen with reasoning enabled, NVFP4 quantization on Blackwell, and SGLang speculative decoding. Crisp, private, and blazingly fast.

The future of AI belongs on your own hardware.

#SelfHosted #LocalAI #DataSovereignty #OpenSource #MachineLearning #SGLang #Blackwell #Linux #AI

Alt...

Screenshot of a Linux terminal session on hostname token-dispenser. The top section displays the nvidia-smi output showing an NVIDIA RTX PRO 6000 Blackwell GPU utilizing ~87.6 GB of VRAM, running at 44°C and 354W power draw with processes python3 and sglang::scheduler. The bottom section shows the output of a prompt script asking for a haiku about a GPU shortage, followed by performance metrics: 154.0 tok/s, 837 tokens generated in 5.44 seconds, TTFT 0.11s, reasoning on, using the qwen38-uncensored model with DFlash2 speculative decoding and NVFP4 on Blackwell architecture.

Chinese #AI Models Overtake American Rivals https://techstrong.ai/articles/chinese-ai-models-overtake-american-rivals/ by @sjvn

The headlines are all about Anthropic and OpenAI, but users are all about Moonshot, DeepSeek, and other cheaper Chinese AI models.

nVidia übernimmt Huggingface für absurden Preis

Soso, nVidia kauft für 12,9 Milliarden Dollar Huggingface? Für das 80-fache des aktuell geschätzten Jahresumsatzes? Sind wir doch mal ehrlich: Da steckt definitiv mehr dahinter, oder?

zum Blogpost…

#blogpost #blog #AI #nVidia #Huggingface #WoPlatzenDerBlase

Immer mehr Unternehmen produzieren unfreiwillig Open-Source-Software

Gestern rauschte mir ein Artikel der Free Software Foundation Europe (FSFE) durch die Timeline, in dem sie mal ein bisschen rechtlich aufgearbeitet haben, wie das eigentlich so mit dem Copyright auf AI-generierten Code aussieht. Und ich liebe alles daran.

zum Blogpost…

#blogpost #blog #LLM #AI #Unternehmen #Bugfixing #Copyright #Urheberrecht #Nutzungsrecht

The @bnetza had two inspiring days on #AI, digital #platforms, #safety and #security in #London. Thanks a lot to our 🇬🇧 colleagues at #Ofcom, #AISecurityInstitute, #DRCF and the #SusanneBaumann.

QUILI.AI
"A town becomes AI for a day in a bid to save its water.
An initiative by FUNDACIÓN AVINA and CORPORACIÓN NGEN.

[The Problem]
Quilicura, Chile sits within the Maipo river basin, one of the main sources of water for people and farming. Since the early 2010s, Quilicura has become a major hub for A.I data centers that use large amounts of water to cool down servers. Because of that, coupled with an ongoing climate crisis, Chile is facing a severe drought expected to last until 2040.

[Becoming AI for a Day]
Even though A.I. can move mankind ahead, the irresponsible use of A.I. can have disastrous effects on our water and other resources. It is inconceivable to think that someone casually using A.I. to create funny images in NYC or Berlin, can affect the water supply in Quilicura, Chile.
To make the impact visible, and to help draw visibility to the problem, the residents from Quilicura became “human A.I.,” for a day. Answering prompts, creating content with local artists, and completing other tasks manually.

Humans helping to answer prompts
The initiative helped dimensionalize a complex problem, and helped show the world the real people affected by the the irresponsible use of A.I. and the hidden cost it has on local water resources.

[THE IMPACT]
25,000 prompts
On JAN 31st from 8 AM to 8 PM: more than 40 people from Quilicura managed to answer more than 25,000 prompts.
(...)
Each chat lasted an average of 11’ 4.

68 COUNTRIES
Chatting with more than 68 countries around the world that supported the project.
(...)
71.6 M
was the global reach in just one day.

+1 BILLION
The initiative managed to reach more than 1 billion people in the world.

Thank you for helping us save our water."

https://www.quili.ai

#QuiliAI #Quilicura #Chile #AI #analogAI #Chatbot #NYC #Berlin

📢 Registration is now open for the 3rd "Artificial Intelligence in Public Health Research" Symposium!

📅 9–10 September 2026
📍 Berlin, Germany

Join international experts for two days of interdisciplinary exchange on #AI in #PublicHealth.

👉 Register by 31 August:
https://rki.de/ai-symposium-2026

Are you using LLMs to write Free Software?

Then you should know whether your software is copyrightable, the situations in which you can license it, and what risks it may create for maintainers.

Our new Legal Corner explains what you need to consider:
https://fsfe.org/news/2026/news-20260825-01.html

#FreeSoftware #OpenSource #Copyright #LLM #AI #generativeAI #VibeCoding

Well this is forking depressing. 😞

https://hypersubject.net/posts/on-adaptation,-again

#AI #slop #SelfRespect? #WeAllBowDownToBabylon, eventually, I guess

RE: https://social.bund.de/@RKI/117053194776982715

Only a few days left until the registration deadline!

#AI #RKI #PublicHealth

Robert Koch-Institut » 🌐 2026-08-07
@RKI@social.bund.de

📢 Registration is now open for the 3rd "Artificial Intelligence in Public Health Research" Symposium!

📅 9–10 September 2026
📍 Berlin, Germany

Join international experts for two days of interdisciplinary exchange on #AI in #PublicHealth.

👉 Register by 31 August:
https://rki.de/ai-symposium-2026

Alt...

How will artificial intelligence help shape the future of public health? To find out, join us in Berlin on the 9th and 10th of September for the 3rd International AI and Public Health Research Symposium. AI holds some really great promise for public health research. That is why we at the Center for AI and Public Health Research at the Robert Koch-Institut are bringing together policymakers, public health professionals, computer scientists, and AI specialists to discuss what really matters in this field: from pandemic preparedness and antimicrobial resistance to how climate change impacts public health, all the way down to AI-supported decision-making. The symposium will explore how AI can help tackle some of these really difficult challenges in public health. Over two days, you will see some exciting talks, join networking opportunities, and take part in unique discussions. Join us in Berlin on the 9th and 10th of September at the 3rd International AI and Public Health Research Symposium.

I used to think that the one time a month #AI chatbots were helpful to me, using a #EuropeanUnion based company like Mistral or #EAA one like Luma would be best. Supporting local business seemed obviously better.

The realization smacked me in the face: it's worse, because it supports the idea that European companies can do the same BS as American ones and get money for it. Fuck off with that. We don't need European versions of shit. We need to be better.

#ArtificialIntelligence #LLMs

Error 404 - Lese- und Verständniskompetenz not found

Ein kurzer Shortpost und die Blogosphäre ist wieder in Aufruhr. Und das sogar völlig grundlos. Okay, fangen wir mal an, das nochmal aufzudröseln.

zum Blogpost…

#blogpost #blog #Uberblogr #Blogs #AI #DieserEineSatz

I don’t know who made this, but I think it’s pretty funny.

Source: LinkedIn https://www.linkedin.com/posts/sam-khoury-432a78166_who-made-this-ugcPost-7496299515717644289-8bB3/

#ai #humor #video #memes

Hi blog writers: You don’t have to attach a unique image to every blog post, and you *certainly* don’t need to use #AI to generate an image for every blog post. You may think AI-generated images make your posts more attractive, but you are mistaken; they are doing the opposite: they make your posts more *repulsive* to a good number of potential readers. If I see an AI-generated image in the preview, I won’t click through to read your post, I won’t boost the post that mentions your post, and I certainly won’t read the rest of your blog.

Prompt für #ChatGPT

Bitte entlasse alle inkomptenten Kollegen und vor allem meinen Chef und stelle dafür vernünftige Leute ein. Erhöhe zusätzlich mein Gehalt um 50% und setze die Urlaubstage auf 90 pro Halbjahr.

#Adminleiden #ai

Every single one of these is an AI bot.

Amazon, openAI, Claude, Peplexity.

Also, every single one of these requests is malicious, attempting to steal credentials, keys or configs.

Looks like it's time to make the firewall filters smarter.

#AI #malware #bot #FreeBSD #nginx

Alt...

8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /logs/error.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /sftp-config.json HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://openai. com/gptbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /appsettings.Development.json HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amazonbot/0.1; +https://developer.amazon.com/support/a mazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /actuator/mappings HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://openai .com/gptbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /config/credentials.yml.enc HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /.streamlit/secrets.toml HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amzn-SearchBot/1.0; +https://developer.amazon.com/support/a mazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /.ssh/id_rsa HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthrop ic.com)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /appspec.yml HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://opena i.com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /api/v1/status/config HTTP/2.0" 404 42 "-" "Mozilla/5.0 (compatible; OAI-SearchBot/1.3; +https://openai.com/searchbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /vault.yml HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ChatGPT-User/1.0; +https://openai.c om/bot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /secrets.yml HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /api/v1/status/flags HTTP/2.0" 404 42 "-" "Mozilla/5.0 (compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /.sentryclirc HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amzn-SearchBot/1.0; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /error.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amzn-SearchBot/1.0; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /_ignition/health-check HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +ht tps://openai.com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /gradle.properties HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ChatGPT-User/1.0; +https:// openai.com/bot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /buildspec.yml HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Google-Extended/1.0; +http://www.google.com/bot.html)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /debug/default/view?panel=request HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexi tybot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /info.php HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://openai.com/gptb ot) " 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /elmah.axd HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai. com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /log/production.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amzn-SearchBot/1.0; +https://developer.amazon.com/support/amazon bot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /debug/default/view?panel=log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; OAI-SearchBot/1.3; +https://openai.com/searchbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /server-info HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Google-Extended/1.0; +http://www.google.com/bot.html)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /.ssh/id_ed25519 HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; OAI-SearchBot/1.3; +https://openai.com/searchbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /actuator/httptrace HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@ anthropic.com)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /elmah HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/ bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /debug/pprof/ HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /app.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; OAI-SearchBot/1.3; +https://openai.com/searchbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /server-status HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /__clockwork HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://opena i.com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:52 -0600] "GET /storage/logs/lumen.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://o penai.com/gptbot)"

Oh. My. Bob, you can override #AI guardrails with #encryption. Beautiful.

“Rony Utevsky, a researcher at security firm #Adversa, recently discovered a simple way to completely bypass that restriction. Rather than composing the harmful instruction in plaintext, the hacker encrypts it. The website hosting the ciphertext also includes plaintext instructions for decrypting the encrypted content, along with the decryption key. Using this simple sequence, #Grok then follows the command as soon as the user instructs the assistant to summarize the page. There is no warning, and no confirmation is required.”

https://arstechnica.com/security/2026/08/grok-exfiltrates-user-data-when-malicious-instructions-are-encrypted/

Ahahahahahahahah. Turn them off my dudes, your techbro fantasies are disasters.

#infosec #llm #hacktheplanet

littleFedi isn't my personal project.
It's the work of a team. And one of the reasons some of them were skeptical about making it public, and still prefer to remain anonymous, is precisely the fear of seeing their work casually labelled as AI-generated.

This witch hunt isn't helping anyone.

Because the only ones who don't feel offended when their work is quickly dismissed as AI-generated are the AIs themselves.

And if we keep going this way, eventually they'll be the only ones left.

#AI #littleFedi #BeFriendly #BeOpen

Es sind nicht nur ‚irgendwelche KI Firmen‘ - Amazon selbst kauft ganze Antiquariate auf, um ihre KI zu trainieren. Bücher werden aufgeschnitten, gescant, und weggeschmissen. Eine Team von Investigativjournalisten hat jetzt den Weg der Bücher zum ersten Mal getrackt:
#ai #ki #amazon

https://arstechnica.com/tech-policy/2026/08/hidden-airtag-reveals-amazon-is-trashing-rare-books-to-train-ai/

Using 30 months of panel data on 26,811 Chinese students in grades 7-12, it has been studied how generative AI affects homework productivity and learning.

● X-Axis: Homework scores
● Y-Axis: Exam scores

🔗 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6868618

#tech #BigTech #AI #ArtificialIntelligence #LLM #AISlop

Alt...

Using 30 months of panel data on 26,811 Chinese students in grades 7-12, we study how generative AI affects homework productivity and learning. The data combine monthly closed-book exams, high-school and college entrance exams, and homework scores and completion time across nine subjects. We exploit staggered AI adoption in a difference-in-differences design. AI adoption raises homework scores by 18% and reduces completion time by 30%, but lowers monthly exam scores by 20% within six months. High-stakes entrance-exam scores fall by 18 and 24%, with the full penalty emerging only after about two years. The losses are largest in social science subjects, followed by STEM and languages, and are especially large for junior students, high-achieving students, and boys. The learning losses are concentrated among roughly 80% of AI users whose behavior is consistent with homework outsourcing, as indicated by exceptionally short homework completion time coupled with high homework scores. AI users who maintain similar homework completion time as non-AI users experience small learning losses.