@Tutanota @nextcloud @notesnook @CryptPad @MistralAI

AI slop hurts your soul.

It makes you: distrustful, dismissive of creative works, uncreative, derivative, imaginatively stunted, unreflective, dependant, it robs you of the satisfaction and growth granted by the creative process and it poisons you with all the same cultural biases taken from the training data.

AI robs people of imagination and the drive to create and instead produces output that is pastiche on steroid, uninspired unchallenging slop, every little detail crafted without personality or opinion with nothing to say except "this might vaguely remind you of something good that already existed"

I am also particularly concerned about the rise of chatbots pretending to be humans for political purposes.

https://alextheward.com/blog/llms_are_a_cognitohazard/

#LLM #AISlop #GenAI #DeadInternet #Art

As someone whose job description has expanded to include: "Prepare offers for specialized IT systems including server hardware", I have this to say:

The #LLM bubble cannot burst soon enough.

I may regret this at some point, but I felt the need to put down in writing how I feel about this moment in the tech industry.

It is not kind. You may well be insulted by it. If you are... then you really should question yourself.

https://www.garfieldtech.com/blog/selfish-ai

#AI #LLM #Programming

@kagihq hey Kagi, where are your plans and options without GenAI/LLM? Why does having a search service require and force GenAI?

Provide a Starter plan without GenAI for $3/mo and I'll sign up today. I refuse to use, fund, or otherwise support a service that forces GenAI.

Also, as others have mentioned, where are the APKs for installation by folks that don't use Google Play?

#GenAI #LLM #rAIpe #Kagi #KagiHQ

#WritersCoffeeClub 4/2: How much detail do you use to describe your settings?

The setting in question is mostly 19th century Germany and surrounding areas - which is as alien to a modern English-language audience as many fantasy worlds.

Thus, I have an appendix giving some cultural context, and the tales I translate have numerous additional commentary and footnotes. This is actually a large portion of my work, in addition to the translations themselves. This is also part of the reason why I refuse to use #LLM for translation (beyond the usual ethical objections) - I want the readers to understand the context of the tales, and AI would inevitably get that wrong.

Not only are workloads increasing, by hollowing out their junior ranks, businesses are putting themselves squarely in the path of a slow-burning demographic time bomb as seniors begin to retire in record numbers.

"There won’t be an endless supply of experienced hires to fall back on, so everyone will be fighting for the limited, increasingly expensive talent with domain expertise"

What exactly is the tech-industry's endgame here?

https://www.fastcompany.com/91483431/companies-replaced-entry-level-workers-with-ai

#noai #techbros #ai #llm #technology #programming #sysadmin #linux #indieweb #capitalism #reading #art #adhd

"AI"-users are spamming so many shitty PRs that GitHub is considering allowing repository owners to outright disable the pull request system on their repos to prevent low-quality contributions from reaching them.

https://github.com/orgs/community/discussions/185387

#AI #LLM #Copilot #Claude #ChatGPT #GitHub #Microsoft #Microslop

💬 The talk examined how persuasion increasingly operates indirectly: speakers influence how LLMs interpret language, while model-generated summaries, commentary, or analyses in turn shape human decisions, for example in financial communication, political contexts, and social media.

⚖️ A key focus was on emerging risks in #LLM mediated communication, including shifts in power and responsibility, as well as phenomena such as neutrality collapse and stance amplification in agent-based systems.

💬 The talk examined how persuasion increasingly operates indirectly: speakers influence how LLMs interpret language, while model-generated summaries, commentary, or analyses in turn shape human decisions, for example in financial communication, political contexts, and social media.

⚖️ A key focus was on emerging risks in #LLM mediated communication, including shifts in power and responsibility, as well as phenomena such as neutrality collapse and stance amplification in agent-based systems.

"While LLMs offer immediate convenience, our findings highlight potential cognitive costs. Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels. These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI's role in learning."

https://www.media.mit.edu/publications/your-brain-on-chatgpt/

#MOLE #AI #LLM #cognition

"While LLMs offer immediate convenience, our findings highlight potential cognitive costs. Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels. These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI's role in learning."

https://www.media.mit.edu/publications/your-brain-on-chatgpt/

#MOLE #AI #LLM #cognition

I used to actually enjoy llm/gen-ai back hen it was still really bad.. like getting these super creepy images or ridiculously bad poems and what have you?

Anyone else?

#ai #genai #llm

pairwiseLLM v1.1.0 offers a unified framework for generating and analyzing pairwise comparisons of writing quality with LLMs. It supports live and batch workflows across multiple providers and models results using Bradley–Terry or Elo methods to derive quality scores. The package includes introductory and advanced vignettes.

https://cran.r-project.org/package=pairwiseLLM

#RStats #LLM #TextEvaluation

I approve of @nlnet holding their ground on requiring the logging of LLM chats:

https://nlnet.nl/foundation/policies/generativeAI/archive/feedback/

"We are open to feedback and we also have to learn ourselves. This is a fast-changing field with many moving parts."

I have feedback!

1) Thanks for having a spine.

2) OpenCode has a convenient feature to export chat logs (in a nice, git-friendly Markdown format). Awesome I say! See Screenshot; the feature is called "Export session transcript" (shortcut: ctrl+x x). Avail yourself of it, my dudes!

#OpenSource #OpenCode #LLM #prompt

"AI agents—specifically tools like Claude Code—are inherently vulnerable to a "nightmare" security flaw: Indirect Prompt Injection"

#AI #CyberSecurity #PromptInjection #AIAgents #LLM #Programming #InfoSec #TechSecurity #ClaudeCode

https://youtu.be/_3okhTwa7w4

pairwiseLLM v1.1.0 offers a unified framework for generating and analyzing pairwise comparisons of writing quality with LLMs. It supports live and batch workflows across multiple providers and models results using Bradley–Terry or Elo methods to derive quality scores. The package includes introductory and advanced vignettes.

https://cran.r-project.org/package=pairwiseLLM

#RStats #LLM #TextEvaluation