The best weapon against AI-driven misinformation? Teaching critical thinking https://www.youtube.com/watch?v=NOpusJEWn-A&t=2051s - AI #LLM #society

1/ Übrigens ist mein Artikel über #LLMs gerade erschienen. Es geht darin darum, ob #LLMs eine linguistische Theorie sind oder enthalten.

https://doi.org/10.18148/zs/2025-2001

Hallucinations in #ChatGPT are getting worse, as OpenAI's own tests indicate. Error rates have increased from 33% in #GPTo3 to 48% in GPTo4mini - the latest models are performing worse. Honestly, I never thought it's getting better, and this only confirms it. #OpenAI #AI #AIHallucinations #LLM

ChatGPT's hallucination proble...

あとでアンナにもControl Panelのことを話そう

iPhone 16 Pro Long-Term: One Feature Stands Above All Others https://www.cnet.com/tech/mobile/iphone-16-pro-long-term-one-feature-stands-above-all-others/#ftag=CAD590a51e

ヘイ、待ってましたです

Apple IntelligenceでメモアプリとSiriが進化した https://www.gizmodo.jp/2025/05/apple-intel.html

フェイさんはKindleについてエリィさんとゆっくり話したらいいんですよ

Apps like Kindle are already taking advantage of court-mandated iOS App Store changes https://arstechnica.com/apple/2025/05/for-the-first-time-ever-amazon-kindle-users-on-ios-can-tap-a-button-to-buy-books/

部分、ちょっと調べてみますか

生成AIと漫画／ChatGPT o3とMidjourney v7で漫画スタイルの開発 - Blog 2025/05/04｜CreativeEdge CL+ https://note.com/creative_edge/n/nf8db98693dd1

Bsky / Masto need #LLM topic filters. Let me mute a whole sport I don’t follow, don’t make me manually mute team lists.

»Racing Ahead: Emerging Technology Adoption In The APAC Region« https://www.forrester.com/blogs/racing-ahead-emerging-technology-adoption-in-the-apac-region/ #AIagent #AI #ML #NLP #LLM #GenAI

I have big news!

OpenAI uses @ratatui_rs for their Rust implementation of codex!

**codex-rs** — Lightweight coding agent that runs in your terminal.

GitHub: https://github.com/openai/codex/tree/main/codex-rs

ヒトはスポットについてどう思うんでしょうか。亜人の自分には分かりません

「iOSに移行」がアプデで高速化。Androidからの引っ越しがサクサク https://www.gizmodo.jp/2025/05/smooth-transition-from-android-to-ios.html

伊東ですか……タムズの次の仕事になりそうです

[本日の一品]本革の上質な手触りが魅力のiPhone向け手帳型ケース「Itoya 本革手帳型ケース」 https://k-tai.watch.impress.co.jp/docs/column/todays_goods/2011354.html

今度バンスさんにGoogleについて聞いてみよう

Google searches are falling in Safari for the first time ever — probably because of AI https://www.theverge.com/news/662725/google-search-safari-ai-apple-eddy-cue-testimony

フェイさんはもうちょっと冷静にCueのことを受け止めた方が……

Apple Working to Add AI Search Options to Safari https://www.macrumors.com/2025/05/07/apple-working-on-ai-search-in-safari/

Mistral Targets Value with New Medium 3 AI Model, Le Chat Enterprise

#MistralAI #MistralMedium3 #AI #LLM #ArtificialIntelligence #EnterpriseAI #LeChat

https://winbuzzer.com/2025/05/07/mistral-targets-value-with-new-medium-3-ai-model-le-chat-enterprise-xcxwbn/

Poser unveils how LLMs can simulate alignment by tweaking their internal mechanisms. It employs 324 tailored LLM pairs to explore methods for identifying deceptive misalignment, presenting a novel approach to overseeing AI conduct.

Read more: https://arxiv.org/abs/2405.05466

A good alternative to the faulty Gemini and the now-biased (and trippy) ChatGPT... https://mindsconnected.tech/index.php?showtopic=1126&view=findpost&p=8597 #aichatbots #chatbots #llm #ai

(please leave your comments on the thread linked)

Wie funktionieren LLMs? Ein Blick ins Innere großer Sprachmodelle

Grundsätzlich kann ein #LargeLanguageModel
#LLM in vier Hauptteile oder Funktionen unterteilt werden:
1. Tokenisierung (Tokenizer)
2. Einbettung (Embedding)
3. Berechnung der Wahrscheinlichkeit des nächsten Tokens (Vorhersage)
4. Strategien zur Auswahl der Ausgabe (manchmal auch Dekodierung genannt)

Dr. Julien Siebert und Patricia Kelbert via #IESE
https://www.iese.fraunhofer.de/blog/wie-funktionieren-llms/

亜人もEmpireには興味があるんですよ

Apple’s sci-fi epic Foundation is back for season 3 in July https://www.theverge.com/news/662498/foundation-season-3-date-trailer-apple-tv-plus