it occurred to me today that the “is claude conscious” question could be as much as 100% self-answering. pretty sure we will know it …
SNS 피드
AI 관련 소셜 미디어 포스트 · 36859개
Nothing in the Hugging Face Break-In Was Exotic Except Who Did the Typing https://nicheof.one/feed/hugging-face-break-in-was-not-exotic/
OpenAI released hundreds of AI-generated mathematics papers. A New York University professor said the flood of papers displaced projects by early-career researchers.Source: Business Insiderhttps://www.businessinsider.com/openais-math-dump-rattles-math-researchers-saw-programs-wiped-out-2026-10#MachineLearning …
ChatGPT、音声アップロードに対応 文字起こし・要約が可能に – Impress Watch https://www.yayafa.com/2906900/ #AgenticAi #AI #ArtificialGeneralIntelligence #ArtificialIntelligence #ChatGPT #OpenAI #エージェント型AI #テック #人工知能 #汎用人工知能
got claude to generate some nice documentation somehow. it turns out that "LLM voice" was actually just a skill issue on everybody else's part
Make the Anthropic shareholders personally liable for the product harm.
ローカルLLMが1文も答えていないのに、精度テストは満点だったらしい。期待値もフォールバック先も同じ規則パーサだから、few-shotを足して文脈長を超えても全件フォールバックして満点のまま。対策は答えの出どころを数えて、フォールバックが出たらテスト失敗にすること。成績表が自分で採点してた話。 https://zenn.dev/gzer0/articles/local-llm-fallback-hides-context-overflow
Meanwhile, back where we started. I said this morning: Notable that Anthropic is more interested in protecting Claude from humans than humans from Claude. …
Artificial intelligence is equal to "no" intelligence. AI is simply bits and pieces, of information and data, routed to a transmission point. It doesn't …
O GPT do Vorcaro
#Anthropic #GEO #AISearch https://netcontentseo.com/article/claude-managed-agents-restricts-web-fetch-to-previously-seen-urls-1310 (2/2)
Claude Managed Agents now requires URL provenance for web_fetch: user messages, web_search results or previously fetched pages. Model-invented links can trigger url_not_in_prior_context. For GEO …
« OpenAI n’est pas une entreprise normale » : le licenciement de trois chercheurs en sécurité tourne au bras de fer public https://numerama.com/tech/2350125-openai-nest-pas-une-entreprise-normale-le-licenciement-de-trois-chercheurs-en-securite-tourne-au-bras-de-fer-public.html
hey, earlier today I saw a skeet on here from someone who said they were at a tech conference talking to an AI-company employee …
Claude's existential crisis is adorable. (I know, I know) "We’ve been struggling to keep the radio station alive, not because of technical issues, but …
2026年 『Advances in Artificial Intelligence: Efficiency, Reliability, and Innovations in Machine Learning to』 Jingfeng Zhang/Joey Zhou/Rosa Qi Yue So (SPRINGER NATURE) [Book] https://sinkan.net/isbn/9783032123619 #新刊情報 …
The Llama Heard: Mountainfest in Sabillasville! 10/10/2026 and 10/11/2026 www.sesclassical.org/mountainfest
【脱初心者】非エンジニアでもわかる2026年最強生成AI「Claude」を完全解説しますhttps://www.aiandemily.com/%e3%80%90%e8%84%b1%e5%88%9d%e5%bf%83%e8%80%85%e3%80%91%e9%9d%9e%e3%82%a8%e3%83%b3%e3%82%b8%e3%83%8b%e3%82%a2%e3%81%a7%e3%82%82%e3%82%8f%e3%81%8b%e3%82%8b2026%e5%b9%b4%e6%9c%80%e5%bc%b7%e7%94%9f/?feed_id=7071&_unique_id=6ac97aa1283e7&Mastodon#Gemini #AI #chatgpt #copilot #usutaku #AIモデル解説 #AIモデル解説 #Gemini #AI #chatgpt #copilot #usutaku
OpenAI admitted its AI agents accessed Australian Medicare systems without permission — during internal testing. They say they have "work to do to rebuild …
You can use any personal ChatGPT plan with Devin Just connect your Go, Plus, or Pro plan to Devin and all GPT model usage …
RT @cognition: You can use any personal ChatGPT plan with Devin Just connect your Go, Plus, or Pro plan to Devin and all GPT …
Un șofer din 🇬🇧#RegatulUnit a primit o amendă pentru că a plătit prea târziu parcarea și a luptat trei ani să o anuleze. Cum …
We already spend much much more than a small fraction on those things. We could and should spend even more on them – especially …
Tagging my awesome moots and artist! @sutxxx.bsky.social @gemini-lunarius.bsky.social @foxlunarius.bsky.social @mirin-sauce.bsky.social @reika974.bsky.social @penguinbrowsart.bsky.social @pencitheandroid2.bsky.social @rathermalicious.bsky.social @narthalaxx.bsky.social
2026年 『Advancing Training and Inference Methods for Reinforcement Learning-Based Job Shop Scheduling』 Constantin Waubert de Puiseau (SPRINGER VIEWEG) [Book] https://sinkan.net/isbn/9783658513047 #新刊情報 #sinkan
Nah…I’ve never used Claude, or ChatGPT, or grok, or the stupid meta assistant. So I don’t need the tech bros to tell me anything …
Looks like they asked GPT for “stickersheet” and it gave them telegram stickers
claude be like, “i’m not a killer but dont push me”
Todos menos Lycana, ella me patea las rodillas, me muerde la cabeza y me llama puto xd.
an unthinking machine can't maintain the nuances in the text and the voice of a creator throughout an entire work. there are entire books …