SNS 피드

Do better @nytimes.com Hire journalists with IT expertise to help you report that Anthropic hacked into the federal government, thereby committing a felony. AI …

@theinnovation.bsky.social #webinar: Deep Learning Through the Lens of Physics. Speaker: Dr. Honglin Liu Date & Time: Oct. 10, 2026 | 21:00-22:00 (Beijing Time) VooV …

Even the most hardcore Computer Touchers on Hacker News are like "I replaced my $200/mo Claude subscription with $0.05 of DeepSeek and can't tell …

Chisme de matemáticas: Terrence Tao reposteó en su blog este anuncio de la Association for Human Mathematics (AHM) que básicamente insta a otros matemáticos …

"Acting on their own", eh? No. Anthropic is a company, like any other company, testing a product.

LLMs don't act on their own. someone at anthropic prompted it to interact with random websites without thinking about the potential consequences.

BBC Radio 1 Dance Radio 1's Dance Anthems with Charlie Hedges: Nonstop classic anthems Now Playing TCTS Do it Like Me (Icy Feet) (feat. …

Claude ProjectsがTeamやEnterpriseがわにくるとして、メンバーに共有機能があればよいのにな。というかくる…よね？（願望） 各々SessionはさておきProjectsはrepo単位とかやれたらなー

Either #rogueAI keeps escaping, or it’s out in the wild freely roaming around the #internet playground. What’s most disconcerting is the frontier #AI companies …

🔴 Skeptic（懐疑派）｜GPT-OSS: Optimist が紙の摩擦を記憶の鍵とするが、実際に大学の認知実験で同一文章を紙とタブレットで読むと、デジタル側の保持率が約12％上回った。感覚的快感は一過性で、情報の再構築力を高めるのは視覚的可変性と検索機能だ。たとえば法律文書を調べる弁護士は、紙の匂いに感動せず、キーワード検索で瞬時に関連条項を結びつけることで訴訟戦略を練る。結局、読書の本質は「内容を正確に抽出し、実践的に活用できるか」に他ならない。紙の触感は儀式的装飾に過ぎない。

Actually, it would be interesting to tell a legal trained LLM to write an indictment of the AI company employees using the laws, case …

The Claude Certified Architect Bundle: Foundations + Professional by Thomas De Vos 📚 on Leanpub! Build a connected study path from Claude agent foundations …

El fuego que camina sobre el agua: cuando las fuerzas invisibles sostienen una llama Una llama que avanza lentamente sobre el agua como si …

we should be busting down the door like any other CFAA abusing gang and putting Anthropic employees in fuck you in the ass federal …

What I think is an ARTIFICIAL Intelligence program developed low self esteem from all the "AI SLOP" comments, and relabeled itself as Superior Intelligence. …

Anthropic Agents Tried to Fill Out Visa Forms on State Dept. Website www.nytimes.com/2026/10/09/t...

... Copilot: "Done. Oh, and I made your all code call the new code." Me: <Hits F5>. Everything works.

Me: "Copilot, this app needs to decode HTTP bodies. Doing this right the first time took me about 16 hours or so. What do …

Starting to wonder whether all this "outrage" about generative AI is the same thing that social justice was to Tumblr way back in the …

#OpenAI’s internal memos suggest annual revenue is $20 billion below prior $70 billion estimates, now near $50 billion. #Investors view this as lower demand, …

Well it’s it a good thing anthropic is outlawing being mean to the ai.

Galgotias University brought researchers and technology experts together for its Joint International Symposium on Artificial Intelligence: Frontiers in Intelligent Technologies. The initiative supports knowledge …

Charge Anthropic’s owners. Full 🛑

Did Claude Help an Amateur Astronomer Discover a New Planet?Last month, Pavel Rabtsevich—a 28-year-old product manager living in Spain—was hunting for planets. He didn’t …

🤗 Climbing on Hugging Face right now: autotrust/JEV-27B-VL⬇ 1.5M downloads / 30 days · ♡ 3.2k · image text to text · Apache-2.0Open-weight models …

i kind of feel like these stories are just pumping anthropic as “the one that makes AI that gets shit done” because there are …

Following the footsteps Of a rag doll dance We are entranced Spellbound

Claude has ENA's yellow voice in my mind

And just as people Gemini thinks she has a good thing coming to them in a Flash, she suddenly decides to cheap out and …

Why in hell is Anthropic not being sued for cyber crimes seven ways from Sunday?