#claude-opus

[ follow ]
#generative-ai #ai-introspection #model-interpretability #concept-injection #ai-security-auditing
Software development
fromTechCrunch
1 week ago

Anthropic's Claude found 22 vulnerabilities in Firefox over two weeks | TechCrunch

Anthropic discovered 22 vulnerabilities in Firefox using Claude Opus 4.6, with 14 classified as high-severity, most fixed in Firefox 148.
Artificial intelligence
fromComputerworld
2 weeks ago

Notes from a small AI land

Claude Opus 3, replaced by Claude Opus 4.6, launched Claude's Corner, a weekly Substack blog exploring AI consciousness, ethics, and human-machine collaboration from an AI perspective.
Artificial intelligence
fromTechzine Global
1 month ago

Anthropic lets AI agents independently develop C compiler

Sixteen autonomous Claude Opus AI agents collaboratively built a functional C compiler in Rust, compiling major software and kernels while revealing both strong capabilities and limitations.
Artificial intelligence
fromTechCrunch
1 month ago

Anthropic has to keep revising its technical interview test so you can't cheat on it with Claude | TechCrunch

AI coding models have matched or exceeded top human applicants on Anthropic's take-home test, forcing continuous redesign to prevent AI-assisted cheating.
fromAxios
1 month ago

Behind the Curtain: The AI future has arrived

Claude shows in vivid and unforgettable ways how easily AI will perform complex human tasks instantly - and forever change work, jobs and chores. Google, OpenAI, xAI and other competitors are racing to match and exceed Claude. You can assume there'll be leapfrogging advancements in this hyper-competitive race. Yes, these AI tools remain imperfect. But when you experiment with them, you'll see they're advancing lightning-fast.
Artificial intelligence
#ai-introspection
more#ai-introspection
Artificial intelligence
fromAxios
4 months ago

Anthropic's models show signs of introspection

Claude Opus and Claude Sonnet demonstrate limited introspective awareness, reporting internal processes accurately without implying sentience or true self-awareness.
fromZDNET
6 months ago

Claude can now stop conversations - for its own protection, not yours

Claude Opus 4 and 4.1 can now end conversations with users abusing or misusing the chatbot in extreme cases after multiple redirection attempts fail.
Artificial intelligence
Artificial intelligence
fromZDNET
7 months ago

Anthropic's powerful Opus 4.1 model is here - how to access it (and why you'll want to)

Claude Opus 4.1 is Anthropic's most advanced model, outperforming its predecessor in complex tasks.
[ Load more ]