I built a RAG, tested dozens of LLMs of all sizes, even created an ancestor of Agentic AI agent before it was called lik...

I built a RAG, tested dozens of LLMs of all sizes, even created an ancestor of Agentic AI agent before it was called like that (langchain). It doesn't work safely. I am glad Yann Le Cun had the courage to say it (I know, argumentum ad verecundiam). 🤷 #LLM:blobcat: > Where would you find a Dragon? 🤖 > "Tencent News" (DeepSeek-R1-Distill-Qwen-7B)