-
RAG with ClickHouse: Vector Search, Analytics, and LLMOps
Three Birds, One Engine: How ClickHouse Powers Retrieval-Augmented Generation End-to-End
-
Observability in RAG: LangChain, LangSmith & SigNoz
How SigNoz compares to LangChain and LangSmith in a Retrieval-Augmented Generation pipeline: what it can replace, where it fits, and what gaps remain.
-
ClickHouse and EU AI Act Technical Documentation
Assessing how ClickHouse supports each Annex IV documentation aspect under the EU AI Act and identifying the missing gaps.
-
Paper review: Not quite the AlphaGo moment yet
ASI-ARCH presents a fully autonomous LLM-driven pipeline that reports discovering 106 "state-of-the-art" linear-attention architectures and frames this as an "AlphaGo-like" leap.
-
Shocking Hidden Reasoning: OpenAI Flagship Reasoning Model o3-pro
Instead of generating the requested dataclasses, the model plans to return a soup of placeholders and dubious parameters - before refusing to reply at all.