FEATURED · COMPANY MEMORY
Company memory
Karpathy and Obsidian: build a memory your business can put to work
From the LLM Wiki idea to a governed knowledge base: retain evidence, validate lessons and reuse them across business processes.
Insights & guides
Understand.
Decide. Move forward.
What to prepare, build and measure to put your knowledge to work with AI.
Artificial intelligence
Enterprise RAG assistants: preparing the move from prototype to production
A convincing demo still needs access control, reliable sources and error handling. These decisions should come before deployment.
Data & AI strategy
A terabyte of documents is not a context window for your AI
Select useful documents, preserve their context and measure costs: the preparation work that makes an AI project usable.
Methods & agents
Skills and hooks: teach a method, control its execution
A business procedure and a programmed check serve different purposes. Design both to make an agent useful in everyday work.
Decisions & workflows
Before AI takes action, define how it should route a request
Classifiers, thresholds and human escalation: separate request routing from the business decision that follows.
Knowledge governance
A useful knowledge base must know what it can show and what it must forget
Access permissions, superseded documents and derived summaries: practical decisions for maintaining an AI-ready enterprise knowledge base.