For the last few years, RAG has been one of the most important concepts in AI. But now there is a...For the last few years, RAG has been one of the most important concepts in AI. But now there is a...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Abu's avatar
Abu Bakar
• Sep 24
For the last few years, RAG has been one of the most important concepts in AI.
But now there is another interesting concept developers should know about:
OKF - Open Knowledge Format.
And no, RAG is not dead.
OKF and RAG actually solve two different problems.
Think about it in a very simple way:
RAG is like searching inside a huge library.
You ask a question.
RAG searches through thousands of documents, finds the most relevant pieces, and gives them to the AI so it can answer with better context.
That works really well.
But there is still one problem:
What if the library itself is messy?
What if there are:
→ duplicate documents → outdated information → conflicting answers → unclear definitions → no clear source of truth
This is where OKF becomes interesting.
OKF focuses on organizing knowledge before an AI agent uses it.
Instead of giving an agent random documents, OKF lets knowledge be stored in a structured, human-readable and agent-friendly format.
In simple words:
RAG asks: “Which information should I retrieve?”
OKF asks: “How should this knowledge be organized, trusted and understood?”
Google Cloud introduced OKF as an open format based on simple Markdown files with structured metadata.
The newer OKF v0.2 also adds ideas around:
→ where the information came from → whether it was verified → how fresh the information is → which version is current → how much the agent should trust it
And this is where things become really interesting.
The future may not be:
RAG vs OKF
It may be:
OKF + RAG
OKF can organize and maintain trusted knowledge.
RAG can search through that knowledge when the amount of information becomes too large to send directly to the model.
So instead of thinking:
“RAG is finished.”
I think the better way to look at it is:
RAG was one important layer. OKF is adding another layer to how AI agents understand knowledge.
And as AI moves from simple chatbots toward real autonomous agents, managing knowledge properly may become just as important as choosing the right LLM.
RAG helped AI find information.
OKF is trying to help AI understand which knowledge it should trust and how that knowledge is organized.
We are slowly moving from:
AI that searches documents
to
AI systems that work with structured, trusted and continuously maintained knowledge.
And that is a very interesting direction for Agentic AI.
#ArtificialIntelligence #AgenticAI #RAG #OKF #GenerativeAI #AIEngineering #LLM #SoftwareEngineering #OpenKnowledgeFormat
agenticaiartificialintelligencerag
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started