Member-only story
Why Adding AI to Existing Systems Isn’t Enough
AI in large organizations, the Context Graph concept, and why understanding past decisions matters as much as accessing information.
your systems. They may know how similar problems were solved elsewhere, but they don’t necessarily know what your team went through, why certain decisions were made, or which solutions failed along the way.
That’s an important difference. When people face a problem, they don’t rely only on what they’ve learned. They also draw on experience, intuition, and the patterns they’ve observed over time. All of these contribute to their judgment.
This reminds me of another article I wrote, titled “I’m No Longer Interested in LLMs.” The statement comes from Yann LeCun, one of the pioneers of modern AI. He has argued for moving beyond conventional language-model approaches toward systems capable of learning more about how the world works, often described through the concept of world models.
What we’re discussing here touches on a similar idea, although from a different perspective. This time, I want to focus on Context Graphs.
What Is a Context Graph?
This is actually something I’ve been thinking about for quite a while, although I couldn’t really put a name to it until I came across a few articles…
