Timeline of Early AGI Projects
A compact timeline connecting classic AI, Novamente, OpenCog, LLMs, and reliable agent workflows.
Archive
What early AGI architecture work can teach builders working with LLMs, RAG, and agentic systems today.
Before LLMs became the dominant interface, AGI researchers worked on architectures that combined symbolic structures, probabilistic reasoning, memory, and action selection.
The modern equivalent is a workflow stack: model, retrieval, tool use, state, evaluation, and human review. The old work is useful when it helps us ask better reliability questions.