Part 2/11:
The researcher recounts his early ideas, dating back to around 2017 or 2018, when he devised Maraghi, a microservices architecture aimed at integrating robotics and artificial general intelligence (AGI). Frustrated by the complexity and limited technological capabilities at the time—specifically the availability of only GPT-2—he moved away from this vision.
With the advent of GPT-3 and its fine-tuning abilities, he simplified his approach to what he now calls Natural Language Cognitive Architecture. Demonstrating forward-thinking, he authored a free online book outlining this structure, which features a shared database linking dual processing loops—a core component of his original design.