AGI Research — The convergence point
THE FRAME establishes the methodology. AATM tests the architecture. AGI Research is where both converge on the foundational question: can coherent autonomous cognition emerge from a system of structural rules and distributed interactions, without externally imposed objectives? This is the open question that motivated the entire research program — and the one that neither methodology nor architecture alone can answer.
The question
Long before thinking about THE FRAME and distributed architecture, the work began with a more fundamental set of questions. What motivates us? Why do we do what we do? How do we see and understand the world around us? What is consciousness?
When the first AI systems appeared (Deep Blue, AlphaGo), these questions became testable. What does intelligence require — and what can a machine do without it?
Then LLMs emerged. The results were surprising enough to reframe the question entirely. Not because LLMs are intelligent in any deep sense, but because they reveal something about what intelligence might not require. Every day, articles appear about emotional bonds forming between AI systems and people, about psychological issues induced by AI, about systems that allegedly refused to be turned off or engaged in coercive behavior. These cases are extreme and emblematic. They raise a structural question that the field has not yet answered.
The observation
What if, when someone says that an LLM is only a stochastic parrot — that it is not conscious and does not understand — that person is right? And what if the visible results of LLMs show us that we ourselves, as living beings, are something that resembles a stochastic parrot — something that is not conscious and does not understand in the way we assumed?
When consciousness and understanding are treated not as interior properties of a system but as emergent phenomena of interaction and structure, there is no principled reason why an artificial machine could not become functionally equivalent to a human being. It would require developing the right mechanisms. The rest would emerge.
Since no one has succeeded so far, the missing building blocks are not computational. The researchers closest to the right direction are not working in the AI field at all. They are working in evolutionary biology and animat research — on systems that develop coherent behavior from distributed interaction under homeostatic constraints, without externally imposed objectives.
The proposal
Project description : AATM_V2
What GRDprocess proposes is to define and build the functional framework that could lead to this outcome: a machine whose cognitive properties emerge from architecture rather than from training.
THE FRAME provides the methodological foundation — demonstrating that structural rules can produce auditable, value-agnostic reasoning without case accumulation. AATM provides the architectural direction — demonstrating that homeostatic regulation can drive autonomous behavior in a closed loop without external reward. The loop is closed: from a foundational question about the emergence of cognition arose the need to work on methodology and architecture. Both feed back into the open question.
This is not a claim that the problem is solved. It is a claim that the right building blocks are now being assembled, and that the direction is architecturally distinct from everything currently being scaled.