· TMK Press team
Mario Garcés featured by Fundación Empresa y Sociedad in the debate on AI scaling and alternative architectures
Mario Garcés on the limits of AI scaling and why new cognitive architectures need explainability, traceability and causal learning by design.
The Mindkind founder contributes to the discussion on the economic and technological limits of current AI scaling and the need to investigate alternative cognitive architectures.
Fundación Empresa y Sociedad has featured Mario Garcés, founder and CEO of THE MINDKIND, in its latest article examining one of the central debates surrounding artificial intelligence: whether AI development should continue accelerating, slow down, or follow a more selective approach based on measurable technological and business outcomes.
The article, titled “La innovación de la innovación #5 | ¿Acelerar o frenar la IA?”, addresses the growing tension between the rapid increase in AI capabilities, the investments required to sustain that growth, and unresolved questions surrounding safety, control and economic returns. In its section dedicated to Mario Garcés, the article highlights two additional dimensions of this debate: the economic and technological limits of AI scaling.
It specifically raises the possibility of:
“investigating other architectures if current systems encounter limits that cannot be solved simply by adding money, data and computing capacity.”
This question is closely aligned with THE MINDKIND’s research strategy. While the dominant AI paradigm has achieved remarkable results through increasingly large models, datasets and computing infrastructures, THE MINDKIND is investigating a different technological path through its ETR Cognitive Architecture.
ETR approaches intelligence from a complex-systems and cognitive perspective. Rather than relying exclusively on statistical scaling, it explores how complex behaviour can emerge from the dynamic interaction of simpler cognitive processes (CELLs) operating over a shared knowledge structure.
This architecture is being designed around properties that THE MINDKIND considers particularly relevant for the development of advanced AI: efficiency, causal relationships, traceability and explainability by design.
The company has not yet achieved Artificial General Intelligence (AGI). Its long-term research objective is to investigate whether alternative cognitive architectures such as ETR can contribute to overcoming limitations that may not be solvable through scale alone.
As AI investment continues to grow, THE MINDKIND believes that the question is no longer only how far existing architectures can be scaled, but also what new architectures should be explored.
More compute can increase capability. The next fundamental breakthrough will require a different architecture.
Related reading
More on the efficient, causal, traceable and explainable-by-design architecture mentioned in this article:
TMK Press team
press@themindkind.com
