Conjecture Institute Fellow Carlos De la Guardia to Unveil Epistemological Approach to AGI in Future Book

Conjecture Institute Fellow Carlos De la Guardia is set to release a future book delving into artificial general intelligence (AGI), focusing on an epistemological framework for understanding and developing advanced AI. The announcement came via a tweet from the official Conjecture Institute account, stating, "Join Conjecture Institute Fellow @dela3499 talk about his (future) book on artificial general intelligence!" This initiative highlights a distinct philosophical perspective on AGI development.

De la Guardia, an AGI researcher with a background in robotics and longevity, centers his work on comparative epistemology, seeking to understand what makes human minds uniquely powerful. His research addresses five fundamental questions: how new things are created, which new things survive, the form ideas take, how resources are used, and the physical implementation of these processes. He draws insights from diverse fields, including language, logic, computation, causation, emotion, evolution, and epistemology.

The Conjecture Institute, a non-profit organization, is dedicated to advancing critical rationalism, a philosophy emphasizing that knowledge grows through conjecture and criticism. The institute posits that creating AGI necessitates "the right theory of knowledge," rather than solely relying on Bayesian calculations or increasing computational power. This aligns directly with De la Guardia's focus on the foundational mechanics of knowledge creation.

De la Guardia’s upcoming book aims to tackle profound mysteries surrounding human intelligence and its potential for enhancement. He explores how a deeper understanding of minds could lead to advancements such as faster thinking and even optional immortality through mental upgrades and backups. His work underscores an optimistic outlook on humanity's capacity for progress and the transformative potential of AGI.

The book is currently being developed, with sections shared weekly on his Substack, offering a detailed exploration of its themes and arguments. This structured release allows for ongoing engagement with his research, which seeks to break down complex AGI problems into more manageable components. The project is supported by the Conjecture Institute, reflecting their shared vision for knowledge-driven AI development.