From: jimruttshow8596
The Role of Feedback Loops in Complex Systems
Feedback loops are fundamental to the operation and emergence of higher levels of complexity in systems [00:40:08]. These loops are not merely modern concepts but have been recognized throughout history in various intellectual traditions, including in the writings of Aristotle [00:41:40]. They form the core principles of cybernetics and control theory [00:42:03].
Feedback Loops in Societal Systems
In complexity science, feedback loops help illustrate the distinction between focusing on individual components (“the dancers”) and the collective interactions and patterns that arise (“the dance”) [00:39:42]. For example, a business company, while abstract and not physically tangible, is “real in the sense that they have traction in the physical world” [00:45:07]. This reality emerges from coordinated actions operating on signals within boundaries, crucially supported by internal feedback loops [00:40:04].
A contemporary critique of society’s operating system highlights the dominance of the money on money return loop, which optimizes for financial gain and influences the behavior of individuals, thereby creating top-down causality [00:40:28].
Feedback Loops in Cognition and AI
The human mind itself is understood as having intricate feedback mechanisms. Early psychological models, like those proposed by Dietrich Dorner, incorporated “systems of competing springs in the mind that pull and push against each other and keep it in some dynamic balance” [00:41:51]. This cybernetic idea of a feedback loop is central to understanding autonomous motivation within cognitive architectures [01:40:43].
In the realm of Artificial Intelligence (AI), particularly with deep learning, the concept of feedback loops is crucial. While many early deep learning models were primarily feed-forward, the desire for recurrences has been present since the beginning [01:33:08]. Modern architectures, like the transformer models (e.g., GPT-2, GPT-3), implicitly incorporate feedback by building internal representations of an entire image or text sequence during prediction [01:36:33]. The introduction of “recurrent links” means that predictions from later layers inform and become context for earlier layers, leading to a “fully interconnected” system rather than a simple hierarchical one [01:32:08]. This interconnectedness, where information flows “everywhere backwards and forwards,” is believed to be how the human mind is structured [01:32:41].
The Dynamic Nature of Reality and Self-Regulation
Feedback loops also play a role in how individuals perceive and interact with reality. One’s “model that you make and the actions that you perform as a result of that change you regulate in a different way and as a result reality will now look different to you” [01:03:58]. This suggests a self-regulatory feedback loop where internal models influence external actions, which in turn shape perceived reality.
In terms of AI systems, the aspiration is to move beyond merely predicting reality from the past to also predicting it “from the perspective of the future that we want to have,” thereby limiting the search space to desired outcomes [01:37:11]. This reflects a more active, goal-driven form of feedback in intelligent systems.
The ability of a system to self-organize and adapt to the realities of its operating environment is also dependent on these feedback mechanisms [01:48:06]. For instance, in distributed computing, a system needs to understand network costs and organize itself to maximize efficiency, a dynamic process enabled by feedback.