‘The Brain, In Theory,’ an Excerpt

Why It Matters

Understanding the brain’s limits as a programmable device refines computational neuroscience and tempers AI analogies, influencing both research directions and philosophical debates about mind and machine.

Key Takeaways

• Evolution shapes brain indirectly, not via direct weight programming.
• Synaptic plasticity follows biological rules, not arbitrary user instructions.
• Neural network universality relies on modifiable parameters absent in biology.
• Brain functions as information‑processing system, not literal computer.
• Computationalism emphasizes computation, not programmability.

Pulse Analysis

The notion that the brain is a programmable entity persists partly because of the seductive analogy between neural circuits and digital computers. However, evolution does not edit synaptic weights like a software engineer; it operates through genetic instructions that guide protein production and developmental processes, producing structural changes over generations. This indirect influence means there is no one‑to‑one mapping from genome to specific synaptic configurations, debunking the idea that natural selection "writes" programs into the brain.

Artificial neural networks achieve universal function approximation by allowing a modeler to set weight parameters arbitrarily. In biological tissue, however, synaptic strengths are both dynamic variables and the result of constrained plasticity mechanisms such as Hebbian learning, spike‑timing‑dependent plasticity, and neuromodulation. The artificial distinction between fixed parameters and evolving states does not translate to living neural tissue, where any change is governed by biochemical rules rather than external instruction. Recognizing this gap prevents over‑extension of AI concepts to neurobiology and clarifies why brain‑inspired hardware must respect the limits of biologically plausible learning.

Shifting the perspective from "brain as computer" to "brain as information‑processing system" has practical consequences for research and industry. It encourages scientists to focus on how neural circuits transform sensory inputs into adaptive outputs, rather than seeking a universal programming language for cognition. This framing aligns with contemporary computational neuroscience, which models predictive coding, hierarchical inference, and embodied interaction. For AI developers, it underscores the value of learning algorithms that emulate biologically grounded plasticity instead of relying on brute‑force weight tuning, fostering more robust, adaptable, and energy‑efficient technologies.