TECHNOLOGY
Neural Networks
The class of machine-learning architectures loosely modeled on biological neurons — the substrate of the current AI revolution and the opposite of Asimov's designed-then-programmed positronic brain.
Neural networks are computing systems composed of interconnected nodes that learn by adjusting weights on their connections based on exposure to data. Modern deep neural networks — trained on text, images, audio, and code — are the technology behind every significant contemporary AI system, from image classifiers to large language models. Unlike Asimov's fictional positronic brains, real neural networks are neither designed top-down nor inspectable by their creators.
In The You On AI Field Guide
The Orange Pill Asimov volume uses neural networks as the point of contrast with the positronic brain. Where Asimov's robots have explicit rules operating on a designed substrate, real AI systems have implicit preferences operating on a learned substrate. This asymmetry is not a shortcoming to be fixed; it is a fundamental property of the architecture that made modern AI possible at all.
Every technique for understanding neural network behavior — prompt engineering, interpretability research, red-teaming, RLHF — is a workaround for the fact that the system's "reasoning"