Latest

A normal camera takes a photo. An event camera only records change — each pixel fires independently the moment its light level shifts. That retina-like design cuts latency to microseconds and slashes power consumption.

Most chips are always on, burning power regardless of activity. Intel's Loihi takes a radically different approach: compute only when neurons fire, just like a brain — and the efficiency gains are dramatic.

Your neurons don't output a float between 0 and 1. They fire — or they don't — and when they fire matters as much as whether they do. The gap between spiking and continuous computation is wider than most AI researchers admit.

Your brain runs on roughly 20 watts. A GPU cluster doing comparable work burns thousands. Neuromorphic chips close that gap by computing with spikes instead of floating-point operations — mimicking the mechanism the brain actually uses.

Hyperdimensional computing encodes concepts, binds them, and retrieves them using simple operations on very large vectors. The approach is robust, fast, and brain-like. The deeper lesson: coherence in biological systems might emerge from the same high-dimensional geometry.

Two frameworks developed independently converge on similar ideas about how minds model the world. Hyperdimensional computing uses algebraic operations on vast vectors; active inference uses probabilistic predictions. Together they explain how brains do so much with so little energy.

Hyperdimensional Computing for Cognitive Architectures Series: Hyperdimensional Computing | Part: 7 of 9 In 1956, the first summer conference on artificial intelligence at Dartmouth launched a field based on a single working hypothesis: intelligence is symbol manipulation. Build the right data structures, implement the right algorithms, and cognition would emerge. Sixty-eight