about mnovum

A century of information theory and computer science has shown that knowledge and learning are not aspirational ideals to pursue, but instead, they are concrete design problems to solve. While machines and humans operate under vastly different constraints, the conditions of intelligence are structurally equivalent. Intelligence emerges from reliably accessible memories, pattern detection, and recombination. Until recently, machines have struggled with pattern detection and recombination, while completely dominating the memory component. Meanwhile, pattern detection and recombination are things humans do without even trying. But for humans, the bottleneck has always been organized memories- and that is what we are all about.

Unlike the conditions under which our education system evolved, we live in a world saturated with high quality knowledge with answers to almost every conceivable question. It takes decades of education just to learn what is left to be learned. Incremental advances in increasingly constrained domains lead to diminishing returns— and this precisely the kind of work machines now perform best. Our vision for human learning focuses on rapid understanding of knowledge domains, so that patterns between them emerge. While there is no formula innovation and creativity (they are by definition unknown unknowns), we can be certain that detected patterns and recombination emerge from known knowns- retrieved from memory.

The topic of education often evokes vague and grandiose values that are out of touch with the challenges facing contemporary education. For our part, human learning holds together only when it is intrinsically interesting, rewarding to master, and self-perpetuating. As a design problem, they are these constraints that drive our solutions to domain expertise. We hope you enjoy them.