Why the Soviet experiment still would have failed even with better computers Some thoughts on why AI-based central planning of the kind @RokoMijic describes won't work ** 1 Rational allocation requires economic calculation. Removing free price formation hinders the cost-benefit calculations that guide efficient resource use toward ever-changing consumer preferences. Mkts provide a crucial surround of computational guidance & feedback 2 Firms are islands of central planning within market ecosystems. They may allocate internal resources based on targets, but rely on market price signals & competitive feedback to determine what outputs and opportunities merit investment 3 The economic knowledge most relevant for coordination consists of highly dispersed, subjective insights into individual goals, skills, habits, & specific circumstantial details--not aggregated datasets. Critical innovation manifests as situational judgment and tacit skills applied flexibly. Mkt structures harness this undisclosed expertise while price shifts connect decentralized responses, transmitting localized insights across the system 4 Through prices, markets incentivize knowledge discovery and generation. They reward self-interested risk-taking, disclosing knowledge about needs, costs, innovations etc. Aligning decentralized participation with secure property rights radically enhances system insights (including about creative destruction opportunities) and produces knowledge that would not have existed otherwise 5 Imposing fixed prescriptive plans limits emergent discoveries. Markets enable coordination precisely bc they afford flexibility & let individuals apply expertise and adapt responses using their local judgment. This fosters decentralized learning beyond any single prescribed vision ***note that these critiques say nothing about the totalizing tendencies of such a system & its potential effects on freedom & the flattening of human life