Machine intelligence quality improves with data volume, unlike industrial production

Evidence

Google’s machine intelligence capabilities feed on behavioral surplus, and the more surplus they consume, the more accurate the prediction products that result. Wired magazine’s founding editor, Kevin Kelly, once suggested that although it seems like Google is committed to developing its artificial intelligence capabilities to improve Search, it’s more likely that Google develops Search as a means of continuously training its evolving AI capabilities. This is the essence of the machine intelligence project. As the ultimate tapeworm, the machine’s intelligence depends upon how much data it eats. In this important respect the new means of production differs fundamentally from the industrial model, in which there is a tension between quantity and quality. Machine intelligence is the synthesis of this tension, for it reaches its full potential for quality only as it approximates totality. — Shoshana Zuboff, The Age of Surveillance Capitalism, 301

Categorization

  • XREF: Connects to the reader’s interest in behavioral surplus as the defining mechanism of surveillance capitalism; the quality/fullness inversion is a genuinely counterintuitive economic fact. — connects to: Training models on absurd datasets exposes AI’s true capabilities — , The Braided Reading EPUB: 5-Year…, 1863; AI’s apparent generality exceeds what headline metrics justify — , The Braided Reading EPUB: 5-Year…, 1783; Big data and AI outperform human judgment in high-stakes decisions. — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law: A Definitive Look at…, 1171 (+1 more)

Broader Connections