Seeing the Future
AI and machine learning are enabling more accurate predictions across fields like weather and economics. This predictive power offers benefits for decision-making but also risks from over-reliance on flawed models, reshaping understanding of the future and human agency.
Background
- The article is by Jackson Price, an independent writer on Substack who covers technology, AI, and society.
- Substack is a newsletter platform where writers can publish directly to subscribers, bypassing traditional media gatekeepers.
- The piece explores how experts and investors try to anticipate technological breakthroughs — particularly in AI — and why predictions so often fail. It draws on the history of tech forecasting (e.g. Moore's Law, the dot-com boom, the "AI winter" periods when progress stalled).
- It also discusses how the current wave of generative AI (ChatGPT, image generators, etc.) is reshaping the landscape, but that seeing the future clearly remains extremely difficult even for insiders.
- Key context: the reader should know that AI has gone through several boom-bust cycles of hype and disappointment; that large language models (LLMs) like GPT-4 are the latest major advance; and that there is deep disagreement among experts about whether AGI (human-level AI) is near or still far away.