Conference Presentation, Fireside Chat, Interview
Uncharted: Big Data as a Lens on Human Culture
- Erez Aydin envisions big data as an engaging "adventure" and predicts that automated n-gram analysis of English books spanning five centuries will allow the examination of cultural trends and the detection of censorship patterns without reconstructing underlying texts.
- Regarding computational capabilities, the outlook maintains that computers will remain significantly inferior to humans in reading for the foreseeable future and are unlikely to achieve strong AI status capable of effectively comprehending complex human communication.
- Predictions on information velocity indicate that the time for inventions to reach 25% of peak prominence has decreased from approximately 65 years to 26 years, while celebrity fame now peaks earlier at age 29 compared to 37 in 1871, with a corresponding acceleration in the rate of being forgotten.
- Future language trends are projected to see the word "sustainable" appear once per page by 2036, once per sentence by 2061, and compose entire sentences exclusively by 2109.
- The summary forecasts that censorship can be objectively measured via frequency declines, citing specific examples such as the rapid disappearance of references to the 1989 Tiananmen Square events in Chinese literature and historical cases like Marc Chagall.
- Career trajectories suggest actors as the optimal path for early fame, politicians for maximizing fame through delayed gratification, and mathematicians as an inefficient choice for public recognition until one's 70s or 80s.
- Market and innovation plans anticipate a market for betting on n-grams based on cultural inertia, expect the next 20 major data visualization projects to be created by teenagers, and foresee the application of these techniques to quantify social trends in stock markets and politics.
- Risks and challenges include the accelerating decay of public memory, the persistent societal difficulty in quantifying people for hiring processes, and the potential for automated tools to be used for both censorship detection and the suppression of freedom of information.
- Speculative future scenarios, such as a "technological singularity" creating a "silicon hive mind," are acknowledged but noted as beyond the speaker's expertise.
- Google's scanned book collection is expected to eventually encompass roughly 25% of all published books, serving as a foundation for the described analytical capabilities.