EE 292P: Atoms, Bits (Feb 24)
Attendees: Dustin J Ross Date: February 24, 2026 Type: Class Session
Summary
Course Context & Structure
- EE 292P: Atoms, Bits, and National Interest class session
- Guest speaker: John Hartley (economist, Stanford PhD, Hoover Institution policy fellow)
- Background: Goldman Sachs, World Bank, IMF, US Congress Joint Economic Committee
- Runs podcast on capitalism, multiple research affiliations including UT Austin
- Class progression: semiconductors → computing → applications → innovation → society → economics → geopolitics
AI Adoption Research Findings
- Survey methodology: nationally representative surveys across countries asking about generative AI use at work
- Global adoption rates: ~40-45% in Anglosphere countries (US, New Zealand, Australia, UK)
- Lower adoption in Italy, Netherlands, Sweden, Japan
- US timeline trends:
- ChatGPT launch: November 2022
- Massive surge: early 2024
- Adoption plateau/decline: second half 2024
- Recent recovery trend
Adoption Patterns by Demographics
- Education correlation: higher education = higher AI tool usage
- Industry breakdown (highest to lowest usage):
- Information services/programming
- Management/consulting companies
- Real estate
- Education, construction
- Agriculture, mining, government/military, healthcare (lowest)
Productivity Impact Analysis
- Time savings data across tasks:
- Writing: 25 minutes with AI vs 80 minutes without
- Significant time savings (~1 hour) across various tasks
- User behavior: 85% report AI helps complete tasks faster (complement), 15% report AI completes entire tasks (substitute)
- Income distribution effects: U-shaped curve with highest time savings for lowest income and highest income workers
Labor Market Effects Study
- Methodology: difference-in-differences analysis comparing high AI-exposure vs low AI-exposure occupations
- Key findings: minimal statistically significant labor market impact so far
- No major changes in job postings, employment levels
- Slight negative wage effect but economically small
- Conclusion: limited evidence of massive job displacement to date
Economic Theory Implications
- Creative destruction concept (Schumpeter): dynamic markets allow new technologies to disrupt established companies
- Policy tension: protecting existing companies vs allowing innovation
- Current debate: open-weight models vs closed foundation models
- Productivity-wage relationship questions: productivity gains not always immediately reflected in wages