Interview
How The AI Fight Will Play Out | Dean Ball, author of America's AI plan
- General superintelligence is anticipated within five to ten years, with robust superintelligence in specific cognitive domains relative to top humans arriving in one to ten years.
- Physical infrastructure integration of AI systems is expected to become significantly more common within twenty years, though the process will be delayed by diffusion rates and capital upgrade cycles.
- A compute constraint is projected to limit AI development for at least five years and potentially up to ten years.
- Fully automated researchers are expected to eventually face diminishing returns across all areas of scientific research, even with accelerated navigation of micro-paradigm S-curves.
- Gross domestic product is projected to be substantially higher by 2040, though this economic growth may not correlate with increased human well-being.
- Future social structures may see the majority of humans relegated to "gift economy" roles or "rent seeking," while AI performs the vast majority of cognitive labor.
- The federal government is expected to shrink to an 18th-century scale, automating technocratic implementation to focus on high-level political decisions rather than current institutional extensions.
- Federal AI adoption is predicted to be slow due to legal barriers such as the Freedom of Information Act, data privacy laws, and rigid IT budgeting, despite initial usage by staff on personal devices.
- Open source AI is expected to be framed as a competition between companies and humans rather than a competition between models.
- Legal risks for AI developers are anticipated to increase as the liability system is expected to lack the protections of Section 230, unlike social media platforms.
- Various industries and professional groups may establish regulatory moats through interest group politics, potentially leading to over-regulation as AI advances.
- The U.S. and China may reach an agreement involving intense competition in AI applications alongside a mutual assurance treaty to prevent premature superintelligence training or deployment.
- The political landscape may feature a "critical left" remaining skeptical of AGI capabilities, while the political right becomes more prone to "AI doom" narratives and alliances with anti-tech populists.
- Tension is expected between the "new tech right" and the traditional "MAGA wing" regarding AI, with the resolution dependent on economic growth and labor market outcomes.
- Public perception of AI policy is predicted to become more toxic and divided, characterized by a small technocratic group of early adopters and a larger, less nuanced population entering the discourse.
- Former President Donald Trump is characterized as likely to be a dovish figure regarding China, more willing to negotiate than the previous administration.
- Personal expectations include raising children with enduring human traits like charisma, sprezzatura, and gravitas to maintain relevance in an AI-dominated future.