Dwarkesh PodcastAlex Tabarrok - Prizes, Prices, and Public Goods
CHAPTERS
- 0:00 – 0:33
State capacity libertarianism: small government that can execute
Tabarrok opens with a core political philosophy: preferring limited government while insisting it must competently perform essential functions when crises hit. The framing sets up later critiques of pandemic response and institutional dysfunction.
- •Small-government preference paired with high expectations for execution
- •Government competence and timing matter as much as size
- •Foreshadows later discussion of CDC/FDA/Congress failures
- 0:33 – 2:38
Grand Innovation Prize for COVID: speeding vaccines, tests, and therapeutics
Tabarrok explains why pandemic innovation needs extra incentives beyond normal market rewards, especially to accelerate timelines. He lays out tools like big prizes, advance market commitments, and paying for at-risk manufacturing capacity.
- •Private incentives to move fast are weaker than social incentives during a pandemic
- •Key bottleneck: firms won’t build capacity until success is likely/approved
- •Policy options: prizes, guaranteed prices, and subsidizing manufacturing buildout
- •Portfolio approach: many shots on goal because most candidates fail
- 2:38 – 8:43
Prizes vs upfront funding: capital constraints, VCs, and “throw money at the problem”
The conversation turns to a practical challenge: great ideas may lack funding to compete for a prize. Tabarrok argues that in a crisis, speed dominates—so a mix of mechanisms and intentional waste may be justified given trillion-dollar stakes.
- •Government can’t pick winners well; neither can VCs quickly in complex biotech
- •VC due diligence and information problems take time—time we don’t have
- •Pandemic policy should accept waste to buy speed and diversification
- •Example risk: promising candidates (e.g., AstraZeneca trial pause) can stumble
- 8:43 – 14:09
Why grants dominate and when prizes/fast grants outperform
Tabarrok discusses why modern research funding relies more on grants than prizes, despite historical precedent for prizes. He cites evidence that “less strings attached” funding (e.g., Howard Hughes-style) can yield higher-impact research, and highlights rapid COVID “fast grants.”
- •19th-century prizes were more common; 20th-century shifted toward grants
- •Possible explanation: grants concentrate power among fund allocators
- •Evidence: flexible, trust-based grants correlate with higher citations/patents
- •COVID fast grants show speed advantages vs traditional NIH processes
- 14:09 – 21:41
The Baumol effect: why education, healthcare, and repairs keep getting pricier
Tabarrok introduces Baumol’s cost disease as a unifying explanation for rising prices in labor-intensive sectors with slow productivity growth. He reframes the issue: it’s often not dysfunction in the “stagnant” sector, but success in the fast-growing “progressive” sector raising opportunity costs.
- •Two-sector view: progressive productivity vs stagnant productivity
- •Relative prices rise for sectors that can’t scale labor productivity easily
- •String quartet example: same labor time, far higher opportunity cost today
- •Prediction: when overall productivity slows, price pressure in stagnant sectors eases
- 21:41 – 27:50
Wage stagnation, skilled labor, and automation: who gets squeezed
Dwarkesh presses on wage stagnation and AI fears; Tabarrok refines Baumol by emphasizing skilled labor as the key scarce input. He argues automation may not eliminate work overall, but it can sharply harm less-skilled labor, analogizing to horses after automobiles.
- •Twist on Baumol: skilled labor costs drive many “stagnant” sectors
- •Education/healthcare rely heavily on skilled labor with rising outside options
- •Service share rises naturally; doesn’t imply “no automation disruption”
- •Distributional risk: less-skilled wages pressured by automation, trade, remote work
- 27:50 – 31:35
Caplan vs Tabarrok on education: signaling, apprenticeships, and rebalancing pathways
Tabarrok concedes significant signaling in college, but argues the U.S. system is misallocated—too many non-completers and too much push toward college vs vocational routes. He contrasts U.S. patterns with German-style apprenticeships and criticizes mismatches between majors and labor-market demand.
- •Agreement: college includes substantial signaling
- •U.S. imbalance: lower high-school completion + very high college enrollment
- •Europe: higher completion and stronger apprenticeship/training pathways
- •Mismatch: growth in low-return majors; domestic CS flat while foreign students fill gaps
- 31:35 – 35:12
Scaling education online: MRU, AI tutors, and the coming teacher superstar market
Tabarrok predicts online delivery will reduce the number of teachers while increasing reach and potentially raising pay for top performers—like sports stars. He argues AI-assisted tutoring and adaptive assessment can personalize learning and dramatically increase productivity in education.
- •Online scale changes economics: one instructor can teach globally
- •Likely outcome: fewer teachers overall, higher rewards for the best
- •AI tutors can match human tutors in trials; adaptive feedback targets misconceptions
- •MOOCs won’t replace universities outright, but universities will move online
- 35:12 – 48:49
Why ‘this time might be different’: thresholds, user control, and automatic translation gains
Dwarkesh challenges techno-optimism with historical analogies (radio/TV/books). Tabarrok responds that combined technologies can cross a threshold (Newton vs iPhone), and highlights concrete advantages of online learning: pausing, pacing, and improving via external AI (e.g., better captions/translation).
- •Past ed-tech hype cautions humility; disruption isn’t guaranteed
- •Threshold effects: small tech improvements can trigger big adoption shifts
- •Online advantages: rewind, variable speed, self-paced learning
- •Progress in translation/captioning can improve courses automatically over time
- 48:49 – 52:13
Declining research productivity: low-hanging fruit and worrying innovation math
The discussion shifts to evidence that research is getting harder: more researchers are needed to sustain the same rate of technological progress (e.g., semiconductors). Tabarrok worries that even “progressive” sectors face diminishing returns, though he hopes for future leaps (e.g., quantum, new general-purpose tech).
- •Bloom et al.: sharply rising research inputs needed to maintain growth rates
- •Semiconductors/pharma show higher R&D costs per unit of progress
- •Possible explanations: low-hanging fruit exhausted, regulation, or tech lifecycle dynamics
- •Hope rests on new “quantum leap” technologies and global scientist growth
- 52:13 – 58:41
Dominant Assurance Contracts: turning public-good funding into a dominant strategy
Tabarrok lays out his mechanism for improving public goods provision: like Kickstarter but with a refund bonus if the threshold isn’t met. This changes incentives so contributing can be individually rational regardless of others’ actions, and lab evidence suggests it can significantly increase successful projects.
- •Public goods suffer from free-riding (non-rival, non-excludable)
- •Assurance contracts: pay only if threshold reached (Kickstarter-like)
- •Dominant assurance: add a refund bonus when threshold fails
- •Lab results: can roughly double successful projects vs standard assurance contracts
- 58:41 – 1:06:02
Future of governance: online experimentation, RadicalxChange, and futarchy
From DACs, the conversation broadens to governance innovation: mechanisms, not just ideologies. Tabarrok argues online communities and blockchains enable rapid experimentation with collective decision systems, and he endorses running trials of Robin Hanson’s futarchy (policy chosen by prediction markets).
- •Democracy as a governance mechanism: good at limiting worst abuses, weaker at preference aggregation
- •Online worlds lower switching costs (Tiebout-like), enabling many governance experiments
- •Blockchains create thousands of small-scale governance trials—most will fail, some may teach us
- •Futarchy is presented as a rare, genuinely new governance design worth experimenting with
- 1:06:02 – 1:21:39
Communication and influence: teaching economics at scale, policy frustration, and an ‘innovation state’ agenda
Tabarrok reflects on which media reach the most people (especially MRU) and jokes about “beating” Adam Smith in total students taught. He then pivots to policy: frustration with pandemic institutional failures, a call for state capacity, and a proposal to shift from a ‘warfare-welfare state’ toward an innovation-focused state—potentially spurred by international competition.
- •MRU’s global reach may surpass traditional books/textbooks in impact over time
- •Economists can sometimes influence policy, but adoption is slow and inconsistent
- •Pandemic exposed failures of CDC/FDA/Congress; need competent execution (state capacity libertarianism)
- •Rebalance budgets toward R&D: from warfare/welfare toward innovation; competition can catalyze effort
- 1:21:39 – 1:26:01
Advice to a 20-year-old: race with the machine by building complementary skills
Tabarrok closes with career advice tuned to rapid technological change. He stresses that returns to skill are rising and recommends building capabilities complementary to technology—especially data analysis—while noting design, marketing, and human-centered application remain valuable when paired with tech.
- •Older advice depreciates faster as the world changes
- •Returns to skill rising; invest in education that complements technology
- •Focus areas: CS, data science, causal inference, ML, engineering
- •Combine artistic impulses with tech via design and human-centric product building
