Will Bryan Caplan win his bet with Matthew Barnett on whether an AI can pass his exams in 2029?
💡 What the odds say
The market puts this at about a 6% chance — very unlikely.
No money — just record your call and see if you were right. Yes is at 6% right now.
Bryan Caplan's track record of winning bets on future predictions, combined with his early skepticism after testing ChatGPT in 2023, underpins the market's strong belief that AI will not pass his exams by 2029, making the 6% Yes price a reflection of Caplan's confidence rather than a neutral assessment of AI progress.
What's driving it
- • The Guardian article (Apr 2023) highlighted Caplan's perfect record on past bets and his immediate skepticism after testing ChatGPT, which likely set a strong prior against AI passing his exams.
- • Caplan's bet with Matthew Barnett is structured around his own exams, which are designed to test deep economic reasoning, a domain where current AI systems still struggle significantly.
- • The 94% No price reflects the market's assessment that AI will not reach the required level of nuanced understanding and original argumentation by 2029, given the slow pace of progress in such specialized tasks.
The case for YES
- • If AI capabilities accelerate unexpectedly, such as through a breakthrough in reasoning or a new architecture, it could pass Caplan's exams by 2029, especially if the exams are multiple-choice or have predictable patterns.
- • Caplan's past bets were on different domains (e.g., fertility, economic growth) and may not directly translate to AI performance; his skepticism could be misplaced if AI improves faster than he anticipates.
- • The bet's long time horizon (2029) allows for multiple generations of AI models, and even modest annual improvements could compound to meet the exam threshold.
The case for NO
- • Caplan's exams likely require original argumentation and deep understanding of economic theory, which current AI (including GPT-4) fails to demonstrate consistently, as noted in the 2023 Guardian article.
- • The market heavily discounts the possibility of AI reaching human-level reasoning in specialized fields within 3 years, given the lack of clear breakthroughs in causal reasoning or long-form logic.
- • Caplan has a strong incentive to design exams that are resistant to AI cheating, and he has publicly stated his confidence that AI will not pass, which aligns with the 94% No odds.
What to watch
- • Release of a major AI model (e.g., GPT-5 or equivalent) in 2027 that demonstrates near-human performance on economic reasoning benchmarks could shift odds toward Yes.
- • Publication of Caplan's exam results or a public test of AI on his exams before 2029 would provide direct evidence and likely move odds significantly.
- • Any announcement from Caplan or Barnett about modifying the bet terms or releasing preliminary results would act as a catalyst, with direction depending on the outcome.
AI-generated · grounded in recent news + odds · informational only, not advice. Verify on the source platform.
Data from Manifold’s public API, for informational purposes only. PredictPal is not affiliated with any platform and does not facilitate trading.
Discussion
Loading…
How it resolves
Resolved by whoever created the market, at their discretion per the question's description. It's play-money (Mana) and not tied to an official source — treat it as a community forecast.
ⓘ A market settles under its own written rules, which can lag what looks decided in the news — so the price may not move to 100% the moment an outcome seems obvious.
View the official rules on Manifold ↗Related markets
Which of these Language Models will beat me at chess?
The field is extremely concentrated on 'any model announced before 2034' at 88%, but the long tail of 22 candidates above 5% suggests bettors see many plausible paths to a 1900-rated human being beaten by a future LLM, with the single biggest recent shift being the 2025-08-15 Business Insider report that OpenAI's o3 swept a chess tournament against xAI's Grok 4, likely boosting confidence in near-term AI chess ability.
52 outcomes
Apple Announces AI Glasses by September 30, 2026
2 outcomes
Will Anthropic release its next Mythos-class model to the public by August 31, 2026?
Yes ≈ 46% chance