Tracking (Expert/Influential) Predictions About AI

Key Takeaways

• Aggregates predictions from Metaculus, Good Judgment, Manifold, web sources
• Includes non‑opt‑in experts like researchers, public intellectuals
• Provides per‑person history UI, enhancing transparency
• Addresses incentive misalignment causing vague AI forecasts
• Risks concentration of trust; mitigated by diversified tracking

Pulse Analysis

Forecasting AI development has become a cornerstone of strategic planning, yet most existing platforms rely on participants who voluntarily submit predictions. This creates a blind spot for high‑profile voices—lab leaders, futurists, and commentators—who shape public expectations without formal accountability. By scraping interviews, podcasts, and social media, the proposed tracker would fill that gap, offering a centralized repository that captures both formal and informal forecasts. Such a database would enable analysts to compare track records, calibrate confidence, and identify systematic biases across the AI forecasting ecosystem.

For investors, regulators, and corporate strategists, the value lies in a clearer signal‑to‑noise ratio. When a well‑known AI researcher consistently overestimates timelines, stakeholders can adjust capital allocation or policy stances accordingly. Conversely, accurate forecasters gain credibility, fostering a merit‑based reputation system that rewards precision over hype. The platform’s per‑person UI would let users quickly assess an individual’s historical accuracy, supporting more nuanced deference than simple aggregate scores. This granular insight is especially critical as artificial general intelligence timelines tighten and the cost of misjudgment escalates.

Implementing the tracker raises challenges, notably the risk of creating new trust hierarchies that amplify a few dominant voices. Mitigation strategies include displaying raw prediction data, weighting outcomes by uncertainty, and encouraging a diversified pool of tracked experts. Open‑source moderation and transparent scoring algorithms can further guard against manipulation. The initiative invites developers and forecasters to collaborate on a prototype, aiming to launch a functional MVP over a weekend and spark broader community adoption. By institutionalizing accountability, the project aspires to elevate AI epistemics and improve collective foresight ahead of transformative breakthroughs.