Fork 1: Will AI Progress Fizzle Before Transformative Impact?
The opening gate of Scott Aaronson & Boaz Barak's decision tree: Does AI hit fundamental data, energy, or architectural walls (YES → AI-Fizzle, ~3%), or does it achieve revolution-scale transformative impact before 2050 (NO → 97%)?
In 'Five Worlds of AI' (April 2023), Scott Aaronson and Boaz Barak adapt Russell Impagliazzo's famous 1995 complexity taxonomy to map humanity's AI future.
The first branching question asks: 'Will AI progress fizzle?' By late 2026, with frontier coding and reasoning agents already deployed across enterprise software and scientific workflows, the Metaculus community assigns ~97% probability that AI will NOT fizzle before 2050, and ~3% probability that fundamental sample-efficiency, out-of-distribution fragility, or energy bottlenecks cap AI below transformative impact.
Verification Note: Metaculus Question #20683 community distribution assigns ~3% to AI-Fizzle and ~97% to the four transformative worlds combined.
Will AI progress fizzle out before 2050 (delivering less civilizational impact than the Industrial or Computer revolutions)?
Resolution Criteria: Resolves to AI-Fizzle (3% baseline) if two consecutive annual Metaculus panel votes or the final 2050 vote conclude AI failed to trigger a revolution-scale transformation.
Even if scaling hit a wall tomorrow, integrating existing frontier models into the global economy already virtually rules out a true 'AI-Fizzle' by 2050.
When Boaz Barak and I defined AI-Fizzle in April 2023, we compared it to nuclear power or supersonic flight—technologies that worked but plateaued below a general societal revolution. With 24+ years remaining until 2050 and agentic coding/research systems already transforming software and science in 2026, the burden of proof for a 25-year global stall is extraordinarily high.
Why 3% for AI-Fizzle is calibrated: Metaculus's two-consecutive-annual-vote rule heavily favors an early 'Futurama' resolution in the 2030s before any multi-decade plateau could be declared.
Under Sylvan's resolution mechanics on Metaculus #20683, the question resolves early if the annual panel votes unanimously for the same world two Januarys in a row. Because AI-Fizzle is the absence of a revolution, the panel is unlikely to declare AI-Fizzle unanimously in 2032, whereas a booming tool-AI economy in the 2030s can lock in a Futurama resolution early.