Ilya SutskeverCo-founder of Safe Superintelligence and former OpenAI chief scientist.
Our goal is to solve the core technical challenges of superintelligence alignment in four years.
OpenAI, "Introducing Superalignment" (co-led by Sutskever) · Jul 5, 2023 · source
Normalized claim
The core technical challenges of superintelligence alignment will be solved by July 2027.
Confidence
Sounded fairly sure
Deadline
due Jul 5, 2027 (our reading)
Resolution rule
Hit if by July 2027 OpenAI or Sutskever's Safe Superintelligence publishes a method broadly accepted by alignment researchers as solving scalable oversight of superhuman systems. Miss otherwise; the OpenAI Superalignment team was dissolved in May 2024.
Crowd vote
Sign in to vote and see the crowd.
Your call
Will it happen? Your call locks at the deadline and is scored when the verdict lands. No odds, no money — just your record.
Forecaster
Ilya Sutskever
Co-founder of Safe Superintelligence and former OpenAI chief scientist.
Are you this person? Request a correction