We’re Building AI We Can’t Understand and Trusting Its Creators to Warn Us
中文摘要
前OpenAI预测员认为超级智能或在2030年前出现，探讨了开发不可理解的AI及过度依赖开发者预警的风险。
English Summary
An ex-OpenAI forecaster predicts superintelligence by 2030, highlighting risks of building incomprehensible AI and relying solely on creators for safety warnings.
Original Excerpt
A former OpenAI forecaster believes superintelligence could arrive before 2030. His extinction estimate is debatable. His argument for… Continue reading on Medium »