Unknown
His post on RL and inference compute scaling for frontier AI models is the central focus of analysis and critique.
How media typically covers Toby Ord
Toby Ord as author
The shift from pre-training to reinforcement learning for frontier models reduces information efficiency by 1,000 to 1,000,000x, potentially constraining future scaling progress despite unlocking new reasoning capabilities.
“Author of "The Extreme Inefficiency of RL for Frontier Models"”