Automated AI research yields massive speedup
full automation of ai r&d likely generates 3.5 years of progress in one year even without a software-only singularity.
this analysis indicates that full automation of ai research delivers substantial acceleration even without a critical software intelligence explosion. a model with an r-value of 0.7 predicts achieving 3.5 years of progress in the first year without additional compute scaling. further gains arise from a feedback loop where increased compute improves the ai researchers themselves, maximizing hardware utilization and amplifying the impact of further compute investments by 2x-4x.