Tag
A 75-page paper by over 30 researchers categorizes AI self-improvement into five levels and reviews 491 papers, finding that most current advancements are stuck at lower levels.
The paper introduces a model for recursive AI self-improvement, defining a recursive reproduction number to determine when incremental improvements in AI research become self-amplifying or dampening across development cycles.
This newsletter discusses a study on AI's limitations in recursive self-improvement and covers the causes behind recent heatwaves, along with other tech news including OpenAI's safety pause.
Former top researchers from OpenAI, DeepMind, and Meta jointly founded the startup Recursive, focusing on letting AI self-improve. In just 6 months with fewer than 30 people, they achieved a valuation of $4.65 billion.