“Biological anchors” is about bounding, not pinpointing, AI timelines

Abstract

The author assesses the validity of “biological anchors” as a tool to predict the development of transformative AI. They argue that while the framework, which relies on extrapolating from current trends in AI development, has its limitations, it is useful for establishing bounds on potential timelines for transformative AI development. The author discusses several weaknesses of the model, such as its overemphasis on computing power, but finds it valuable for establishing plausible scenarios and for providing a basis for making informed judgments on the probability of transformative AI within given timeframes. They further emphasize that the biological anchors framework should not be interpreted as a precise prediction but rather as a tool for generating reasonable bounds on a range of possibilities. – AI-generated abstract.