This post lays out legible arguments for its position, which I consider to be one of the best ways to drive conversations forward, short of demonstrating convincing empirical results (which seem like they’d be difficult to obtain in this domain). In this case, I hope that future conversations about sharing LLM capabilities focus more on object-level details, e.g. what evidence would bear on the argument about LM agent “overhang”.
Curated.
This post lays out legible arguments for its position, which I consider to be one of the best ways to drive conversations forward, short of demonstrating convincing empirical results (which seem like they’d be difficult to obtain in this domain). In this case, I hope that future conversations about sharing LLM capabilities focus more on object-level details, e.g. what evidence would bear on the argument about LM agent “overhang”.