Complexity in Language Models
Tim discusses the exponential complexity of problems in relation to language models, questioning why they can effectively understand lengthy texts despite this complexity. Jan highlights that many challenges can be simplified, revealing that the actual decision-making required is minimal, suggesting that language models could potentially scale to larger problem sizes without significant issues.In this clip
From this podcast

Machine Learning Street Talk (MLST)
Daniel Franzen & Jan Disselhoff - ARC Prize 2024 winners
Related Questions