Can System 1 LLMs Ever Show System 2 Thinking?
How to make System 1 LLMs show System 2 competencies?
A Wiley paper asks how fast, intuitive LLMs might display deliberate, System 2 competencies. The Hacker News discussion weighs whether prompting, architecture, or training could bridge the gap between quick pattern matching and slow reasoning.
How to make System 1 LLMs show System 2 competencies?