3 ms·

A lot of the advancement boils down to LLMs reprompting themselves with better prompts to get better answers.

• chimprich 1y ago
  Like an inner conversation? That seems a lot like how I think when I consider a challenging problem.