I feel like whenever I use ChatGPT, Claude or Gemini, I rarely get a focused, short and to the point answer. What the LLM spits out is super useful but I feel sometimes I’m just looking for like a super quick response : a yes/no, maybe 4-5 bullet points or a quick table.
Apparently this is a research topic in AI and something that AI researchers are working on and LLMs spitting out long responses has to do with how they are designed and work - they generally generate longer responses.
Ideas?