1 comment

[ 0.20 ms ] story [ 9.9 ms ] thread
Super interesting to see how - for the same set of prompts - GPT and Claude rely on just 50-60 domains while Perplexity and Gemini cite 100-140.