Comment by nyellin

No, it was designed on paper by someone with no understanding of prompt caching and no consideration of latency or token costs