Speculative Programmatic Tool Calling

Alex Zhang · · rlm applied_ai

This feels like a natural evolution of the RLM paradigm. With most interpreted languages, you don’t need to wait for the entire code block to be generated for you to start executing calls.