we’re making improvements to sonnet 3.7 in @cursor_ai 0.47, allowing you to delegate even harder tasks to agent and unlock the full potential of the model 3.7 is incredible when you let it operate like intended: long tasks, complex problems and hard thinking let me explain↓