Is this like Jupyter, where the state of the interpreter is preserved? Or if you ask follow up questions does every interpreter block have to calculate everythi

Is this like Jupyter, where the state of the interpreter is preserved? Or if you ask follow up questions does every interpreter block have to calculate everything from scratch? I'm still on the waiting list, or I would test it myself.