Many reactivity frameworks (e.g. observablehq, shiny) recompute intelligently - they're aware of what parts of the calculation has changed and needs recomputing

Many reactivity frameworks (e.g. observablehq, shiny) recompute intelligently - they're aware of what parts of the calculation has changed and needs recomputing. Haven't checked with ipyflow, but this idea would help mitigate some of your concerns