we could imagine a scenario where we have an embryo in a test tube and are able to document every time it reproduces a new cell. using crispr-like technology, we could in theory go in and edit every single cell in real-time, as each error develops.
but, we can't know what those errors are until they happen - they don't come from your parents, they are the result of random errors occurring in the embryo. so, we won't know which embryos to monitor for errors as they are developing.
and, of course, this is a somewhat dystopic way to envision reproduction.
i have a better idea: let's ask ourselves if environmental causes (like introducing toxic chemicals into the environment) may have an effect on that error rate, and focus on that, instead.