I actually don't really like Record types in the way people/library maintainers often use them - the type-checker asserts that values are actually present for all the specified keys, which is fine if the objects with the Record type really were exhaustive; but instead I often see them used where the reality of the data is a Partial<Record> - some keys are missing. Something about the abstraction causes people to misuse it frequently.
You pretty much always have to define your record type with `| undefined` tacked on to the value type parameter. With that, the problem mostly goes away.
