I see ... in all nearby Everett branches he would have the same kind of cancer, but make different decisions. Further out, the cancers would only be subtly different from each other. Further out still, he might not have cancer at all, rather than different kinds of cancer.
Yes, it would have been much more accurate to speak of a perfect inference machine with the same information; only then could I integrate over all the possible cancers he 'might have had' as I intended. Thanks.
In nearby Everett branches Jobs would have made the same decisions.
EDIT: I think sometimes the concept of Everett branches are used in a misleading way on this website when people use the language of large-universe cosmology.
From Quora: