The predicting computer
I agree with you on the following
* Our knowledge on laws of nature comes from observations of the past or present.
* If we are interested in knowing why things are the way the are at (i.e. how the became the way) current, we must study the past.
But given knowledge about laws of nature (regardless of how we got that knowledge) and given knowledge about the present state, we can predict the future - if laws or nature only gives one possible outcome.
Novelty is just an illusion in such a universe (I would say thats terrible spontaneously), it something that appears to be unpredictable but was bound to happen all the time. Was that your point too?
If there are any laws of nature for everything, then the knowledge about them - all of them - is enough. What preceeded them in another universe does not affect us today and is there for not of interest in order to predict things.