My understanding is that gravitational systems are not predictable in the long term (this is called the N-body problem in astronomy). As the SI has perhaps a 600-billion-body problem to solve at least 100 billion years into the future, where some of the bodies may not follow "natural laws" (i.e. they are SIs that decide to change course for independent reasons), this could very well take a significant amount of computation.
A system with a load of variables that we can't track at this tech level is not "unpredictable", it is "very hard to predict", if you want to be correct.
This footnote also confirms that this email message has been swept by MIMEsweeper for the presence of computer viruses.