You have to consider the scope of the knowledge we are able to obtain about the deep past. We certainly know nothing about the particular events of a given day 2 billion years ago, but we can determine lots of broad/granular knowledge based on the available evidence.

When it comes to physics, we are able to verify both the accuracy of models in the present and the consistency of these models with evidence from the past, all the way back to the inflationary period after the "big bang". The universality of physics across time and space is generally a necessary assumption to seek any understanding of the universe across time and space. This does not, however, mean that this assumption is baseless. A universe inconsistent with this assumption would almost certainly produce evidence of this inconsistency, which has never been shown to be the case. To the contrary, tests of universality on observations from the distant universe based on physical constants measured here on Earth have only ever been consistent with universality.

We can wax philosophical all day about the unknowability of all things outside our own minds; but ultimately we must work with the tools at our disposal, fallible though they may be, to understand our world. The same tools that allow us to know that a ball will fall when thrown also allow us to understand the early, and distant, universe.