There will be a historical post-facto rationalization of what went down for anything happening in the recent past, but the meat we draw from historical analogies is often poorly applicable since the divergence between the map and the territory is so large. The difference between reality and a game of chess is indeed paramount here, but for different reasons. The past is so complex that the lessons we draw from it are likely to just be distorted narratives that are already just a reflection of the present.
>To be honest I'm not even sure I can think of a way you can completely ignore history and still be able to do useful work of any kind.
In Levandowski's own example, he learned what existed in his time and then built useful things from there without worrying too much about the details of past beyond the directly visible iceberg spire in the present. This is broadly applicable to many fields and there are many people who did the same. It's tautologically true that disciplines that need to draw on the past to move on will do so, such as biology, but that doesn't say all that much about engineering for instance.
Since when is reality like a game of chess? Setting aside human history for a moment, just think about natural history (Levandowski mentioned dinosaurs and Neanderthals). Evolutionary biology is to a significant extent a historical discipline - it looks at the record and tries to come up with ways in which animals have changed over time under all sorts of pressures. Can that endeavor be done a-historically? How about even something like cosmology - the origins of the universe etc.?
The universe is not a Markov process. Thinking of it as such is incredibly limiting as an intellectual paradigm.
To be honest I'm not even sure I can think of a way you can completely ignore history and still be able to do useful work of any kind.