I loved the video too and initially felt the way you felt. But as I thought more about it, I don’t think that value is found in doing something that nobody ever has done before, that might be nay impossible for the majority of humanity. I think that new for the individual is very valuable, like, how to solve quadratic equations has been solved for a long time, but it’s still new to new people all the time!
But even more powerful is the application of the not-new to a new setting! One might think that because, say, proper data management has been “solved” (as in, there are tools out there like SQL and graph DBs and such), it is solved for everybody, but it only takes one glance at huge international companies that are held together by an assorted array of excel files moving around in a semi-random way with no data types, enormous space for human error, and some other atrocities for data management… it’d be of great value for such a fictional organization that has never existed to do something “new” for them, use and actual database!
Looking at the cellular automatons video, I felt it captures the early sense of wonder learning about math and programming. But also, the despair in what problems are actually kind of open. What if I design a cellular automatons where every neighborhood is in the shape of a Disney character and every value it can take corresponds to the release date of one movie that was in the global top 10 of highest grossing movies? Well, turns out that's a special case of a Heisendorff-Circlydot automatons, which was solved using topolographical methods in the 80's and the only open question have been proven to be equivalent to proving the consistency of Riemann.
Unless say you give each cell a bounded notebook with a given number of page, in which case the problem falls into a weird computability class between the gaps of the Chomsky hierarchy. And then it's thesis material :-)