I'm thinking that a marriage of physics, mathematics and CS might be necessary to overcome our limits in understanding these structures. Something like an IBM Watson for physicists, where a computer is fed with all informations we have and solves an optimisation problem to come up with a unified theory explaining all the phenomenons with the least complex solution (i.e. the least universal constants). Another requirement would be to have 42 as an error code for all possible failures in the calculation ;).
We know quantum constants. Is a cosmological bivector so different? Should be observable?
If only we could measure universal expansion accurately at small timescales, we could listen to the rain on the roof (horizon).