One can of course question the words "everything" and "tendency". These tendencies are certainly outside the mainstream research programs however, so much is clear. So maybe it should be left at that and not dicussed here?
But I felt hit by the question and focused on the why part.
We have for example attemps for gravity, And also Ariels more general approach, including attemps for quantum logic.
https://en.wikipedia.org/wiki/Entropic_gravity
Verlindes gravity -
https://arxiv.org/abs/1001.0785
From Information Geometry to Newtonian Dynamics, Ariel C,
https://arxiv.org/abs/0710.1071,
https://arxiv.org/abs/0808.1260, https://www.albany.edu/physics/ACaticha-EIFP-book.pdf
My own personal research is also very much in this direction, and there are parallells to the ideas to abstract physical process as "computations", and when you combined that with statistical inference, there are many very exciting deep possibilites to "explain" things in a very deep way. So even if i agree that this is not mainstream, there exists such a tendencies and the logic behind them is imo sound.
The ultimate vision here is that the LAWS of physics themselves are to be understood as equations to state, ie. as equilibrium conditions. Its then easy to understand the advantages this may have for unification of laws IF successful. So there is great potential.
But to put the finger on the main problem. None so far has been terribly successful and i think its because many attempt to start with an universal objective microstructure, but this unavoidable IMO will cause problems with many things. So I think the "entropic forces" must be understand to be conditional upon observers, and thus - like any computation - relative to the hardware. What is natural or easy depdends on the hardware.
I will leave it there.
/Fredrik
