OK - Here goes my best shot at answering the Anonymous 1/07 question...
Lisi wrote down the Action S, not the Entropy S. It is simple enough to define Entropy. Following the methods of Claude Shannon, we could define S = (-k)*(Sum over i of (pi*ln(pi))) where pi is the occupation probability of the ith particle state.
The problem is deriving this from the first principles of an Exceptional TOE, but I think it can be done.
The first step is to determine the correct Exceptional TOE. E8 is nice, but E8, E10, E11, E12, E14, etc. all lead to different particle multiplets and branes. The correct gravity-brane and spinors (or twisters) will reproduce General Relativity (including Minkowski's metric) and Dirac's Large Number. Anything else is a waste of time.
The second step is to take the limit as this Exceptional TOE approaches E-infinity. I think this intersection is the key to the statistical nature of Thermodynamics (including Entropy) and the Path Integral formulation of Quantum Mechanics. I think that this is also the root of String Theory's 10^500 parameters, the Planck scale, Einstein's Hidden Variables, and Dirac's Large Number - some of the greatest mysteries of Modern Physics. Please refer to the upcoming papers by Leonard Malinowki in CS&F for more details.
Lisi's "Theory of Everything" is not intended to include "everything" at this point, but it (or a similar Exceptional TOE) certainly has the capacity to include such.
And I would like to restate that an Exceptional TOE is not necessarily opposed to String Theory. The correct Exceptional TOE should be able to reproduce parts of String Theory (if both the Exceptional TOE and String Theory are true, then the Correspondence Principle would demand this). The wrong Exceptional TOE will understand its deficiencies relative to String Theory and will, therefore, oppose String Theory.
Sincerely, Ray Munroe