Abstract: Song and Bernevig proposed a topological heavy fermion description of the physics of magic angle twisted bilayer graphene (MATBG), describing the hybridization of two electron fluids: a lattice of localized moiré “atoms’’ with a topological relativistic conduction sea to capture the seemingly dual localized and delocalized nature of MATBG in a mixed real and momentum space model that is faithful to the symmetries of the system. We explore the implications of this model in the normal state, synthesizing insights from heavy-fermion physics and MATBG experiments. We [1] identify a key discrepancy between measured and calculated onsite Coulomb interactions, implicating renormalization effects beyond the current treatments of the model. With these considerations in mind, we consider a topological heavy fermion model with a single, renormalised onsite interaction between the localized f-electrons, containing a phenomenological heavy-fermion binding potential on the moiré AA sites that deepens with total filling factor. This feature allows the simplified model to capture both the quantum dot behavior and the observed stable local moments alongside the periodic reset of the chemical potential with filling, associated with the sloshing of electrons out of the topological relativistic sea into the localized state. There are two salient differences between MATBG and conventional bulk heavy-fermion physics that we glean from our approach, the first is that the relativistic character of the topological conduction sea in MATBG profoundly affects the Kondo effect, leading to two characteristic energy scales: a high-temperature scale corresponding to the onset of the hybridization between the localized electrons and the topological conduction electrons; and a low-temperature scale associated with the emergence of flat-band conference and scaling toward a conventional heavy Fermi liquid. Secondly, contrary to rare-earth materials, the optical phonon dynamics are fast compared to valence fluctuations of the localized f-electrons. We demonstrate [2] in a toy single impurity Anderson model that onsite interactions can be renormalized provided that optical phonons are sufficiently “fast’’. This suggests that fast optical phonons in MATBG may provide the necessary source for onsite Coulomb renormalizations to reconcile the ab-initio and experimental discrepancy.