Why do mature faults slip at low stress without measurable heating?
In plain words
Rock friction measured in the lab predicts that a sliding fault like the San Andreas should heat the ground around it. No such heat is seen, so the fault must be weaker than lab rocks.
Precise statement
The absence of a frictional heat-flow anomaly over the San Andreas fault implies mean shear stress below about $2e8\,\mathrm{dyn}/\mathrm{cm}^{2}$, while Byerlee friction (coefficient 0.6 to 0.85) at seismogenic depth predicts about $1e9\,\mathrm{dyn}/\mathrm{cm}^{2}$. Determine whether mature faults are statically weak (weak clay minerals, high pore pressure) or only dynamically weak during slip. An answer is a measured in situ shear stress and friction profile consistent with heat flow and stress orientation data. The residual question concerns locked, seismogenic fault segments.
What would settle it
Deep borehole stress, pore-pressure and temperature measurements on mature faults combined with friction tests on recovered fault rock.
Status in the literature
SAFOD core showed the creeping San Andreas segment is statically weak through saponite-bearing gouge with friction coefficient below about 0.15 (Lockner and coauthors, Nature 2011, https://doi.org/10.1038/nature09927); whether locked seismogenic segments are statically or only dynamically weak remains open.