One potential solution to an initial permanent colony on Mars would be a series of modules with dirt floors and walls and slightly convex, transparent roofs. Assuming the interior temperature was artificially maintained at a human comfortable 22 C, it would take significant time for the walls and floors to come close to thermal equilibrium. However, once they did, they would provide nearly infinite thermal insulation on all sides, except the top. Their large thermal inertia would also be helpful in mitigating Mar's large diurnal temperature cycles. Lastly, they would block the constant bombardment of galactic cosmic rays (GCRs) on all sides, except the top.
Considering these aspects, the idea of human habitat modules built with dirt floors and walls (possibly covered with some sort of masonry) seems worth exploring further. The problem is, what we think of as dirt on Earth, usually contains a significant component of organic detritus. Assuming the existence on Mars of regolith, sand, clay, and bedrock, if you tried to dig a big hole, what would you be digging in? It would seem that a combination of sand and clays as a dirt, or dirt substitute, could be structurally sound. So my question is: is there dirt or its equivalent on Mars? Maybe a better question might be "How can you create a large, deep hole in the ground on Mars?".