how are you applying concepts like "energy-dense" to, say, sunlight or geothermal?

Energy density refers only to fuels and energy storage media and doesn't have much to do with grid-scale investment, although it's important for things like transport where you have to move your power source along with you. (Short version: hydrocarbons beat everything else, although batteries are getting better.)

The usual framework for comparing things like solar or geothermal energy to fossil fuels, from a development or policy standpoint, is energy return on investment. (Short version: coal beats everything but hydroelectric, but nuclear and renewables are competitive with oil and gas. Also, ethanol and biodiesel suck.)

at the caveman level the fossil fuels are pretty much useless

Coal was used as fuel before the Roman empire. It didn't lead to an industrial revolution until someone figured out a way to turn it into mechanical energy substituting for human labor instead of just a heat source in a society where that could be made profitable due to a scarcity of labor. That was the easiest, surface-exposed deposits, yes, but you hardly need any infrastructure at all to extract the energy, and even mechanical energy extraction just needs a boiler and some pistons and valves. This was also true of peat in what is now the Netherlands during the early second millennium.

your imagination with respect to future technology seems severely limited. ... This entirely depends on the technology level.

What does 'technology level' even mean? There's just things people have figured out how to do and things people haven't. And technology is not energy and you cannot just substitute technology for easy energy, it is not a question of technology level but instead the energy gradients that can be fed into technology.

And how are you applying concepts like "energy-dense" to, say, sunlight or geothermal?

Mostly in terms of true costs and capital (not just dollars) needed to access it, combined with how much you can concentrate the energy at the point of extraction infrastructure. For coal or oil you can get fantastic wattages through small devices. For solar you can get high wattages per square meter in direct sunlight, which you don't get on much of the earth's surface for long and you never get for more than a few hours at a time. Incredibly useful, letting you run information technology and some lights at night and modest food refrigeration off a personal footprint, but not providing the constant torrent of cheap energy we have grown accustomed to. Geothermal energy flux is often high in particular areas where it makes great sense (imagine Iceland as a future industrial powerhouse due to all that cheap thermal energy gradient), over most of the earth not so much.

Sunlight is probably our best bet for large chunks the future of technological civilization over most of the earth's surface. It is still not dense. It's still damn useful.