Regarding whether it's "pretty limited", yes and no, imo:
Yes, compared to, say, the size of a continent, the amount of new land is tiny.
However, new land is usually made in highly sought-after (high-density, high-value etc) areas [citation needed]. So it would make better sense to compare the surface of the new land to the surface of the adjacent area.
For example, the polders in the Netherlands significantly augment the country's surface. The artificial islands in Tokyo Bay significantly add to the real estate of the harbour. The artificial islands in Dubai and Bahrain significantly add to the coastline.
I think that this is enough for OP's analogy between domain parking and land speculation to reasonably stand.
A bit on the pedantic side, but you can make more land; there are dozens of artificial islands around the world, some of which have resulted from the need for more land.
The costs, of course, as well as other considerations (legal...) are of a different scale.
I've got one from Bosch (a name brand), and use it to check for steel reinforcements in concrete and power wires in walls (and, rarely, studs). I can't double check the readings non destructively, so I evaluate the device according to whether in the end I hit something or not when drilling.
The readings are imprecise at best, with lots of false positives and possibly some false negatives (no way to know).
The tool is useful, if repeated readings are coupled with domain expertise, common sense and intuition.
If you ask me if the tool, by itself, works, I'll reply "maybe". All I know is that, aided by the tool, I haven't done any damage yet.
I don't think it's the tool's fault though. Rather, it's the combination of non-destructive testing and the limited feedback.