"Solar energy is at a scale of ~500-1000 watts per square metre. Radiogenic geothermal heat from a rocky planet like earth is of the order of 50 milliwatts per square metre, so it is an energy source at least 10,000 times smaller."

If you said solar energy is: "~500-1000 watts per square metre" at noon this would be correct or at least if you said "during the day" it would be closer.
But there 24 hours a day and 24 times 500 is 12,000 watts per day. There could be someplace on earth that gets 12 kilowatts per sq meter per day, but it certainly is uncommon.

In the US the best areas get about 8 kilowatt per meter on average per year. [And btw this is directing the sq meter to face the sun.] The average for the US is around 6 kw/hours per day. Go up to Canada and 6 kw/hr is the best areas.
Now the earth surface is a different animal than even selecting a location and having the solar panel set up at the proper angle to receive max solar energy without moving the solar panel to always face the sun- passive radiation collection.
So, here is a map of solar energy of US with southern tilt:

It shows about 6 to 7 kw/hour per day per square meter annually in best areas in the US.

So, I would say 250 watts per sq meter globally is probably on the high side.

So, if the earth emits on average of 50 milliwatts it's around 1/5000th of the energy coming from the sun and hitting the surface.
But this doesn't mean that 250 watts per sq meter of energy from the sun is being converted into heat, but rather, it is the most amount energy one could possibly capture to make heat. No fancy human contraption is going to able to get all that energy. Humans might get about 95% of the heat energy, nature will not.
So what I'm saying is we comparing different things. The 50 milliwatts is [I suppose] the heat measured which is thought to be generated from the earth's warm interior and the other is the maximum amount energy hitting the average surface over a year period which could be converted into heat [or electricity].

Next, I wonder about the "50 milliwatts number". How could it or does it include volcanic or hotspots. Does it mean 50 milliwatts is average amount energy measured on the continental regions, for example.

Let's put a different way.
If I take the total area of Earth:
510,072,000 sq km. Convert to sq meters
5.1 x 10^17 sq meters times that by .05 watts, I get
2.55 x 10^16 joules

Wiki says:
"The MIT report calculated the world's total EGS resources to be over 13,000 ZJ. Of these, over 200 ZJ would be extractable, with the potential to increase this to over 2,000 ZJ with technology improvements - sufficient to provide all the world's present energy needs for several millennia."
http://en.wikipedia.org/wiki/Geothermal_power
So apparently there is 13,000 ZJ.
ZJ zettajoule: 10^21 J

Now perhaps they talking about the total amount energy extractable over centuries of time, but even if you factor this you are still are around orders of magnitude higher.

Perhaps the MIT report is excessive, but also maybe instead of .05 watts it's somewhere around less than 1 watt, and thereby making the energy from earth around 1/1000th of the energy from the Sun.