Cape Station, future home of an enhanced geothermal power plant, in Utah

Pipes, steam, turbines...

We need better peltier devices.

I thought exactly the same thing. Peltier's themselves don't really have a pathway for becoming as efficient as steam turbines but there are other methods currently in research. One promising idea is to use sodium vapor in a fuel cell style device: http://dx.doi.org/10.1016/j.jpowsour.2017.10.022

The maximum COP for a peltier device is 1, in practice it’s far below that.

Heat pumps go way beyond a COP of 1; an open-loop cooling system with an evaporative cooling tower can have a COP of 7. A closed loop heat pump alone can have a COP of 4.

Peltier devices are a dead end for moving heat around outside of specialized applications where you can’t drag around two heat exchangers, a valve, and a pump (like active cooling clothing). It’s impossible for them to even approach the efficiency of resistive heating (COP of 1).

The aim here is to get power from temperature differential, not vice-versa