[ars]

"transform some of the energy from radioactive decay"

Not so simple. You need a hot source and a cold sink to transform energy. Where's your cold sink? This thing is intended to melt what's around it, and the outside of the probe is not that different in temperature from the inside.

[gipp]

Erm. If it's melting the rock around it, then the rock is your cold sink. By the time the temperature of the rock catches up with the probe, it's already molten anyway.

[dredmorbius]

More specifically: the interior of the probe is 2000F, the surrounding rock melts at 1000F. That's a pretty good temperature gradient right there.

[ars]

The rock would not work as a cold sink - the cold rock is too far from the probe. The rock right near the probe is at almost the exact same temperature.

[kijin]

And that's exactly why the probe will take 30 years to move 100km. It's going to be a very slow process.

[cperciva]

You need a hot source and a cold sink to transform energy.

Only if it's thermal energy. Given the small amount of power needed, they could easily use a betavoltaic generator.

[InclinedPlane]

Co-60 decays via beta (and then gamma) decay. It's a straightforward matter to convert the beta decay into electrical power.

[ars]

Is it really that simple? You can't have any active electronics - they would melt.

And the tungsten is touching the cobalt, with no air gap (i.e. to opportunity to harvest power from the electrons returning to the cobalt to neutralize charge).

How would you harvest the electrons?

[InclinedPlane]

Of course it's not simple, it's a very challenging project. But in terms of providing electrical power, if it's possible to have electronics that work at all then it would be possible to power them via beta decay. This is already an established technology in the form of "betavoltaics", which have powered existing devices (such as pace makers) in the past.

http://en.wikipedia.org/wiki/Betavoltaics