Sincere apologies. I really CBA tearing the whole thing apart, but for starters, let's consider a hypothetical 100 W battery powered by Co-60. Let's be generous and suppose that there's a way to capture all of the radiation emitted using a tiny device and to convert it to electrical energy. I'll omit the electron anti-neutrino emissions as even the original writer would (probably) accept that intercepting these might be tricky.
100 W requires at least 240 TBq of Co-60. Co-60 is a high energy gamma-emitter, which is one reason why it's not used in RTGs. It's difficult to shield. Another good reason is it's half-life of 5.27 years, which would mean having to replace the battery rather regularly.
To get the external radiation down to a level generally accepted as safe would require arounda 55 cm radius of steel shielding, even if the electrical conversion device were miniscule. A 110 cm steel sphere weighs a bit over 5 T. A realistic device which occupies a finite volume would weigh more than the boat but might be useful as an anchor.
I could go on, but won't...