Indeed. Either this was some sort of fission device (utterly implausible, IMHO - where's the moderator, for just one thing), or the "wave of heat" mentioned in the article is, shall we say, embellishment.
It sounds like the device was being transported in pieces. Each piece would be warm. Assemble it and you have far more concentrated heat. RTGs also work on temperature differential--when operating there's going to be a system to eject that heat to get a maximum temperature gradient. I can easily picture something like a fan blowing air over a radiator to maximize the power output--and there's your "wave of heat".
I take your point about it warming up after assembly, though I doubt it was fan-cooled. This was a SNAP-19C device, primarily used for Nimbus satellites, where, of course, fans could not be used. A manual is available [1], where it mentions an operating temperature of 380°F (193°C) in space, and 260°F (127°C) in the atmosphere (presumably in an environment around standard sea level temperature and pressure; the air near the top of Nanda Devi is less dense but also colder, and presumably it is often windy.) I guess it would be like standing next to an operating stove (though, of course, in mountaineering clothing.)
The manual also warns, when checking it out, against letting the temperature change too fast: no more than 35°F (19.5°C) in 15 minutes. And don't have it assembled without an electrical load, or it will overheat.
Then there's the question of whether the electronics were fan-cooled, but an RTG is so thermally inefficient that its own waste heat must dominate that dissipated by any device it powers.
Yes, this was an RTG. An RTG is not a fission device. An RTG uses the heat from the radioactive decay (alpha decay, in this case, with some contribution from the beta decay of the daughter nuclides, once they have formed) of the fuel. As khuey said above, this cannot be started or stopped.
