As a short, but helpful, side discussion consider the breeding of Pu-239 in a reactor for use in nuclear weapons. Pu-239 works great in a nuclear bomb but other isotopes of Plutonium cause problems, specifically the high spontaneous fission rate of Pu-240 can create so many background neutrons that having a significant amount of it in a bomb can vastly increase the "pre-detonation" (fizzle) risk, which causes the bomb to have very low yield (on the same scale as the chemical explosives in the bomb). As Plutonium remains in a reactor subject to neutron flux it will naturally breed other isotopes, which is why there is a difference between weapons grade and reactor grade Plutonium. Reactor grade Plutonium can contain any variation of isotopes, essentially all of them are suitable for use in a reactor, but weapons grade Plutonium can only have a very small amount of Pu-240. And for this reason the way that weapons grade Plutonium is made is by removing fuel rods from a reactor on very short time scales and reprocessing it to remove the Plutonium then forming it back into new fuel rods and so forth. This is a very costly and complex endeavor which is why you can't just use ordinary power reactors for generating weapons grade Plutonium.
Interestingly, Pu-238 also exists in reactor grade Plutonium in substantial amounts (1% or more of the Plutonium, so nearly an entire tonne per year worldwide). However, because it's mixed up with all of the other Plutonium it would require extremely costly isotopic separation.
So, how is this relevant to Thorium reactors? Well, let's go back and look at what actually happens when Pu-238 is produced in a Thorium reactor. You start with Th-232, and under neutron flux you'll breed U-233, which is the reactor's main fission fuel. If U-233 is hit by a neutron and absorbs it instead of fissioning then you can end up with U-234, U-235, U-236, Np-237 (through a decay), and Pu-238 (through another decay). However, it's not as though things don't stop there. Pu-238 will breed into Pu-239, Pu-240, and so on, just as in a conventional Uranium power reactor.
So here you have the same problem as producing weapons grade Plutonium, you have a process that you need to stop before it goes too far, and in order to do that you need to pull the Plutonium out of the reactor at regular, short intervals. Otherwise you'll just get a buildup of ordinary reactor grade Plutonium. However, the problem is worse here because instead of being the product of just one neutron reaction (natural U-238 becoming Pu-239) the breeding of RTG grade Plutonium is the product of a long chain of reactions (taking 6 steps between the isotopes in the reactor fuel at the start and the production of Pu-238). This means the amount of production and the time scales of production are very much not helpful from the perspective of pure Pu-238 production, especially if you want to also operate a power reactor cost-effectively at the same time.
(Edit: also, there's an additional problem, because Thorium reactor fuel contains U-232, which has a short half-life and has decay products that are prodigious gamma ray sources, making reprocessing and handling even more difficult than ordinary reactor fuel. You can handle Plutonium in a glove box, but with used Thorium fuel containing U-232 you'd need to handle it via robotic manipulators in a heavily shielded area distant from humans, except that gamma radiation kills electronics like nobody's business, which is a bit of a catch-22. This is one of the key technical hurdles of Thorium reactor designs in general.)
In short, the fact that Thorium reactors produce Pu-238 isn't helpful, because existing reactors do too, but in either case it's hard to get at.
Contamination with plutonium-240 was a big problem during the Manhattan Project. They'd done the preliminary calculations based on cyclotron-produced plutonium-239, which was pure enough that a gun design with a long tube was feasible. ("Thin Man")
The plutonium-239 that came out of the Hanford reactor had a much higher contamination rate. This required a much faster assembly of the critical mass in order to avoid a fizzle. The gun design would have to be so long that it couldn't fit on a plane. Whereas implosion could provide the required rapid assembly in a limited space. Thus, "Thin Man" was out, and "Fat Man" was in. Implosion was no longer a secondary option -- it was the only option.
This isn't specific to thorium, but could liquid fueled reactors allow for chemical separation of Pu-238, as it's the first Pu isotope in the chain?
Obviously the ability to do something like that is still far off; I'm not actually sure how we previously made usable Pu-238. Is it isotopic separation from breeder reactors?
