This is news to me. Not doubting you - curious. Can you show me some links on that? I suspect so they can iterate quickly. They're complex shapes; more like exhaust vents than sinks and they're curved in two directions. They'd be problematic to cast. I agree with you largely. but turns out the powder thats was near the laser gets affected by the heat and if you pour it back in you get all sorts of additional material oddities in the next part.
Idk why you would make 3D printed heat sinks... probably because you have too much money but ok it works fine in any application where you dont need structural material properties.
Here is a decent one talking about the decreased tensile strength of products made from recycled polyamides. Though I'd like to hear HGL's input as well. Going to a different setting and hoping it's a good analogy, It seems to make sense from the organic synthesis standpoint. Once you allow it to happen, runaway polymerization is one of the easiest ways to obliterate your product (e.g. decarboxylation of salicylic acid to phenol). All you need is heat and impurities that act akin to nucleation sites for growing crystals.
That paper is about polymers, though, and the whole fun of SLS is powdered metal. I have no doubts that getting a polymer melty-adjacent makes the remelt janky. If you're talking about steel, aluminum or titanium, though, you're at basic-bitch cast level anyway.