When looking at the idea of pellet extrusion, it sounds like an even better idea than filaments.
Why? filaments are basically plastic extrusion with extra steps. But pellet extruders skip that step.
Also, the science and engineering of pellet extrusion is deeply established and known.
Then why do pellet extruders, at least at desktop FDM scale, suck so bad?

In short, philosophically :

1. Industrial technology, miniaturized without context

2. High power appliance, engineered like a toy

3. Marrying the knowledge of FDM, ignoring polymer physics which leads to

4. Scarce public knowledge

Metafuse were not exempt from this. In fact :
1. Up to V5, we did not know rigidity is king (aluminium was good for filament hotends, not 20Nm torque)
2. Single-flight screws have inherent eccentricity. It literally generates screw bending cantilever forces.
3. We avoided heat treatment for simplicity.
4. Wirings were ‘jumble it up and hide it somewhere’
5. Gravity is not the best pellet feeding method
6. We thought everyone knew what a Z-Probe was, so consequently,
7. We didn’t explicitly say “most 3d printers” means “not Bambu” which came with headaches the users never even know existed
8. We underestimated the advancements in overall 3d printing and added features late and lastly,
9. We spent time and money undoing AI-generated wisdom (we used all of them)
10. A fixed barrel is better for maintenance. Removing a hot barrel is more complicated than removing a hot screw.
11. Most of the technology in Materium were developed to approach filament simplicity. Most users never asked for this.

These are issues that would otherwise never be discovered if we hesitate about replacing filament extruders.
Even if we never get there, this wiki should explain why for future readers.