Each block of flash can be written only a limited number of times, so flash drives (SSDs, cards, USB sticks) all have more blocks than are visible as part of the disk. Drives internally rotate active blocks in and out of the spare pool to try to keep the number of writes to each similar ('wear levelling'). When you write to a flash drive — including trying to overwrite data to destroy it as someone might on a magnetic disk — it will generally pull a block from the spare pool for the new data, and put the old block in the spare pool.

The spare pool is invisible to the OS, but it is reasonable to assume that there are ‘secret’ commands to access it — not because some TLA demands it, but because the hardware/firmware engineers need it for development and debugging.

BUT there is a great big BUT. Writing flash is a two-step process. Programming flash can only change a 1 bit to a 0. Before this, there has to be a slower erase step, that sets the block to all 1s. In order to avoid this performance-killing overhead on every write, flash drives erase as much as possible (whether spare pool blocks or TRIMmed visible blocks) in the background as soon as possible.