The Transformer Algorithm Is Not as Complex as Many People Think

The world that we are in is learnable largely because of information preservation. The super molecule of the cell is extremely more complicated than an organic molecule and that's because it would not be able to stay stable if it wouldn't anticipate the future to a small degree. A continuous world is inherently easier to learn than a static world which consists of disconnected frames but this requires our systems being real time so they can deal with data that comes in in real time.