Hacker News Reader

These models are trained on sequences of words, not told the letters and left to get on with it.

It continues to amaze how people are incapable of following even the most trivial cases of abstract reasoning.