It's not 'magic' though but previously LLMs have performed very badly on longer multiplication, 'insight' is the wrong word but I'm saying maybe they're not wil

It's not 'magic' though but previously LLMs have performed very badly on longer multiplication, 'insight' is the wrong word but I'm saying maybe they're not wildly better at this calculation... maybe they are just optimising these well known jagged edges.