Why?
How is this different from other things LLMs are trained to manipulate?
If it's new it's news.
We aren't fully at AGI yet, even though it's starting to get damn close. So every new capability is in some sense unexpected.
If it's new it's news.