3 ms·

I think the real question isn't whether they think like humans, but whether their "discussions" lead to consistent improvement in how they accomplish tasks