| ▲ | alyxya 2 hours ago | |
The first point is along the lines of what I'd expect given that claude code is generally reliable at this point. A model's raw intelligence doesn't seem as important right now compared to being able to support arbitrary length context. | ||