Jun 15

1. 17:16

   Reduce reader brain Modal memory use⁠

   cvpfus +1: Codex
   718f2c4
2. 17:05

   Fix Modal reader brain startup⁠

   cvpfus +1: Codex
   a988b59
3. 16:53

   Clear llama.cpp image entrypoint for Modal⁠

   cvpfus +1: Codex
   1502a42
4. 16:34

   Use prebuilt llama.cpp Modal image⁠

   cvpfus +1: Codex
   f158a05
5. 16:24

   Add CPU Spaces deployment with Modal runtimes⁠

   cvpfus +1: Codex
   41a3f6b