They are seeking to address the fundamental issue that you cannot easily share models in a single file. GGUF seeks to contain “all the information needed to load a model.” While at the early stages, llama.cpp had breaking changes (like this development), the development was to make sure that “new features can be added to GGML without braking compatibility,” ensure there aren’t “more breaking changes,” adds support for non-llama models, and automatic prompt formats.
This is both a QoL improvement for users of llama.cpp, but also streamlining the repo for its developers.