[om8]

llama.cpp is a mess and ollama is right to move on from it

[buyucu]

ollama is still using llama.cpp. they are just denying that they are :)