Talk about LLMs is everywhere, and if you listen to the loudest voices, you might think that you always need lots of memory ...
Thanks to some llama.cpp wizardry, Gemma 4 runs well on my outdated phone ...