A Reddit user asks whether purchasing an RTX 5060 Ti is worthwhile for additional Large Language Model workloads, given their existing setup with an RTX 4090 and Intel i9 13900K.
The poster primarily uses their GPU for training Small Language Models and llama.cpp inference. They note that the 5060's memory bandwidth is significantly lower than the 4090's and question if the extra 16GB of VRAM justifies the cost and effort.