It's really amazing...
MacBook M5 base 32G + Worth 3060 infrastructure.
Since my company does a completely different job, most of the time I'm in a mobile situation, so I installed Hermes+Ollam+LLM+Telegram on both sides and
After searching for a suitable model here and there, it seems that QWEN3.6-35B-A3B is the limit for MacBook and GEMMA4:LATEST for Worth...
But the difference between these two is really, really huge.
I wonder if I'm writing the prompt wrong, but even though I'm following Gemini's guide, Gemma avoids it and shirks logical laziness, so it takes a long time to correct this.
It seems too difficult to take one step further like this... I feel bad for spending a lot of money on a MacBook M5 when I have to use the cloud API...
I'm in a dilemma. ㅜ.ㅜ
