Hermes model debate signals limits for Mac mini servers
In r/, the choice of which model to run with Hermes has become the main debate. In the past week, 78 threads and more than 1,350 comments focused on model choice, which was more than the discussions about memory, security, deployment, and s combined.
The debate has split into four camps. Each camp sees the others as wasting either money or time.
One camp favors local LLMs, meaning models that run on your own machine so data stays at home. A setup with an with 16GB memory and an with 12GB helped fuel the argument that may be useful for light jobs like gathering morning news, but less useful for heavier Hermes work.
Key points
- Hermes model choice produced 78 threads and over 1,350 comments in one week.
- Model choice got more discussion than memory, security, deployment, and s combined.
- The debate has split into four camps with sharply different views on cost and time.
- The local LLM camp cares about keeping data on personal hardware.
- An with 16GB memory may fit light automation, but heavier AI work may need lower expectations or extra compute.