solar.codes
Boutique Software & Hardware
Intentional AI
Blog
-
Using LM Studio
Browse, load, slide GPU to max, serve on 1234.
-
Using Ollama
Pull. Run. Point your tools at localhost.
-
On art and less
Remove until it nears.
-
Local inference — tips
Small adjustments. Large quiet gains.
-
Claude Code will eat Claude
When the tool surpasses the teacher, what remains.
-
Choosing a quantized model
Fit first. Quality second. One rung at a time.
-
Family office, local AI
Wealth stays home. So should judgment.