docs: align quickstart model with CPU-friendly recommendation (qwen3.5:4b) (#287)

Reported via twitter (issue #269): the macOS install guide and other
docs recommend a 4B model on CPU but the README and quickstart pull /
ask a much larger model (qwen3:8b) — confusing and slow for first-time
CPU-only users.

Switch the README quickstart and the docs/getting-started/quickstart.md
'Chat with Any Model' tile to qwen3.5:4b, with an inline note that GPU
users can scale up to qwen3.5:9b or larger. This matches the existing
chat-simple preset and the macOS guide's CPU recommendation.

Closes #269

README.md

@@ -69,13 +69,13 @@ uv run jarvis init
 # 2. Start Ollama and pull a model
 curl -fsSL https://ollama.com/install.sh | sh
 ollama serve &
-ollama pull qwen3:8b
+ollama pull qwen3.5:4b   # CPU-friendly default; use qwen3.5:9b or larger if you have a GPU
 
 # 3. Ask a question
 uv run jarvis ask "What is the capital of France?"
 ```
 
-`jarvis init` auto-detects your hardware and recommends the best engine. Run `uv run jarvis doctor` at any time to diagnose issues.
+`jarvis init` auto-detects your hardware and recommends the best engine and model size. Run `uv run jarvis doctor` at any time to diagnose issues.
 
 ## Starter Configs
 

docs/getting-started/quickstart.md

@@ -14,7 +14,7 @@ OpenJarvis is a modular AI assistant framework. Here's what developers build wit
 === "Chat with Any Model"
 
     ```bash
-    jarvis ask "Explain quantum entanglement" -m qwen3:8b
+    jarvis ask "Explain quantum entanglement" -m qwen3.5:4b   # use qwen3.5:9b or larger on GPU
     ```
 
 === "Agent + Tools"