Fix LLM provider selection in podcast handlers

- Remove hardcoded preferred_provider=huggingface in podcast handlers
- Set preferred_provider=None to respect GPT_PROVIDER env var
- Change default model from Qwen to gpt-oss-120b:cerebras (the model user had access to)
- WaveSpeed will now use gpt-oss-120b model instead of Qwen

backend/api/podcast/handlers/analysis.py

@@ -80,7 +80,7 @@ Return JSON with:
             prompt=prompt,
             user_id=user_id,
             json_struct=None,
-            preferred_provider="huggingface",
+            preferred_provider=None,
             flow_type="premium_tool",
         )
         
@@ -279,7 +279,7 @@ Requirements:
             prompt=prompt,
             user_id=user_id,
             json_struct=None,
-            preferred_provider="huggingface",
+            preferred_provider=None,
             flow_type="premium_tool",
         )
     except HTTPException: