fix(tts): play Opus chunks immediately as they arrive instead of buffering until speaking_end

This makes the voice start playing the first words while the rest of the response is still generating (big win for perceived latency).
Per PLAN item 2.

frontend/src/hooks/useConversation.ts

@@ -141,7 +141,7 @@ export function useConversation() {
             combined.set(chunk, offset);
             offset += chunk.length;
           }
-          audioBufferRef.current = [];
+          // audioBufferRef no longer used for playback (incremental)
           const blob = new Blob([combined], { type: 'audio/ogg' });
           const url = URL.createObjectURL(blob);
           audioRef.current.src = url;