diff --git a/backend/apps/agents/manager/predict_prompts.py b/backend/apps/agents/manager/predict_prompts.py
index f41ebad8..5eb73da6 100644
--- a/backend/apps/agents/manager/predict_prompts.py
+++ b/backend/apps/agents/manager/predict_prompts.py
@@ -17,6 +17,10 @@ logger = logging.getLogger(__name__)
 
 MAX_TOPICS = 24
 MAX_SUGGESTIONS = 5
+# Don't predict someone's next prompt until we ACTUALLY know their patterns. Below this many real
+# past chats, any guess is just noise (onboarding starters alone are what they browsed at setup, not
+# a read on what they want now), so we stay silent and let the neutral placeholder stand.
+MIN_REAL_TOPICS = 4
 # Names the aux title-gen hands out for empty/greeting chats; they carry no topic signal.
 P_SKIP_NAMES = {"untitled", "new chat", "greeting", "chat", ""}
 
@@ -70,8 +74,9 @@ async def predict_prompts(count: int = MAX_SUGGESTIONS) -> List[str]:
             for s in (global_settings.personalized_starters or [])
             if getattr(s, "prompt", None)
         ]
-        # Nothing to personalize from: let the composer keep its static placeholder.
-        if not topics and not starters:
+        # Only predict once there's a real track record. Onboarding starters can enrich a prediction
+        # but never trigger one on their own: a brand-new user hasn't shown us what they want yet.
+        if len(topics) < MIN_REAL_TOPICS:
             return []
 
         aux_model = (await resolve_aux_model(global_settings, preferred_tier="haiku"))[0]
diff --git a/frontend/src/app/pages/Dashboard/DashboardToolbar.tsx b/frontend/src/app/pages/Dashboard/DashboardToolbar.tsx
index 179a0848..c3c438dd 100644
--- a/frontend/src/app/pages/Dashboard/DashboardToolbar.tsx
+++ b/frontend/src/app/pages/Dashboard/DashboardToolbar.tsx
@@ -99,7 +99,6 @@ const DashboardToolbar = React.forwardRef<HTMLDivElement, Props>(
     // load (cached), then one is shown at a time and cycled while the composer sits idle+empty. Empty
     // list (no signal / no provider / error) just leaves the static "What should I do sir..." placeholder.
     const [ghostList, setGhostList] = useState<string[]>([]);
-    const [ghostIdx, setGhostIdx] = useState(0);
     const ghostFetchedRef = useRef(false);
     useEffect(() => {
       if (!inputOpen || ghostFetchedRef.current) return;
@@ -116,14 +115,10 @@ const DashboardToolbar = React.forwardRef<HTMLDivElement, Props>(
         } catch { /* fail open: keep the static placeholder */ }
       })();
     }, [inputOpen]);
-    // Rotate the visible suggestion every few seconds while the composer is open, so the user sees a
-    // few different ideas instead of one. Cheap; the list is already fetched and cached.
-    useEffect(() => {
-      if (!inputOpen || ghostList.length <= 1) return undefined;
-      const t = setInterval(() => setGhostIdx((i) => (i + 1) % ghostList.length), 4500);
-      return () => clearInterval(t);
-    }, [inputOpen, ghostList.length]);
-    const ghostSuggestion = ghostList.length ? ghostList[ghostIdx % ghostList.length] : undefined;
+    // ONE stable suggestion, never a rotating carousel: cycling guesses every few seconds reads as
+    // "the app is throwing darts." The backend only returns anything when it has real usage history to
+    // predict from (see predict_prompts.py), so an empty list just leaves the neutral placeholder.
+    const ghostSuggestion = ghostList.length ? ghostList[0] : undefined;
 
     // Reset defaults on each new compose session so in-session picks don't leak into the next new-chat draft.
     const prevInputOpen = useRef(false);