diff --git a/front_end/src/app/(main)/questions/components/notebook_form.tsx b/front_end/src/app/(main)/questions/components/notebook_form.tsx
index 7f5082341..fa523d1a7 100644
--- a/front_end/src/app/(main)/questions/components/notebook_form.tsx
+++ b/front_end/src/app/(main)/questions/components/notebook_form.tsx
@@ -125,8 +125,8 @@ const NotebookForm: React.FC<Props> = ({
       categories: categoriesList.map((x) => x.id),
       notebook: {
         markdown: data["markdown"],
-        type: news_category_id ? "news" : "discussion",
-        image_url: null,
+        type:
+          post?.notebook?.type ?? (news_category_id ? "news" : "discussion"),
       },
     };
 
diff --git a/posts/services/search.py b/posts/services/search.py
index 00600fcbb..79f12417c 100644
--- a/posts/services/search.py
+++ b/posts/services/search.py
@@ -2,6 +2,7 @@
 import logging
 
 import numpy as np
+from asgiref.sync import async_to_sync
 from django.contrib.postgres.search import SearchVector, SearchQuery
 from django.db.models import Value, Case, When, FloatField, QuerySet
 from pgvector.django import CosineDistance
@@ -66,9 +67,9 @@ def update_post_search_embedding_vector(post: Post):
 
 
 def perform_post_search(qs, search_text: str):
-    embedding_vector, semantic_scores_by_id = asyncio.run(
-        gather_search_results(search_text)
-    )
+    embedding_vector, semantic_scores_by_id = async_to_sync(
+        gather_search_results
+    )(search_text)
     semantic_scores_by_id = semantic_scores_by_id or {}
 
     semantic_whens = [
diff --git a/utils/openai.py b/utils/openai.py
index 618e4e974..6a9ce3a21 100644
--- a/utils/openai.py
+++ b/utils/openai.py
@@ -52,9 +52,10 @@ def generate_text_embed_vector(text: str) -> list[float]:
 
 
 async def generate_text_embed_vector_async(text: str) -> list[float]:
-    response = await get_openai_client_async().embeddings.create(
-        input=text, model=EMBEDDING_MODEL
-    )
+    async with get_openai_client_async() as client:
+        response = await client.embeddings.create(
+            input=text, model=EMBEDDING_MODEL
+        )
     vector = response.data[0].embedding
 
     return vector