diff --git a/pkg/tui/views/chat/update.go b/pkg/tui/views/chat/update.go
index 9b58a3e..be314dc 100644
--- a/pkg/tui/views/chat/update.go
+++ b/pkg/tui/views/chat/update.go
@@ -93,7 +93,7 @@ func (m Model) Update(msg tea.Msg) (Model, tea.Cmd) {
 			// append chunk to existing message
 			m.setMessageContents(last, m.messages[last].Content+msg.Content)
 		} else {
-			// use chunk in new message
+			// use chunk in a new message
 			m.addMessage(api.Message{
 				Role:    api.MessageRoleAssistant,
 				Content: msg.Content,
@@ -177,6 +177,8 @@ func (m Model) Update(msg tea.Msg) (Model, tea.Cmd) {
 	case cursor.BlinkMsg:
 		if m.state == pendingResponse {
 			// ensure we show the updated "wait for response" cursor blink state
+			last := len(m.messages)-1
+			m.messageCache[last] = m.renderMessage(last)
 			m.updateContent()
 		}
 	case msgConversationPersisted:
diff --git a/pkg/tui/views/chat/view.go b/pkg/tui/views/chat/view.go
index 40a1463..30ca344 100644
--- a/pkg/tui/views/chat/view.go
+++ b/pkg/tui/views/chat/view.go
@@ -124,6 +124,9 @@ func (m *Model) renderMessageHeading(i int, message *api.Message) string {
 	return messageHeadingStyle.Render(prefix + user + suffix)
 }
 
+// renderMessages renders the message at the given index as it should be shown
+// *at this moment* - we render differently depending on the current application
+// state (window size, etc, etc).
 func (m *Model) renderMessage(i int) string {
 	msg := &m.messages[i]
 
@@ -138,8 +141,11 @@ func (m *Model) renderMessage(i int) string {
 		}
 	}
 
-	// Show the assistant's cursor
-	if m.state == pendingResponse && i == len(m.messages)-1 && msg.Role == api.MessageRoleAssistant {
+	isLast := i == len(m.messages)-1
+	isAssistant := msg.Role == api.MessageRoleAssistant
+
+	if m.state == pendingResponse && isLast && isAssistant {
+		// Show the assistant's cursor
 		sb.WriteString(m.replyCursor.View())
 	}
 
@@ -236,22 +242,14 @@ func (m *Model) conversationMessagesView() string {
 		sb.WriteString("\n")
 		lineCnt += lipgloss.Height(heading)
 
-		var rendered string
-		if m.state == pendingResponse && i == len(m.messages)-1 {
-			// do a direct render of final (assistant) message to handle the
-			// assistant cursor blink
-			rendered = m.renderMessage(i)
-		} else {
-			rendered = m.messageCache[i]
-		}
-
+		rendered := m.messageCache[i]
 		sb.WriteString(rendered)
 		sb.WriteString("\n")
 		lineCnt += lipgloss.Height(rendered)
 	}
 
 	// Render a placeholder for the incoming assistant reply
-	if m.state == pendingResponse && (len(m.messages) == 0 || m.messages[len(m.messages)-1].Role != api.MessageRoleAssistant) {
+	if m.state == pendingResponse && m.messages[len(m.messages)-1].Role != api.MessageRoleAssistant {
 		heading := m.renderMessageHeading(-1, &api.Message{
 			Role: api.MessageRoleAssistant,
 		})