chore: don't stop inference and hide the stop button when creating a …

…new thread (#4354)

web/hooks/useCreateNewThread.ts

@@ -21,8 +21,6 @@ import { toaster } from '@/containers/Toast'
 
 import { isLocalEngine } from '@/utils/modelEngine'
 
-import { useActiveModel } from './useActiveModel'
-
 import useRecommendedModel from './useRecommendedModel'
 import useSetActiveThread from './useSetActiveThread'
 
@@ -52,21 +50,15 @@ export const useCreateNewThread = () => {
   const [activeAssistant, setActiveAssistant] = useAtom(activeAssistantAtom)
 
   const experimentalEnabled = useAtomValue(experimentalFeatureEnabledAtom)
-  const setIsGeneratingResponse = useSetAtom(isGeneratingResponseAtom)
 
   const threads = useAtomValue(threadsAtom)
-  const { stopInference } = useActiveModel()
 
   const { recommendedModel } = useRecommendedModel()
 
   const requestCreateNewThread = async (
     assistant: (ThreadAssistantInfo & { id: string; name: string }) | Assistant,
     model?: Model | undefined
   ) => {
-    // Stop generating if any
-    setIsGeneratingResponse(false)
-    stopInference()
-
     const defaultModel = model || recommendedModel
 
     if (!model) {