diff --git a/src/tensors/indexmanipulations.jl b/src/tensors/indexmanipulations.jl
index 19208fdd..7e80f938 100644
--- a/src/tensors/indexmanipulations.jl
+++ b/src/tensors/indexmanipulations.jl
@@ -326,18 +326,32 @@ function _add_general_kernel!(tdst, tsrc, p, fusiontreetransform, α, β, backen
     elseif β != 1
         tdst = scale!(tdst, β)
     end
-    if Threads.nthreads() > 1
-        Threads.@sync for s₁ in sectors(codomain(tsrc)), s₂ in sectors(domain(tsrc))
-            _add_sector!(tdst, tsrc, fusiontreemap, s₁, s₂, α, β, backend...)
-        end
-    else
-        for (f₁, f₂) in fusiontrees(tsrc)
-            for ((f₁′, f₂′), coeff) in fusiontreetransform(f₁, f₂)
-                TO.tensoradd!(tdst[f₁′, f₂′], p, tsrc[f₁, f₂], :N, α * coeff, true,
-                              backend...)
-            end
+    
+    # TODO: implement multithreading for general symmetries
+    # Currently disable multithreading for general symmetries, requires more testing and
+    # possibly a different approach. Ideally, we'd loop over output blocks in parallel, to
+    # avoid parallel writing, but this requires the inverse of the fusiontreetransform.
+    
+    for (f₁, f₂) in fusiontrees(tsrc)
+        for ((f₁′, f₂′), coeff) in fusiontreetransform(f₁, f₂)
+            TO.tensoradd!(tdst[f₁′, f₂′], p, tsrc[f₁, f₂], :N, α * coeff, true,
+                          backend...)
         end
     end
+    
+    # if Threads.nthreads() > 1
+    #     Threads.@sync for s₁ in sectors(codomain(tsrc)), s₂ in sectors(domain(tsrc))
+    #         _add_sectors!(tdst, tsrc, fusiontreemap, s₁, s₂, α, β, backend...)
+    #     end
+    # else
+    #     for (f₁, f₂) in fusiontrees(tsrc)
+    #         for ((f₁′, f₂′), coeff) in fusiontreetransform(f₁, f₂)
+    #             TO.tensoradd!(tdst[f₁′, f₂′], p, tsrc[f₁, f₂], :N, α * coeff, true,
+    #                           backend...)
+    #         end
+    #     end
+    # end
+    
     return nothing
 end