diff --git a/src/node_builtins.cc b/src/node_builtins.cc
index cfb4b5ab031dba..e9bf8a5a0e0b7e 100644
--- a/src/node_builtins.cc
+++ b/src/node_builtins.cc
@@ -248,10 +248,10 @@ bool BuiltinLoader::Add(const char* id, std::string_view utf8source) {
   size_t expected_u16_length =
       simdutf::utf16_length_from_utf8(utf8source.data(), utf8source.length());
   auto out = std::make_shared<std::vector<uint16_t>>(expected_u16_length);
-  size_t u16_length = simdutf::convert_utf8_to_utf16le(
-      utf8source.data(),
-      utf8source.length(),
-      reinterpret_cast<char16_t*>(out->data()));
+  size_t u16_length =
+      simdutf::convert_utf8_to_utf16(utf8source.data(),
+                                     utf8source.length(),
+                                     reinterpret_cast<char16_t*>(out->data()));
   out->resize(u16_length);
   return Add(id, UnionBytes(out));
 }
diff --git a/test/cctest/test_util.cc b/test/cctest/test_util.cc
index 443a03117c09fc..5b83e07db3b7c2 100644
--- a/test/cctest/test_util.cc
+++ b/test/cctest/test_util.cc
@@ -299,17 +299,3 @@ TEST(UtilTest, SPrintF) {
   const std::string with_zero = std::string("a") + '\0' + 'b';
   EXPECT_EQ(SPrintF("%s", with_zero), with_zero);
 }
-
-TEST(UtilTest, SimdutfEndiannessDoesNotMeanEndianness) {
-  // In simdutf, "LE" does *not* refer to Little Endian, it refers
-  // to 16-byte code units that are stored using *host* endianness.
-  // This is weird and confusing naming, and so we add this assertion
-  // here to verify that this is actually the case (so that CI tells
-  // us if it changed, because for most people Little Endian is
-  // host endianness, so locally everything would work fine).
-  const char utf8source[] = "\xe7\x8c\xab";
-  char16_t u16output;
-  size_t u16len = simdutf::convert_utf8_to_utf16le(utf8source, 3, &u16output);
-  EXPECT_EQ(u16len, 1u);
-  EXPECT_EQ(u16output, 0x732B);
-}