In rust, hashsets are just hashmaps with a zero-sized type as the value, which results in all the extra getting optimized away.

This is true for Java as well, a `HashSet` is just a `HashMap<T, Object>` where the value is:

  // Dummy value to associate with an Object in the backing Map
  private static final Object PRESENT = new Object();

https://github.com/openjdk/jdk/blob/master/src/java.base/sha...

A thread-safe implementation is available with `ConcurrentHashMap.newKeySet()`, which uses a similar approach:

  public static <K> KeySetView<K,Boolean> newKeySet() {
    return new KeySetView<K,Boolean>
      (new ConcurrentHashMap<K,Boolean>(), Boolean.TRUE);
  }

https://github.com/openjdk/jdk/blob/master/src/java.base/sha...

There probably aren't many languages where that's different, whether a hashset on a hashmap or a treeset on a treemap.

However sets tend to provide support for… set operations. Which currently can't be done generically in go except through interface{} or specialised, neither of which is a great option.

Even if that's the case, the thing we want is syntactic sugar. It'd be nice to just do `sets.New[int]()` rather than manipulate maps manually or create bespoke types.