[BUG]: cuda::std::variant fails to compile w/ clang & libc++

### Is this a duplicate?

- [x] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cccl/issues) for this bug and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md)

### Type of Bug

Compile-time Error

### Component

libcu++

### Describe the bug

emplace_index_init_list_args.pass.cpp fails to compile w/ clang & libc++

It's not clear if the issue is with the test or the way libcu++ interacts w/ the standard library.

### How to Reproduce

Reduced reproducer based on the emplace_index_init_list_args.pass.cpp test: https://godbolt.org/z/Yqf66GjEP

Compiles OK with NVCC, and w/ clang and libstdc++. Fails with ambiguous overload resolution with libc++.

### Expected behavior

cuda::std::variant should be compileable with both libc++ and libstdc++.

### Reproduction link

https://godbolt.org/z/Yqf66GjEP

### Operating System

_No response_

### nvidia-smi output

_No response_

### NVCC version

_No response_