Position: Theory of Mind Benchmarks are Broken for Large Language Models | Read Paper on Bytez