To me, its interesting to see the chatbot's responses when I ask them about how they might resolve the 'simultaneous reveal' issue. For example, Llama doesn't even get the rules of RPS right. When confronted about the concept of a simultaneous reveal, its responses were little more than 'I don't know, lets play again'. No big surprises there.
When I did the same thing with ChatGPT (3.5), I got a sensible discussion about the complexities of simultaneous reveal, and what it would mean to allow a chatbot to process and interact with humans independently of a textual input vehicle.
I'd be keen to see how ChatGPT 4 handles it.