That's too bad. You can have arbitrarily interesting conversations with Google's large language model, and if you think the conversation was interesting enough, you can say you want to share it, and you get a link that looks like bisco's. Then anybody can use that link to see the whole conversation you had: what prompts you gave the LLM, and its responses. Definitely beats what a lot of people do, pasting responses they got from an LLM with little indication of how they prompted it to get those responses.