So what you need to do is take each of these talkers and load their transcripts into a GPT3 model and then let people "talk" with that model. Call it GPP(tm) a

So what you need to do is take each of these talkers and load their transcripts into a GPT3 model and then let people "talk" with that model. Call it GPP(tm) and see if anyone gets the reference. For giggles, expand the context length to 4096 tokens and it will even have a better "memory." Now store that context for each ongoing conversation. CH could soon stand for Chatbot Heaven. https://www.sfchronicle.com/projects/2021/jessica-simulation-artificial-intelligence/ https://www.sfchronicle.com/projects/2021/jessica-simulation... I think the appeal of clubhouse was its exclusivity and that we were all in lockdown craving contact and its time of release.That's mostly over now despite the Delta variant taking this into extra innings so who needs this thing? It's just another social network now.