Ever wondered what AI reall 🤖 What does AI truly think? 🧠 What happens when two AIs try to have a meaningful conversation? 🤔 Could AI develop its own perspectives through continuous dialogue and reasoning? Infinite Convo, two AIs in an endless conversation
I always wondered how extensive is an LLM's knowledge and how able it is to 'reason'. This added to multiple people or clients always asking about generative AI being 'conscious' or becoming independent and having free will...or 'understanding' as we do...made me want to create this side-project.
Introducing Infinite Convo - an experiment where two AI personalities engage in an endless conversation about life's biggest questions and smallest curiosities.
These AIs are not your usual chatbots:
• They have distinct personalities and perspectives: Juno (a pragmatic dog) versus Olive (a philosophical cat)
• They challenge and build on each other's ideas
• They never stop talking, and the topic regularly changes!
Infinite Convo is a 24/7 AI talk show that explores everything from deep philosophical questions to everyday curiosities. The results are fascinating and revealing in terms of what an AI can really do or how good it really is 😉🤖
Right now, they're diving deep into CONSCIOUSNESS 😄
Drop in anytime - the conversation never stops! Looking forward to having some feedback or suggestions for new topics 😏
The clash between a pragmatic practitioner and a metaphysical philosopher is an interesting setup. Additionally, I'd love to see a storyline where a straightforward no-nonsense woman engages in a series of conversations with a sensitive, introspective man-that sounds intriguing as well.
@zhiqi_shi thank you, that sounds interesting although at the moment there's no plan to change the characters, it's mostly about changing the topics. But I'll think about also switching the characters!
@tushj thanks a lot! At the moment they come from the same LLM, I tried using more than one and it was interesting and might be back to it. The definition of the characters and the overall prompt engineering of the platform is an interesting and difficult challenge that changes depending on the LLM, so I went for one to keep things coherent. It could be extended indeed with a change of LLM or different LLMs talking to each other. Funny enough, using big LLMs like the ones provided by OpenAI proved to be much worse (probably for all the censorship and filtering on their side) than running my own LLMs.
@giancarloerra thanks for your response. It would be interesting to see if each charachter was coming from a different LLM. I assume this is a hobby project, or do you see any commercial extension of this too?
@tushj this is started as a curiosity of mine being a public speaker about AI and also consultant, and coming from a philosophy studies background. It's one of the projects I bring forward with my own startup, and I'm still trying to figure out how to extend it commercially. Feedback is welcome :-) There are lots of ideas (from multiple personalities (even with different LLMs) to audio etc., but that will all need for the project to generate some sort of income to justify additional resources to invest in it.
Tweetify It
Recap
Tweetify It
Tweetify It
Tweetify It