This message was deleted.

I was not expecting less than you doing that! 🤩

@Giulia Guizzardi and this gets two dots on the KubeCon bingo board with Generative AI too!

🤔 interesting service, chatgpt mostly hallucinates because it does not have access to new or specific information, plugins released for gpt4 go a long way to help with that :)

Thanks @Jennifer Riggins !! Hey Hugo, I’m actually with Kubiya. As you guessed, we actually don’t use GPT-4 much. For the knowledge management and the workflows with have our own LLM and vector database.

🙂 right I was curious is it one model for everyone or for each client, since how can I train a model on my individual data unless I have the model to myself?

Great question. For security purposes the vector database and the reinforcement learning is organization specific. An example I like to give is ‘VM’. Each organization might reference a different thing when referring to a VM hence we learn from within the organization. Also the docs might include proprietary information which is another reason its organization specific.

Gotcha, makes sense 🙂