Yup that's about right - the tool is LLMs generating infra code or docker code as needed as well as a mechanic that feeds cluster usage/consumption to try to sc

Yup that's about right - the tool is LLMs generating infra code or docker code as needed as well as a mechanic that feeds cluster usage/consumption to try to scale things up and down at the pod & node level. We have humans in the loop to generate more and more labeled data for retraining. Wish I could open it up for trial/general use but we're a bit early for us to open up self-serve