• Stop being a LURKER - join our dealer community and get involved. Sign up and start a conversation.

Reply to thread

Imagine you have a big box of toys called "BigQuery Editions". You can pick and choose different toys (features) to play with, depending on what games you want to play (workload requirements). It's like having a mix and match toy set, so you don't need to pay for toys you don't use. When you're playing and need more toys, the box magically adds more toys (compute capacity) for you, and you only pay for the toys you play with.


Now, think of "BigQuery ML" like a super helpful robot friend that lives in the toy box. This robot can help you build awesome toy structures (ML models) with different types of building blocks (structured, semi-structured, and unstructured data) using simple instructions (SQL). It's super fast and makes building your toy structures really easy. You can also take your toy structures and show them off at other playgrounds (Vertex AI or your own serving layer).