Google Cloud's intro to generative AI is the best starting point for explaining AI to people who come from a traditional software or data background

The specific distinction that lands for technical people: traditional machine learning is discriminative, it classifies or predicts from existing patterns. Generative AI is generative, it produces new outputs from learned distributions. That framing makes the shift from classic ML and data science to generative AI conceptually clear rather than requiring you to start from scratch.

The comparison between classic ML pipelines and generative AI workflows is the translation layer for technical teams evaluating where generative AI fits into their existing systems versus where it replaces them versus where it does neither.

For data engineers, ML engineers and analysts in the community: what is the most significant workflow change generative AI has introduced to your work and has it replaced, augmented or added to what you were already doing?