Generative AI
AI that makes new stuff — text, images, code, audio — rather than just sorting or scoring what already exists.
Think of it like
The difference between a critic who rates paintings and an artist who paints new ones.
Example
Asking a model to write a poem or generate a product mockup is generative AI; asking it to label a photo “cat” is not.
How it actually works
Generative models learn the shape of their training data well enough to sample fresh examples from it. For language, that’s predicting the next token over and over. The output is novel in wording but bounded by what the training distribution taught it.
For product teams
It opened AI to creative and drafting work, which is why it went mainstream so fast.
For engineers
Models that learn a data distribution and sample from it, e.g. autoregressive LLMs or diffusion image models.
Related
- LLM — Language versions are LLMs.
- Next-token prediction — Works by next-token prediction.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome