Generative AI
Generative AI is any AI system that produces new content — text, images, audio, video, code — rather than classifying or predicting from fixed options.
Generative AI is the broad term for AI systems that produce new content rather than predicting a label or score for existing content. It's the family of techniques behind ChatGPT (text), Midjourney (images), Sora (video), Suno (music), and Cursor (code).
The split with "discriminative" AI matters. Discriminative AI says "is this email spam?" (classification) or "what house price does this listing predict?" (regression). Generative AI says "write me an email," "generate an image of a house," "compose a song." Generative AI is harder, more open-ended, and produces outputs that don't have a single correct answer.
The current generative AI landscape, roughly:
- Text — dominated by LLMs (Claude, GPT, Gemini, Llama).
- Images — diffusion models (Midjourney, Stable Diffusion, FLUX, DALL-E 3, Imagen) plus emerging native image LLMs.
- Video — Sora, Veo, Runway, Pika. Quality is improving fast but still expensive and short-form.
- Audio — ElevenLabs (voice), Suno + Udio (music), Whisper (speech-to-text, technically discriminative).
- Code — Claude Code, Cursor, Copilot, v0. Built on LLMs with code-specific training and tooling.
- 3D — early-stage but progressing: text-to-3D for asset generation, real-time scene synthesis.
The economic disruption is concentrated in industries where "skilled human creates artifact" has been the model: marketing, design, writing, coding, video production, music. The disruption isn't "AI replaces humans" — it's "the skill ceiling rises and the median worker is now responsible for more output, with AI as a force multiplier."
Related on ToolMango
Related terms
- LLM (Large Language Model) — A Large Language Model is a neural network trained on huge volumes of text to predict the next token, which produces emergent capabilities like reasoning, code generation, and translation.
- AI agent — An AI agent is an LLM-driven system that takes actions in the world — calling tools, browsing, writing code, finishing tasks — instead of just answering questions.
- Transformer architecture — The transformer is the neural network architecture introduced in 2017 that powers every major LLM — built around the attention mechanism that lets each token weigh all other tokens.
Want to actually build with this?
Our Stack Builder picks the best AI tools for your specific project in under 60 seconds.
Build my stack →