AI use cases and applications
Models
The models on Amazon Bedrock
The story: Amazon Bedrock is like a food court where each stall has a specialty. You could get noodles at most stalls, but you go to a particular one for the best version.
In AI/AWS terms:
| Provider | Model | Stall specialty | Example use |
|---|---|---|---|
| AI21 Labs | Jurassic-2 | Text generation, summarization, paraphrasing, chat, information extraction | Summarize long financial documents, write product descriptions |
| Amazon | Titan | Summarization, classification, open-ended Q&A, information extraction, embeddings, search | Studio-quality ad images, real-time summaries for customer service |
| Anthropic | Claude | Content generation, translation, Q&A, summarization, code explanation and generation | Code generation and debugging, parsing legal documents |
| Stability AI | Stable Diffusion | Photorealistic images from text, improving image quality | Game characters and worlds, marketing assets |
| Cohere | Command | Text generation, information extraction, Q&A, summarization | Support chatbots, product descriptions, summarizing medical text |
| Meta | Llama | Q&A, chat, summarization, paraphrasing, sentiment analysis, text generation | Customer service chatbots |
For the exam: Most models can do general text work, so look for the standout skill in each row.
The two to remember
The story: If you want a picture, go to the stall run by the painter. If you want your company's library organized so you can find anything, go to the stall that builds the map of meanings.
In AI/AWS terms: Image generation points to Stable Diffusion (or Titan Image Generator). Embeddings for search point to Amazon Titan.
For the exam: Images from text = Stable Diffusion. Embeddings for search = Amazon Titan.