Grounding for Gemini with Vertex AI Search and DIY RAG

Learn about grounding with generative AI and discover how to integrate Gemini, multimodal embeddings, and vector search to build a production-ready RAG system.
Github → goo.gle/4b6pcN7
NOTE: Cloud Credits discussed in this session or workshop were for live audiences only
Speaker: Holt Skinner
Watch more:
Check out all the AI videos at Google I/O 2024 → goo.gle/io24-ai-yt
Check out all the Cloud videos at Google I/O 2024 → goo.gle/io24-cloud-yt
Subscribe to Google Cloud Tech → goo.gle/GoogleCloudTech
#GoogleIO
Products Mentioned: Vertex AI
Event: Google I/O 2024

[@attilavass6935]

I'm sad because:
- we can not have public websites crawled regularly in our Chat and Agent type Apps, only in Search type...
- Gemini 1.5 (Pro) is not available for the apps (EU)

[@needpeace5214]

The advanced option allows you to use Flash, but if you customize through agents, it's not available?
And in general, is the advanced variant more efficient because of different frameworks and, for example, the ability to search through images, which is not available in agents?

[@Optimusjf]

Exclent content

[@MrStupiuno]

This is super exciting, but it’s hard to get Gemini these days to reply to anything reliably over time without it saying “As a language model I can’t help you with that”. Even when it’s a simple question that the competitors answer easily.