# @salesforce on Twitter

- **Type:** Video
- **Original URL:** https://twitter.com/salesforce/status/1778831550173565367
- **Gondola URL:** https://gondola.cc/posts/21077640-salesforce-twitter
- **Thumbnail:** https://img.gondola.cc/tr:w-,h-,fo-auto/thumbnails/981aa989cc.jpg
- **Posted:** 2024-04-12T17:04:25.000+00:00
- **Account Owner:** Salesforce (@salesforce) — https://gondola.cc/salesforce

## Caption

On #AskMoreOfAI, we discuss how Retrieval Augmented Generation (RAG) improves your generative AI results with structured and unstructured data sources, like SQL databases and PDFs, and how LLM fine-tuning enhances response accuracy.

Hear the conversation with the CEOs of @Perplexity_AI, @Llama_Index, and @LangChainAI here on X, YouTube, and your favorite podcast platforms.

Episode #15: Workflows & Tooling to Create Trusted AI

Timestamps:
0:00 - Introduction
0:50 - Perplexity AI overview with @AravSrinivas
4:23 - LangChain overview with @hwchase17
6:23 - LIamaIndex overview with @jerryjliu0
7:55 - When do you use RAG versus fine-tuning?
9:22 - Aravind's background in artificial intelligence
11:40 - How do you manage ever-changing models with your tools?
13:40 - Value of open source projects
14:44 - Will any of you build your models in the future?
16:45 - Advantages of post-training ‌models
18:18 - Disadvantages of post-training
19:49 - How Perplexity AI tracks evals
20:30 - Thoughts on few-shot prompting
23:10 - What trends are you seeing with developers?
26:07 - Techniques for using RAG workflows
28:00 - Where should developers start with RAG pipelines?
33:00 - Do developers use AI copilots?
36:55 - How do we educate everyone on AI?
39:22 - What are your predictions for AI in the next year?
44:27 - How do we educate the next generation to succeed in an AI era?
49:22 - Three takeaways from @ClaraShih

## Stats

- **Views:** 0
- **Likes:** 14
- **Shares:** 7
- **Comments:** 2

## Tags

askmoreofai, 15

---
Copyright (c) Gondola