Monthly Roundup: Ray Compiled Graphs, Llama 3.2 and Multimodal AI, and Structured Data for RAG

Ben Lorica and Paco Nathan on Advances in AI Infrastructure, Multimodal Models, and and the state of Frontier Model Providers.

Subscribe: Apple • Spotify • Overcast • Pocket Casts • AntennaPod • Podcast Addict • Amazon •  RSS.

This is our monthly conversation on topics in AI and Technology with Paco Nathan, the founder of Derwen, a boutique consultancy focused on Data and AI.

Subscribe to the Gradient Flow Newsletter

Interview highlights – key sections from the video version:

1. Ray Compiled Graphs
2. SB 1047 Is Vetoed, what next?
3. Structure Is All You Need: enhancing RAG with structured & contextual information
4. Llama 3.2 & the state of frontier model developers
5. vLLM and Ray Data: status reports
6. The Bigger-is-Better Paradigm in AI
7. Recommendations

Relevant Links.

Ray Compiled Graphs

Structure Is All You Need

Llama 3.2

Status Update

• The State of vLLM: Roblox talk (“Challenges and Learnings from Supporting Multi-Modality on vLLM”);  Uber talk (“Large Scale Generative AI Batch Prediction with Ray and vLLM”)
• Ray Data GA

The Bigger-is-Better Paradigm in AI

Recommendations

• Ben:  Unit C: How the Pentagon and Silicon Valley Are Transforming the Future of War and Bach: Music in the Castle of Heaven
• Paco:  AML and real estate

---

Related content:

• A video version of this conversation is available on our YouTube channel.
• Previous monthly roundups with Ben Lorica and Paco Nathan
• Mars Lan → The Security Debate: How Safe is Open-Source Software?
• Yishay Carmiel → Generative AI in Voice Technology

---

If you enjoyed this episode, please support our work by encouraging your friends and colleagues to subscribe to our newsletter: