In this episode of the AI TriPod podcast, the hosts discuss the latest advancements in AI, focusing on the Llama 4 model set, Amazon Nova Sonic, and the Google Cloud Next 2025 agent-to-agent protocol. They explore the features, benchmarks, and implications of these technologies, emphasizing the importance of open-source models and the evolving landscape of AI interactions.

Hosts of the Show

Karthick Arun - Show Host - Highschooler in 10th grade

Loveneet Pal Singh - Show Host - CS Undergrad at ASU

Arun Arunachalam - Show Host - Industry expert with 25 years of experience

Chapters

00:00 Intro

00:42 Introduction to Llama 4 and AI Models

02:45 Benchmarking Llama 4: Performance and Comparisons

05:55 Open Source vs. Proprietary Models

08:45 Cost Efficiency and Value of Llama 4

11:24 Context Window and Its Impact on Performance

14:01 Final Thoughts on Llama 4 and Future Implications

15:02 Benchmarking Model Performance

16:55 Exploring Amazon Nova Sonic

18:29 Voice Cloning and Future Possibilities

20:29 Customer Service Revolution

27:37 Agent to Agent Communication Protocol

32:26 Security Concerns in AI Communication

33:47 Outro

Sources

• 1. ⁠https://www.youtube.com/watch?v=r8d62HsnQA0
  2. https://artificialanalysis.ai/models/llama-4-scout
  3. https://developers.googleblog.com/en/a2a-a-new-era-of-agent-interoperability/
  4. https://github.com/google/A2A
  5. https://www.llama.com/models/llama-4/
  6. https://aws.amazon.com/blogs/aws/introducing-amazon-nova-sonic-human-like-voice-conversations-for-generative-ai-applications/
  7. https://www.theverge.com/news/646968/openai-chatgpt-long-term-memory-upgrade