Explore the latest in AI advancements with Moonlight, a groundbreaking model from Moonshot AI and UCLA, promising efficiency in AI training. Delve into the legal challenges Meta faces over using copyrighted content for AI training, and understand the implications of the EU's AI Act, a pioneering regulation shaping the future of AI governance. This episode offers insights into the technological, ethical, and legal dimensions of AI development.

Sources:

https://www.marktechpost.com/2025/02/22/moonshot-ai-and-ucla-researchers-release-moonlight-a-3b-16b-parameter-mixture-of-expert-moe-model-trained-with-5-7t-tokens-using-muon-optimizer/

https://techcrunch.com/2025/02/22/us-ai-safety-institute-could-face-big-cuts/

https://techcrunch.com/2025/02/21/court-filings-show-meta-staffers-discussed-using-copyrighted-content-for-ai-training/

https://apnews.com/article/ai-act-european-union-chatbots-155157e2be2e42d0f1acca33983d8c82

Outline:

(00:00:00) Introduction

(00:00:42) Moonshot AI and UCLA Researchers Release Moonlight: A 3B/16B-Parameter Mixture-of-Expert (MoE) Model Trained with 5.7T Tokens Using Muon Optimizer

(00:03:28) US AI Safety Institute could face big cuts

(00:06:21) Court filings show Meta staffers discussed using copyrighted content for AI training

(00:09:38) Europe's AI Act: How does it work and what happens next?