Visit Mixture of Experts podcast page to get more AI content
→ https://ibm.biz/~YOqgkKvh7
This week on Mixture of Experts, our team of panelists talk about big infrastructure changes and small models. IBM is teaming up with Together AI on a multi-year deal to build a massive inference cluster on IBM Cloud, powered by NVIDIA's newest chips. The goal: cheaper, faster access to open-source AI models for enterprises, with the cluster expected to go live in early 2027.
Then there's Meta, which just open-sourced Muse Glimmer—a 30B-parameter model small enough to run locally on your laptop's GPU. Think coding, function calling, and agent tasks, all without needing the cloud. But as models get smaller, what does it mean for larger AI providers?
And finally, the story that's got everyone talking: OpenAI says its upcoming Astra model might be hitting "Critical" cybersecurity capability levels, meaning it could potentially find and exploit zero-days on its own. We dig into what that actually means and what guardrails OpenAI is putting up.
All that and more on Mixture of Experts.
00:00 – Intro
1:03 - IBM AI Partnership
11:43 - Meta Muse Glimmer
24:10 - OpenAI’s Astra model
"The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity."
AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/~NEVRq9OsZ
#MetaMuseGlimmer #IBMAIPartnership #openai
AI was used in the creation of the transcript and metadata for this video.