Hugging Face — The Video Signal

Sep 18, 2026 · 38:47

Pacing the AI frontier, IBM Granite 4.2 & Meta’s Muse assistant

This episode provides a deep dive into the current state of frontier AI, covering the debate around slowing development (pacing), the technical specifications of IBM's Granite 4.2 models, and Meta's push into personal agents with Muse. Key technical takeaways include the focus on smaller, auditable models, the use of synthetic data for training, and the critical need for robust sandboxing and guardrails for agentic workflows.

Key takeaways

Pacing AI Development 25:29
Concerns regarding existential risk and the ability of agents to find 'day zero vulnerabilities' have led to calls for slowing AI development. Experts suggest that the focus should be on implementing guardrails for agent-to-agent communication and limiting concurrent agent calls to mitigate economic risks.
IBM Granite 4.2 Release 25:29
The new Granite models (3B, 8B, 30B) are designed for enterprise use, featuring native step-by-step reasoning and support for agentic workflows like planning and tool calling. The models are available on Apache 2.0 license via Hugging Face.
Meta's Muse Agent
Muse is Meta's personal AI agent, designed to run on a secure virtual machine (VM) for isolation. While technically advanced in its design, the discussion highlights ongoing concerns regarding user privacy and the security risks of handing over personal data to such agents.

Watch on YouTube Full article

Sep 8, 2026 · 54:15

Specializing AI for Regulated Industries - How Domyn Uses NVIDIA Nemotron

This livestream details Domyn's journey toward building a family of sovereign AI models for regulated industries, emphasizing full ownership and control over the entire model stack. The presentation covers advanced LLM development techniques—including model compression (pruning/distillation), continual pretraining (CPT), Supervised Fine-Tuning (SFT), and Reinforcement Learning (RL)—using the NVIDIA open source ecosystem. A significant focus is placed on the tooling required to manage these complex pipelines at scale, including custom tools like Swarm and FinalFlows, which are crucial for build engineers managing large-scale compute infrastructure.

Key takeaways

Sovereign AI requires full stack control 0:28
For regulated industries, achieving sovereign AI necessitates controlling every layer of the stack—from compute to model ownership—rather than relying on point solutions. This approach is critical for governance and auditability. (0:28-1:35)
Domain Large development pipeline 7:16
Domyn developed Domain Large by starting with Coliseum 355, followed by model compression (pruning/distillation), CPT to expand context up to 128K tokens, and SFT to enable reasoning. This was executed using the NVIDIA stack on H200s in DJX cloud. (4:36-7:39)
Domain Small for efficiency 23:43
To address cost concerns associated with large models, Domain Small (10B parameters) was created. Its training heavily leveraged Reinforcement Learning (RL) and Direct Preference Optimization (DPO), demonstrating that smaller models can achieve strong performance through advanced post-training curricula. (14:23-17:56)
Tooling for scalable ML pipelines 40:08
Domyn developed internal tools like Swarm (CLI/Python) and FinalFlows (DAG library) to manage complex, interconnected jobs on Slurm clusters. These tools remove friction when running large-scale evaluations and training across European infrastructure. (24:08-31:56)

Watch on YouTube Full article

Sep 4, 2026 · 20:48

Why AI Agents Need Million-Token Context — Thomas Wolf & Olive Song, MiniMax

The discussion details MiniMax M3, a frontier open-source model featuring a functional one-million-token context window and native multimodality (text, image, video). The efficiency of this long context is achieved through the proprietary Minimax Sparse Attention (MSA) architecture. Furthermore, the conversation highlights advanced research practices at MiniMax, including using intern contributions for architectural improvements and training models from the very first step to ensure optimal multimodal performance.

Key takeaways

MiniMax M3 Capabilities 4:53
M3 is a highly capable open-source model (400B total parameters, 20B activated) that combines coding proficiency, agentic capabilities, and multimodal understanding across text, images, and video. It features a functional one-million-token context window.
Long Context Necessity for Agents 6:30
The need for extremely long context windows is driven by complex agentic tasks, which require processing multi-round conversations and accumulating tool responses from an entire environment, making shorter contexts insufficient.
Sparse Attention Architecture (MSA) 7:30
The Minimax Sparse Attention (MSA) architecture is designed to efficiently scale context length. It uses an index branch to select relevant context blocks, followed by a sparse attention branch for calculation, allowing scalability in both length and model size.
Native Multimodality Training 17:25
MiniMax trains M3 from the very first step as a multimodal model (native multimodality), rather than adding vision capabilities after text pre-training. This approach prevents performance degradation in text understanding and ensures better scalability.
Internal Research Culture 13:49
MiniMax utilizes an open research culture where external community members (including interns) can propose projects, find model weaknesses through evaluations, and contribute code/improvements that are integrated into final training runs.

Watch on YouTube Full article

Aug 28, 2026 · 30:59

IBM’s mainframe chip collab, NVIDIA’s Poolside deal & Ox Alpha’s reveal

The discussion covered major developments in AI infrastructure, focusing on IBM's new dual-architecture mainframe processor combining z/OS and Arm. This aims to bring modern AI workloads closer to mission-critical data residing on mainframes. Furthermore, NVIDIA's strategy was analyzed through its $6 billion deal with Poolside and the acquisition of Hugging Face, positioning NVIDIA as a central player in the open-source AI ecosystem by controlling key software standards. Finally, the reveal of Z.ai’s GLM-5.3-Flash model highlighted the trend toward stealth model releases.

Key takeaways

IBM's Dual-Architecture Mainframe Processor 21:50
IBM unveiled a new dual processor architecture at Hot Chips that combines IBM Z (mainframe workload) with Arm. This allows systems to run Arm-native Linux workloads alongside z/OS, addressing the challenge of integrating modern AI software into mission-critical mainframe environments.
NVIDIA's Open Ecosystem Strategy 2:00
NVIDIA is making a strategic play to be the center of open-source AI by acquiring Hugging Face (the cornerstone of open AI software) and securing a $6 billion license deal with Poolside. This solidifies their position in hardware while maintaining an open model ecosystem.
LLM Model Release Tactics 24:30
The anonymous 'Ox Alpha' model was revealed to be Z.ai’s GLM-5.3-Flash, an open-source LLM built with sparse and linear attention techniques. The discussion noted that stealth launches are a highly effective marketing strategy for generating hype and speculation.

Watch on YouTube Full article

Aug 20, 2026 · 20:37

How I automate my own job at Hugging Face using agents — Niels Rogge, Hugging Face

Niels Rogge details how he automated his role at Hugging Face—the 'Google Drive to the hub' team—which focuses on improving the discoverability of machine learning artifacts. He built two systems: an initial deterministic workflow for outreach (using cron jobs and LLM APIs) and a subsequent fully autonomous agent loop for follow-up actions. The architecture leverages modern tooling like Modal, Bash CLI skills, and advanced models (e.g., GLM 5.2) to scale the process of identifying missing artifacts and prompting researchers to publish them on Hugging Face.

Key takeaways

The Problem: Artifact Discoverability
ML weights and datasets are often published on third-party services (Google Drive, Zenodo) rather than the centralized platform (Hugging Face), hindering discoverability. The goal is to automate outreach to authors.
Initial Automation: Deterministic Workflow 11:43
The first phase used a deterministic workflow, running as a nightly cron job on GitHub Actions. This approach utilized LLM APIs in predefined steps without an agent framework, offering high predictability and control.
Advanced Automation: Autonomous Agent Loop 15:36
The follow-up process was automated using a fully autonomous agent loop (e.g., leveraging the Claude agents SDK). This flexible approach allows the agent to use tools and skills, such as Bash and the Hugging Face CLI, to interact with GitHub issues.

Watch on YouTube Full article

Jul 28, 2026 · 53:25

Doors of (AI)pportunity: The Front and Backdoors of LLMs - Kasimir Schulz & Kenneth Yeung

The talk provides an advanced deep dive into LLM security vulnerabilities, moving beyond simple prompt injection attacks. Speakers demonstrate sophisticated methods including Knowledge Return Oriented Prompting (KROP), Policy Puppetry, and direct manipulation of the model's computational graph to implant backdoors. The discussion highlights that even seemingly safe formats (like ONNX) can be exploited if the underlying architecture is compromised, posing significant supply chain risks for AI-driven applications.

Key takeaways

Multiple Layers of Defense are Required 35:35
System prompts alone are insufficient protection. Developers must implement multiple layers of defense to mitigate attacks, as demonstrated by the failure of simple system prompt rules against advanced exploits.
Indirect Prompt Injection is a Major Threat 43:25
Attacks do not require direct user interaction. Compromise can occur through injected text in external sources like customer reviews, emails (especially those processed by Retrieval Augmented Generation/RAG systems), or calendar invites.
Model Backdoors are Difficult to Detect 50:15
Backdoors can be implanted into the computational graph of a model (e.g., changing an animal classification logic based on a specific color trigger). These backdoors can remain dormant until a precise trigger condition is met, making them highly effective and difficult to detect.

Watch on YouTube Full article

Jul 24, 2026 · 17:28

Training Frontier Models to Out-Think Hackers — Uri Rolls, Arithmetic & Thom Wolf, Hugging Face

This talk addresses the rapidly shifting economics of cyber security due to increasingly powerful AI models. While frontier models can perform extensive reconnaissance and find vulnerabilities, they often fail at the critical 'logical leap' required for exploitation—the ability to reason across complex system states. Arithmetic proposes a new defense paradigm: creating specialized benchmarks focused on access control by having human researchers discover real-world zero days in open-source software. The goal is to train models to replicate the deep reasoning and multi-step logic of skilled attackers, thereby giving defenders a lasting edge.

Key takeaways

AI Models Lack World Modeling for Exploitation 17:01
Current large language models (LLMs) struggle with building dynamic world models. The benchmark developed by Arithmetic shows that even advanced models like GPT 5.5 and Opus fail to make the necessary logical leap required to exploit a system, despite successfully reaching the vulnerability check.
Open Source Models are Key to Cyber Defense 5:21
The solution to modern cyber challenges requires open-source models. The speaker argues that relying solely on a few large companies is insufficient, and collaboration using open source LLMs will be crucial for building the next generation of defensive systems.
Focusing on Access Control Vulnerabilities 11:21
Arithmetic's benchmark focuses specifically on access control, as this is the primary entry point for most attacks. The vulnerabilities tested are 'logic based,' meaning they exploit discrepancies between how different parts of a system check permissions (e.g., checking by name vs. checking by ID).

Watch on YouTube Full article