Large language models

Jul 24, 2026

Hugging Face breach: OpenAI’s model breaks containment

This episode of Mixture of Experts explores pivotal AI developments: OpenAI's model breaching containment, Claude's Fable disproving a mathematical conjecture, Moonshot AI's massive 2.8 trillion parameter Kimi K3, and Google's shift to smaller, more efficient Gemini Flash models. The panel discusses AI security, its role in scientific discovery, and the evolving market strategies for model deployment, highlighting the tension between scale and efficiency.

Jul 24, 2026

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

Thom Wolf and Uri Rolls discuss the critical role of AI in cybersecurity, presenting a new benchmark called Masov. They argue that while frontier models excel at reconnaissance, they lack the sophisticated reasoning to exploit complex, logic-based zero-day vulnerabilities, such as a Keycloak name-versus-ID exploit. The solution, they propose, lies in high-quality, open-source AI models trained on real-world zero-day data to enable defenders to outpace attackers and build a new, AI-native security stack.

Jul 23, 2026

The Unreasonable Effectiveness of Separating the Task from the Model — Maxime Rivest, DSPy

DSPy emphasizes separating task definition from model implementation using a "Signature" (inputs/outputs) to enable flexible, optimizable, and scalable AI programs. The framework relies on three pillars—instructions (specs), hard constraints (code), and examples (evals)—to fully specify tasks. DSPy 4.0 introduces DSPy Flex for learning program harnesses and Qualitative Learning for automated, feedback-driven evaluation refinement, offering significant benefits for enterprise applications and addressing "last-mile learning" for future AI systems.

Jul 23, 2026

AI, Corporate Responsibility & Democratic Legitimacy: Extended Q&A • Joanna Bryson • GOTO 2025

Joanna Bryson challenges popular AI assumptions, positing current generative AI as powerful tools for cultural knowledge compression, not autonomous intelligences. She emphasizes that AI's capabilities are nearing the human knowledge frontier, requiring focus on human coordination and governance. Bryson critically examines AI's impact on mental health and law, advocating for data-driven regulation and comprehensible systems. She calls for engineering activism, asserting human agency over technological determinism and stressing the importance of transparency and critical thinking in shaping AI's future.

Jul 22, 2026

Open Models: Kimi K3, Qwen 3.8, Xi's WAIC Speech, Distillation, The Open-Closed Gap, and What's Next

Nathan Lambert and Florian Brand discuss the accelerating open model landscape, focusing on the surge in Chinese models like Kimi K3 and GLM 5.2. They explore the reasons behind China's progress, the evolving US ecosystem, the cybersecurity implications of open-source bans, and debunk common misconceptions about distillation, particularly challenging Ben Thompson's recent claims. The episode concludes with predictions and a frontier model tier list.

Jul 21, 2026

The Real AI Frontier Isn't Smarter Machines (with Catherine Williams)

Dr. Catherine Williams, a former black-hole physicist and early data science leader, explores the field's evolution from Bayesian models to LLMs. She passionately argues that deep mathematical understanding and the ability to build robust mental models are more crucial than ever, even as AI automates technical tasks. Williams also discusses the impact of embeddings, the changing economics of frontier AI, and her work at the nonprofit Candid, advocating for a human-centric approach to intelligence in the age of AI.