Nick Vasiloglou, VP of Research at Relational AI, analyzes the key trends from NeurIPS 2025, highlighting the most impactful and under-the-radar developments for industry professionals. The discussion covers the rise of data markets through real-time attribution, the sophisticated engineering behind capable small language models (SLMs), the explosion of AI for science, and the shift towards post-training models with real-world tools.
Demis Hassabis, CEO of Google DeepMind, outlines the path to AGI, which he predicts by 2030. He discusses the profound impact of AI on science, particularly in revolutionizing drug discovery with systems like AlphaFold, and posits that AI will enable new forms of simulation-based science. Hassabis also delves into the philosophical underpinnings of his work, viewing information as the universe's most fundamental quantity and advocating for developing AGI as a powerful tool before tackling the deeper questions of consciousness.
Demis Hassabis, CEO of Google DeepMind, outlines the remaining challenges on the path to AGI, including memory, continual learning, and true reasoning. He discusses how learnings from AlphaGo are shaping agent development, the strategic importance of powerful small models like Gemma, and his vision for AI as the ultimate tool for scientific discovery, offering a framework for identifying breakthrough opportunities and advice for founders building in the age of AI.
Mistral's Pavan (Voxtral lead) and Guillaume (Chief Scientist) discuss the new Voxtral TTS model, its novel architecture using flow matching for efficient, high-quality speech generation. They elaborate on Mistral's strategy of delivering specialized, open-weight models and the Mistral Forge platform, which empowers enterprises to leverage their proprietary data through fine-tuning for privacy, cost-effectiveness, and superior performance. The conversation also covers Mistral Small, the future of AI for science, and the company's commitment to open-source and foundational research, including formal proving as a proxy for long-horizon reasoning.
Professor Heather Kulik shares her hard-won perspective on applying AI to materials science, from discovering novel polymers with surprising quantum properties to the practical limitations of LLMs and the critical need for integrating deep domain expertise with data-driven methods.
Max Welling connects the dots between quantum gravity, equivariant neural networks, and diffusion models, explaining how these concepts from theoretical physics are now powering a new generation of AI for materials discovery to tackle climate change. He introduces the concept of a "Physics Processing Unit" and details the architecture of his startup, CuspAI.