{"version":"1.0","type":"rich","provider_name":"Acast","provider_url":"https://acast.com","height":250,"width":700,"html":"<iframe src=\"https://embed.acast.com/$/6527e4f1d40c9700125f42de/6afb15ce77e39bab67d4bbbb?\" frameBorder=\"0\" width=\"700\" height=\"250\"></iframe>","title":"Beyond Scaling: Toward Continual and Adaptive Intelligence | Irina Rish","description":"<p>This recording with Irina Rish is from Vision Weekend UK 2026, held in London in June 2026. View the recording along with the slides on <a href=\"https://www.youtube.com/watch?v=Ia5vxdWiSsw\" rel=\"noopener noreferrer\" target=\"_blank\">our YouTube channel</a>.</p><p>Irina Rish is a professor at Université de Montréal and a core academic member of Mila, the Quebec AI Institute. She holds a Canada CIFAR AI Chair and leads the Autonomous AI Lab. Her research covers neural scaling laws, emergent behaviors in large language models, and continual learning: the challenge of building AI systems that accumulate knowledge without forgetting.</p><p><br></p><p>In this talk, Irina makes the case that scaling is not sufficient for genuinely adaptive intelligence. She examines what continual and adaptive learning actually require, why current training paradigms produce systems that fail to generalize outside their training distribution, and what a different approach to AI training would need to look like.</p><p>If you want to join a future Vision Weekend, details and upcoming dates are at <a href=\"https://foresight.org/\" rel=\"noopener noreferrer\" target=\"_blank\">foresight.org</a>.</p>","author_name":"Foresight Institute"}