{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"Neural Newscast","title":"Google Gemini 3.1 and LeCun’s $1B World Model Bet [Model Behavior]","html":"<iframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless src=\"https://share.transistor.fm/e/8ed1edf4\"></iframe>","width":"100%","height":180,"duration":498,"description":"This episode explores the divergence in strategies toward superintelligence, contrasting Sam Altman’s focus on scaling transformers with Yann LeCun’s $1.03 billion bet on world models through his new venture, AMI Labs. We also examine Google’s latest integration of Gemini 3.1 into the Android ecosystem, featuring new agentic tools like Auto Browse, the Nano Banana image editing suite, and Rambler for Gboard. Finally, the show provides a technical breakdown of Gemini 3.1 Pro versus the open-source Gemma 4, highlighting significant performance gaps in reasoning speed and context window capacity. Key topics include the fragmentation of AI features across premium mobile devices and the rise of local, privacy-focused model execution.","thumbnail_url":"https://img.transistorcdn.com/mkCnMvKg2YZJk2kZMcI1a1R5MdeCfMFSDLiEp95sLBs/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS84ZmVm/ZGJhOGNlMGI4ZDQ3/NGFlYzg3ZTk5NDVm/MDg5Zi5wbmc.webp","thumbnail_width":300,"thumbnail_height":300}