BROADCASTS.com

Our 173rd episode with a summary and discussion of last week's big AI news!\nWith hosts Andrey Kurenkov (https://twitter.com/andrey_kurenkov) and Jeremie Harris (https://twitter.com/jeremiecharris)\nSee full episode notes here.\nRead out our text newsletter and comment on the podcast at https://lastweekin.ai/\nIf you would like to become a sponsor for the newsletter, podcast, or both, please fill out this form.\nEmail us your questions and feedback at contact@lastweekinai.com and/or\xa0hello@gladstone.ai\nIn this episode of Last Week in AI, we explore the latest advancements and debates in the AI field, including Google's release of Gemini 1.5, Meta's upcoming LLaMA 3, and Runway's Gen 3 Alpha video model. We discuss emerging AI features, legal disputes over data usage, and China's competition in AI. The conversation spans innovative research developments, cost considerations of AI architectures, and policy changes like the U.S. Supreme Court striking down Chevron deference. We also cover U.S. export controls on AI chips to China, workforce development in the semiconductor industry, and Bridgewater's new AI-driven financial fund, evaluating the broader financial and regulatory impacts of AI technologies.\n\xa0\nTimestamps + links:\n(00:00:00) Intro / Banter\nTools & Apps(00:03:24) Google opens up Gemini 1.5 Flash, Pro with 2M tokens to the public\n(00:08:47) Meta is about to launch its biggest Llama model yet \u2014 here\u2019s why it\u2019s a big deal\n(00:12:38) Runway\u2019s Gen-3 Alpha AI video model now available \u2013 but there\u2019s a catch\n(00:16:28) This is Google AI, and it's coming to the Pixel 9\n(00:17:30) AI Firm ElevenLabs Sets Audio Reader Pact With Judy Garland, James Dean, Burt Reynolds and Laurence Olivier Estates\n(00:20:06) Perplexity\u2019s \u2018Pro Search\u2019 AI upgrade makes it better at math and research\n(00:23:12) Gemini\u2019s data-analyzing abilities aren\u2019t as good as Google claims\n\nApplications & Business(00:26:38) Quora\u2019s Chatbot Platform Poe Allows Users to Download Paywalled Articles on Demand\n(00:32:04) Huawei and Wuhan Xinxin to develop high-bandwidth memory chips amid US restrictions\n(00:34:57) Alibaba\u2019s large language model tops global ranking of AI developer platform Hugging Face\n(00:39:01) Here comes a Meta Ray-Bans challenger with ChatGPT-4o and a camera\n(00:43:35) Apple\u2019s Phil Schiller is reportedly joining OpenAI\u2019s board\n(00:47:26) AI Video Startup Runway Looking to Raise $450 Million\n\nProjects & Open Source(00:48:10) Kyutai Open Sources Moshi: A Real-Time Native Multimodal Foundation AI Model that can Listen and Speak\n(00:50:44) MMEvalPro: Calibrating Multimodal Benchmarks Towards Trustworthy and Efficient Evaluation\n(00:53:47) Anthropic Pushes for Third-Party AI Model Evaluations\n(00:57:29) Mozilla Llamafile, Builders Projects Shine at AI Engineers World's Fair\n\nResearch & Advancements(00:59:26) Researchers upend AI status quo by eliminating matrix multiplication in LLMs\n(01:05:55) AI Agents That Matter\n(01:12:09) WARP: On the Benefits of Weight Averaged Rewarded Policies\n(01:17:20) Scaling Synthetic Data Creation with 1,000,000,000 Personas\n(01:24:16) Found in the Middle: Calibrating Positional Attention Bias Improves Long Context Utilization\n\nPolicy & Safety(01:26:32) With Chevron\u2019s demise, AI regulation seems dead in the water\n(01:33:40) Nvidia to make $12bn from AI chips in China this year despite US controls\n(01:37:52) Uncle Sam relies on manual processes to oversee restrictions on Huawei, other Chinese tech players\n(01:40:57) U.S. government addresses critical workforce shortages for the semiconductor industry with new program\n(01:42:42) Bridgewater starts $2 billion fund that uses machine learning for decision-making and will include models from OpenAI, Anthropic and Perplexity\n\n(01:47:57) Outro