Last Week in AI

Our 190th episode with a summary and discussion of last week's* big AI news! *and sometimes last last week's Hosted by Andrey Kurenkov and Jeremie Harris. Note from Andrey: this one is coming out a bit later than planned, apologies! Next one will be coming out sooner. Feel free to email us your questions and feedback at [email protected] and/or [email protected] Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Sponsors: The Generator - An interdisciplinary AI lab empowering innovators from all fields to bring visionary ideas to life by harnessing the capabilities of artificial intelligence In this episode: * OpenAI's pitch for a $100 billion data center and AI strategy plan outlines infrastructure and regulatory needs, emphasizing AI's foundational role akin to electricity.  * Google's Gemini model challenges OpenAI's dominance, showing strong performance in chatbot arenas alongside generative AI advancements.  * DeepMind's AlphaFold3 gets open-sourced for academic use, while new chips from NVIDIA and Google show significant performance boosts.  * Anthropic and TSMC updates highlight strategic funding, regulation influences, and the complex dynamics of AI hardware and international policy. If you would like to become a sponsor for the newsletter, podcast, or both, please fill out this form. Timestamps + Links: (00:00:00) Intro / Banter (00:02:44) News Preview (00:03:34) Sponsor Break Tools & Apps (00:04:36) OpenAI, Google and Anthropic Are Struggling to Build More Advanced AI (00:16:22) OpenAI Nears Launch of AI Agent Tool to Automate Tasks for Users (00:19:14) Google drops new Gemini model and it goes straight to the top of the LLM leaderboard (00:19:14)  Chinese AI startup takes aim at OpenAI's Sora with image-to-video tool launch (00:20:04) Introducing the Forge Reasoning API Beta and Nous Chat: An Evolution in LLM Inference Applications & Business (00:23:47) OpenAI Discusses AI Data Center That Could Cost $100 Billion (00:26:48) Elon Musk's massive AI data center gets unlocked — xAI gets approved for 150MW of power, enabling all 100,000 GPUs to run concurrently (00:29:34) Newest Google and Nvidia Chips Speed AI Training (00:34:45) Ex-OpenAI CTO Murati’s New Team Takes Shape (00:34:45) Amazon Discussing New Multibillion-Dollar Investment in Anthropic Projects & Open Source (00:37:52) Google DeepMind open-sources AlphaFold 3, ushering in a new era for drug discovery and molecular biology (00:41:29) Near plans to build world’s largest 1.4T parameter open-source AI model Research & Advancements (00:45:38) The Super Weight in Large Language Models (00:55:42) Compositional Abilities Emerge Multiplicatively: Exploring Diffusion Models on a Synthetic Task (01:03:47) Mixture-of-Transformers: A Sparse and Scalable Architecture for Multi-Modal Foundation Models (01:08:14) Contextualized Evaluations: Taking the Guesswork Out of Language Model Evaluations Policy & Safety (01:11:14) The Code of Practice for general-purpose AI offers a unique opportunity for the EU (01:15:38) Three Sketches of ASL-4 Safety Case Components (01:23:05) U.S Department of Commerce finalizes $6.6 billion CHIPS Act funding for TSMC Fab 21 Arizona site , TSMC cannot make 2nm chips abroad now: MOEA (01:26:21) OpenAI to present plans for U.S. AI strategy and an alliance to compete with China (01:30:42) OpenAI loses another lead safety researcher, Lilian Weng (01:33:00) Outro

Our 189th episode with a summary and discussion of last week's big AI news! Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at [email protected] and/or [email protected] Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. In this episode: * OpenAI's acquisition of chat.com and internal shifts, including hardware lead hire and hardware model leaks, signal significant strategy pivots and challenges with model scaling and security.  * Saudi Arabia plans a $100 billion AI initiative aiming to rival UAE's tech hub, highlighting the region's escalating AI investments.  * U.S. penalties on GlobalFoundries for violating sanctions against SMIC underline ongoing challenges in enforcing AI-chip export controls.  * Anthropic collaborates with Palantir and AWS to integrate CLAWD into defense environments, marking a significant policy shift for the company. Sponsors: The Generator - An interdisciplinary AI lab empowering innovators from all fields to bring visionary ideas to life by harnessing the capabilities of artificial intelligence. The AI safety book “Uncontrollable" which is not a doomer book, but instead lays out the reasonable case for AI safety and what we can do about it. Max TEGMARK said that “Uncontrollable” is a captivating, balanced, and remarkably up-to-date book on the most important issue of our time" - find it on Amazon today! If you would like to become a sponsor for the newsletter, podcast, or both, please fill out this form. Timestamps + Links: (00:00:00) Intro / Banter (00:01:28) News Preview (00:02:10) Response to listener comments (00:05:02) Sponsor Break Tools & Apps (00:07:31) OpenAI Introduces ‘Predicted Outputs’ Feature: Speeding Up GPT-4o by ~5x for Tasks like Editing Docs or Refactoring Code (00:11:55) Anthropic’s Haiku 3.5 surprises experts with an “intelligence” price increase (00:17:10) Introducing FLUX1.1 [pro] Ultra and Raw Modes (00:19:11) X is testing a free version of Grok AI chatbot in select regions Applications & Business (00:21:39) OpenAI acquired Chat.com (00:23:40) Saudis Plan $100 Billion AI Powerhouse to Rival UAE Tech Hub (00:28:28) Meta’s former hardware lead for Orion is joining OpenAI (00:31:38) OpenAI Accidentally Leaked Its Upcoming o1 Model to Anyone With a Certain Web Address (00:35:50) Nvidia Rides AI Wave to Pass Apple as World’s Largest Company Projects & Open Source (00:37:53) ‘Unrestricted’ AI group Nous Research launches first chatbot — with guardrails (00:41:48) FrontierMath: The Benchmark that Highlights AI’s Limits in Mathematics (00:46:29) Hunyuan-Large: An Open-Source MoE Model with 52 Billion Activated Parameters by Tencent Research & Advancements (00:49:55) Applying “Golden Gate Claude” mechanistic interpretability techniques to protein language models. (00:58:3) Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-wise LoRA (01:05:55) From Naptime to Big Sleep: Using Large Language Models To Catch Vulnerabilities In Real-World Code (01:10:22) OpenAI reportedly developing new strategies to deal with AI improvement slowdown Policy & Safety (01:19:52)  What Donald Trump’s Win Means For AI (01:28:44) Fab Whack-A-Mole: Chinese Companies are Evading U.S. Sanctions (01:33:57) US fines GlobalFoundries for shipping chips to sanctioned Chinese firm (01:36:55) Anthropic teams up with Palantir and AWS to sell its AI to defense customers (01:39:23) Outro

Our 188th episode with a summary and discussion of last week's big AI news! Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at [email protected] and/or [email protected] Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. This episode was sponsored by The Generator. If you would like to become a sponsor for the newsletter, podcast, or both, please fill out this form. In this episode: * Meta's open-source models utilized by China's military prompt regulatory adjustments; US agencies gain access to counterbalance.  * OpenAI partners with Broadcom and AMD to develop custom AI hardware, aiming for profitability and reducing inference costs.  * Physical Intelligence unveils a generalist robot control policy with a $400M funding boost, showcasing significant advancements in zero-shot task performance.  * New U.S. regulation mandates quarterly reporting for large AI model training and computing cluster acquisitions, aiming to bolster national security. Timestamps + Links: (00:00:00) Intro / Banter (00:02:16) News Preview (00:03:05) Response to listener comments / corrections (00:05:00) Sponsor Break Tools & Apps (00:06:28) OpenAI’s search engine is now live in ChatGPT (00:12:18) Image Playground, ChatGPT, and more Apple Intelligence features roll out in beta (00:14:34) GitHub Copilot will support models from Anthropic, Google, and OpenAI (00:19:00) Introducing the analysis tool in Claude.ai (00:21:34) ElevenLabs Introduces Voice Design: A New AI Feature that Generates a Unique Voice from a Text Prompt Alone (00:24:18) Midjourney's new web editor lets you tweak images uploaded from your PC (00:26:02) Watch out, Midjourney — Recraft just announced new AI image generator model Applications & Business (00:29:57) Meta strikes multi-year AI deal with Reuters (00:33:15) OpenAI will start using AMD chips and could make its own AI hardware in 2026 (00:40:47) Elon Musk's xAI in talks to raise funding valuing it at $40 billion, WSJ reports (00:46:07) Physical Intelligence, a Robot A.I. Specialist, Raises Millions From Bezos (00:48:32) Waymo ramps up robotaxi push with $5.6 bn in funding (00:49:11) Alphabet's Waymo Serving Over 150,000 Paid Robotaxi Rides Every Week Now, Surging 50% In 2 Months Projects & Open Source (00:51:23) Meta AI Silently Releases NotebookLlama: An Open Version of Google’s NotebookLM (00:54:59) Meta Releases Quantized Llama 3.2 with 4x Inference Speed on Android Phones (00:59:16) OpenAI Releases SimpleQA: A New AI Benchmark that Measures the Factuality of Language Models Research & Advancements (01:08:19) This Is a Glimpse of the Future of AI Robot (01:15:06) Can Language Models Replace Programmers? REPOCOD Says 'Not Yet' (01:19:01) Brain-like Functional Organization within Large Language Models (01:21:20) Decart’s AI simulates a real-time, playable version of Minecraft (01:25:39) Raising the bar on SWE-bench Verified with Claude 3.5 Sonnet Policy & Safety (01:29:06) Commerce just proposed the most significant federal AI regulation to date – and no one noticed (01:35:04)Anthropic warns of AI catastrophe if governments don't regulate in 18 months (01:39:32) Open Source Bites Back as China’s Military Makes Full Use of Meta AI (01:46:35) Meta says it’s making its Llama models available for US national security applications (01:48:16) Outro

Our 187th episode with a summary and discussion of last week's big AI news, now with Jeremie co-hosting once again! Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at [email protected] and/or [email protected] Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. This episode was sponsored by The Generator. If you would like to become a sponsor for the newsletter, podcast, or both, please fill out this form. Timestamps + Links: (00:00:00) Intro / Banter (00:03:07) Response to listener comments / corrections (00:05:13) Sponsor Read) Tools & Apps (00:06:22) Anthropic’s latest AI update can use a computer on its own (00:18:09) AI video startup Genmo launches Mochi 1, an open source rival to Runway, Kling, and others (00:20:37) Canva has a shiny new text-to-image generator (00:23:35) Canvas Beta brings Remix, Extend, and Magic Fill to Ideogram users (00:26:16) StabilityAI releases Stable Diffusion 3.5  (00:28:27) Bringing Agentic Workflows into Inflection for Enterprise Applications & Business (00:32:35) Crusoe’s $3.4B joint venture to build AI data center campus with up to 100,000 GPUs (00:39:08) Anthropic reportedly in early talks to raise new funding on up to $40B valuation (00:45:47) Longtime policy researcher Miles Brundage leaves OpenAI (00:49:53) NVIDIA’s Blackwell GB200 AI Servers Ready For Mass Deployment In December (00:52:41) Foxconn building Nvidia superchip facility in Mexico, executives say (00:55:27) xAI, Elon Musk’s AI startup, launches an API Projects & Open Source (00:58:32) INTELLECT-1: The First Decentralized 10-Billion-Parameter AI Model Training (01:06:34) Meta FAIR Releases Eight New AI Research Artifacts—Models, Datasets, and Tools to Inspire the AI Community (01:10:02) Google DeepMind is making its AI text watermark open source Research & Advancements (01:13:21) OpenAI researchers develop new model that speeds up media generation by 50X (01:17:54) How much AI compute is out there, and who owns it? (01:25:28) Rewarding Progress: Scaling Automated Process Verifiers for LLM Reasoning (01:33:30) Inference Scaling for Long-Context Retrieval Augmented Generation Policy & Safety (01:41:50) Announcing our updated Responsible Scaling Policy (01:48:52) Anthropic is testing AI’s capacity for sabotage (01:56:30) OpenAI asked US to approve energy-guzzling 5GW data centers, report says (02:00:05) US Probes TSMC’s Dealings with Huawei (02:03:03) TikTok owner ByteDance taps TSMC to make its own AI GPUs to stop relying on Nvidia — the company has reportedly spent over $2 billion on Nvidia AI GPUs (02:06:37) Outro