Last Week in AI

Our 222st episode with a summary and discussion of last week's big AI news! Recorded on 10/03/2025 Hosted by Andrey Kurenkov and co-hosted by Jon Krohn Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ In this episode: ChatGPT parent company OpenAI announces Sora 2 with AI video appAnthropic releases Claude Sonnet 4.5 in latest bid for AI agents and coding supremacy | The VergeMeta launches 'Vibes,' a short-form video feed of AI slop | TechCrunchOpenAI launches ChatGPT Pulse to proactively write you morning briefs | TechCrunchOpenAI rolls out safety routing system, parental controls on ChatGPT | TechCrunchThe Latest Gemini 2.5 Flash-Lite Preview is Now the Fastest Proprietary Model (External Tests) and 50% Fewer Output Tokens - MarkTechPostMicrosoft just added AI agents to Word, Excel, and PowerPoint - how to use them | ZDNET OpenAI takes on Google, Amazon with new agentic shopping system | TechCrunchExclusive: Mira Murati’s Stealth AI Lab Launches Its First Product | WIREDOpenAI is the world's most valuable private company after private stock sale | TechCrunchElon Musk’s xAI accuses OpenAI of stealing trade secrets in new lawsuit | Technology | The GuardianFormer OpenAI and DeepMind researchers raise whopping $300M seed to automate science | TechCrunch [2509.16941] SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks? [2509.17196] Evolution of Concepts in Language Model Pre-Training[2509.19284] What Characterizes Effective Reasoning? Revisiting Length, Review, and Structure of CoT[2507.02954] Advanced Financial Reasoning at Scale: A Comprehensive Evaluation of Large Language Models on CFA Level III[2509.24552] Short window attention enables long-term memorization SB 53, the landmark AI transparency bill, is now law in California | The VergeElon Musk's xAI offers Grok to federal government for 42 cents | TechCrunchCharacter.AI removes Disney characters from platform after studio issues warningSpotify's Attempt to Fight AI Slop Falls on Its FaceSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Our 221st episode with a summary and discussion of last week's big AI news! Recorded on 09/19/2025 Note: we transitioned to a new RSS feed and it seems this did not make it to there, so this may be posted about 2 weeks past the release date. Hosted by Andrey Kurenkov and co-hosted by Michelle Lee Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ In this episode: Timestamps: OpenAI upgrades Codex with a new version of GPT-5Google Injects Gemini Into Chrome as AI Browsers Go Mainstream | WIREDAnthropic’s Claude can now make you a spreadsheet or slide deck. | The VergeLuma AI's New Ray3 Video Generator Can 'Think' Before Creating - CNETOpenAI secures Microsoft's blessing to transition its for-profit arm | TechCrunchMicrosoft to lessen reliance on OpenAI by buying AI from rival Anthropic | TechCrunchFigure AI passes $1B with Series C funding toward humanoid robot development - The Robot ReportChina’s Unitree plans $7 billion IPO valuation as humanoid robot race heats upTesla's robotaxi plans for Nevada move forward with testing permit | TechCrunchAmazon's Zoox jumps into U.S. robotaxi race with Las Vegas launchReplit hits $3B valuation on $150M annualized revenue | TechCrunchPerplexity reportedly raised $200M at $20B valuation | TechCrunch[2509.07604] K2-Think: A Parameter-Efficient Reasoning System[2509.09614] LoCoBench: A Benchmark for Long-Context Large Language Models in Complex Software Engineering[2509.15155] Self-Improving Embodied Foundation Models[2509.13805] Towards a Physics Foundation Model[2509.12129] Embodied Navigation Foundation ModelAnthropic endorses California's AI safety bill, SB 53 | TechCrunchWarner Bros. Sues Midjourney, Joins Studios' AI Copyright BattleRolling Stone Publisher Sues Google Over AI Overview Summaries See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Our 221st episode with a summary and discussion of last week's big AI news! Recorded on 09/19/2025 Hosted by Andrey Kurenkov and co-hosted by Michelle Lee Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ In this episode: Timestamps: OpenAI upgrades Codex with a new version of GPT-5Google Injects Gemini Into Chrome as AI Browsers Go Mainstream | WIREDAnthropic’s Claude can now make you a spreadsheet or slide deck. | The VergeLuma AI's New Ray3 Video Generator Can 'Think' Before Creating - CNET OpenAI secures Microsoft's blessing to transition its for-profit arm | TechCrunchMicrosoft to lessen reliance on OpenAI by buying AI from rival Anthropic | TechCrunchFigure AI passes $1B with Series C funding toward humanoid robot development - The Robot ReportChina’s Unitree plans $7 billion IPO valuation as humanoid robot race heats upTesla's robotaxi plans for Nevada move forward with testing permit | TechCrunchAmazon's Zoox jumps into U.S. robotaxi race with Las Vegas launchReplit hits $3B valuation on $150M annualized revenue | TechCrunchPerplexity reportedly raised $200M at $20B valuation | TechCrunch [2509.07604] K2-Think: A Parameter-Efficient Reasoning System[2509.09614] LoCoBench: A Benchmark for Long-Context Large Language Models in Complex Software Engineering [2509.15155] Self-Improving Embodied Foundation Models[2509.13805] Towards a Physics Foundation Model[2509.12129] Embodied Navigation Foundation Model Anthropic endorses California's AI safety bill, SB 53 | TechCrunchWarner Bros. Sues Midjourney, Joins Studios' AI Copyright BattleRolling Stone Publisher Sues Google Over AI Overview Summaries

Our 220th episode with a summary and discussion of last week's big AI news! Recorded on 08/30/2025 Check out Andrey's work over at Astrocade , sign up to be an ambassador here Hosted by Andrey Kurenkov and co-hosted by Daniel Bashir Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ Timestamps + Links: Google Gemini's AI image model gets a 'bananas' upgrade | TechCrunchAnthropic launches a Claude AI agent that lives in Chrome | TechCrunchAnthropic’s Claude chatbot can now remember your past conversations | The VergeGoogle Launches AI ‘Guided Learning’ Tool to Teach UsersApple Intelligence’s ChatGPT integration will use GPT-5 starting with iOS 26 | The VergeOpenAI Adds New Features to Codex, Like IDE Extension and GitHub Code ReviewsLovable projects $1B in ARR within next 12 months | TechCrunchDecart hits $3.1 billion valuation on $100 million raise to power real-time interacti | CtechCohere raises $500M to beat back generative AI rivals | TechCrunchPony AI, Nearing Full-Year Robotaxi Goal, Eyes European Markets - BloombergCo-founder of Elon Musk's xAI departs the company | TechCrunchMeta AI Just Released DINOv3: A State-of-the-Art Computer Vision Model Trained with Self-Supervised Learning, Generating High-Resolution Image Features - MarkTechPostGLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation ModelsChina’s DeepSeek Releases V3.1, Boosting AI Model’s Capabilities - BloombergOpen weight LLMs exhibit inconsistent performance across providersMicrosoft Released VibeVoice-1.5B: An Open-Source Text-to-Speech Model that can Synthesize up to 90 Minutes of Speech with Four Distinct Speakers - MarkTechPostDeep Think with ConfidenceGenerative AI reshapes U.S. job market, Stanford study showsInside the US Government's Unpublished Report on AI Safety | WIREDU.S. Government to Take Cut of Nvidia and AMD A.I. Chip Sales to China - The New York TimesAnthropic Settles High-Profile AI Copyright Lawsuit Brought by Book AuthorsAI companion apps on track to pull in $120M in 2025 | TechCrunch

Our 219th episode with a summary and discussion of last week's big AI news! Recorded on 08/08/2025 Check out Andrey's work over at Astrocade , sign up to be an ambassador here Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ In this episode: Timestamps + Links: Astrocade rolls out AI agent-powered game creation experience so anyone can create games OpenAI’s GPT-5 is hereAnthropic Releases Claude Opus 4.1 With Agentic, Coding and Reasoning UpgradesGoogle rolls out Gemini Deep Think AI, a reasoning model that tests multiple ideas in parallel | TechCrunchGrok Imagine, xAI’s new AI image and video generator, lets you make NSFW content | TechCrunch Meta, Microsoft stocks rise on strong earnings and AI spending boomOpenAI to Establish Stargate Norway With 230MW Data Center - BloombergAnthropic Revenue Pace Nears $5 Billion in Run-Up to Mega Round — The InformationOpenAI Hits $12 Billion Annualized RevenueNoma Security raises $100 million to defend against AI agent vulnerabilities | Ctech OpenAI Just Released Its First Open-Weight Models Since GPT-2Falcon-H1: A Family of Hybrid-Head Language Models Redefining Efficiency and PerformanceMeta CLIP 2: A Worldwide Scaling RecipeBFL and Krea release FLUX.1 Krea: Open image model designed for realism Google’s Newest AI Model Acts like a Satellite to Track Climate Change | WIREDGoogle’s new AI model creates video game worlds in real timeAlphaGo Moment for Model Architecture DiscoveryMETR evaluates Grok 4 Estimating Worst-Case Frontier Risks of Open-Weight LLMsAnthropic's AI 'Vaccine': Train It With Evil to Make It Good - Business InsiderAnthropic unveils 'auditing agents' to test for AI misalignment | VentureBeatOptimizing The Final Output Can Obfuscate CoT (Research Note)Why China isn’t about to leap ahead of the West on computeInside the Summit Where China Pitched Its AI Agenda to the World | WIREDNvidia H20 GPUs reportedly caught up in U.S. Commerce Department's worst export license backlog in 30 years — billions of dollars worth of GPUs and other products in limbo due to staffing cuts, communication issues | Tom's Hardware

Our 218th episode with a summary and discussion of last week's big AI news! Recorded on 07/25/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. In this episode: Timestamps + Links: GitHub Introduces Vibe Coding with Spark: Revolutionizing Intelligent App Development in a Flash - MarkTechPostFigma’s AI app building tool is now available for everyone | The VergeTwo major AI coding tools wiped out user data after making cascading mistakes - Ars TechnicaGoogle's AI Overviews have 2B monthly users, AI Mode 100M in the US and India | TechCrunch Leaked Memo: Anthropic CEO Says the Company Will Pursue Gulf State Investments After AllMira Murati says her startup Thinking Machines will release new product in ‘months’ with ‘significant open source component’Waymo responds to Tesla’s dick joke with a bigger Austin robotaxi map | The Verge MegaScience: Pushing the Frontiers of Post-Training Datasets for Science ReasoningTikTok Researchers Introduce SWE-Perf: The First Benchmark for Repository-Level Code Performance Optimization - MarkTechPost Subliminal Learning: Language models transmit behavioral traits via hidden signals in dataInverse Scaling in Test-Time ComputeScaling Laws for Optimal Data Mixtures White House Unveils America’s AI Action PlanChain of Thought Monitorability: A New and Fragile Opportunity for AI SafetySelf-preservation or Instruction Ambiguity? Examining the Causes of Shutdown ResistancePeople Are Being Involuntarily Committed, Jailed After Spiraling Into "ChatGPT Psychosis"Meta refuses to sign EU’s AI code of practice

Our 217th episode with a summary and discussion of last week's big AI news! Recorded on 07/17/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. In this episode: Timestamps + Links: Tools & Apps OpenAI’s new ChatGPT Agent can control an entire computer and do tasks for youAlibaba-backed Moonshot releases new Kimi AI model that beats ChatGPT, Claude in coding — and it costs lessAmazon targets vibe-coding chaos with new 'Kiro' AI software development tool – GeekWireAnthropic tightens usage limits for Claude Code – without telling usersMistral's Le Chat chatbot gets a productivity push with new ‘deep research' mode | TechCrunchI spent 24 hours flirting with Elon Musk’s AI girlfriendUber is close to completing its quest to become the ultimate robotaxi app | The Verge Applications & Business OpenAI’s Windsurf deal is off — and Windsurf’s CEO is going to Google | The VergeCognition, maker of the AI coding agent Devin, acquires Windsurf | TechCrunchAnthropic hired back two of its employees — just two weeks after they left for a competitor. | The VergeAnother High-Profile OpenAI Researcher Departs for Meta | WIREDMeta Hires Two Key Apple (AAPL) AI Experts After Poaching Their Boss - BloombergMira Murati's Thinking Machines Lab is worth $12B in seed round | TechCrunchLovable becomes a unicorn with $200M Series A just 8 months after launch | TechCrunchSpaceX commits $2 billion to xAI as Musk steps up AI ambitions: Report | World News - Business Standard Research & Advancements A former OpenAI engineer describes what it’s really like to work there | TechCrunchReasoning or Memorization? Unreliable Results of Reinforcement Learning Due to Data Contamination Policy & Safety Anthropic, Google, OpenAI, xAI granted up to $200 million from DoDCalifornia State Senator Scott Wiener Pushes Bill to Regulate AI Companies - BloombergAI 'Nudify' Websites Are Raking in Millions of Dollars | WIREDInside ICE’s Supercharged Facial Recognition App of 200 Million Images Synthetic Media & Art Video game actors' strike officially ends after AI deal

Our 216th episode with a summary and discussion of last week's big AI news! Recorded on 07/11/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. In this episode: Timestamps + Links: Tools & Apps Elon Musk's xAI launches Grok 4 alongside a $300 monthly subscription | TechCrunchElon Musk’s AI chatbot is suddenly posting antisemitic tropesPerplexity launches Comet, an AI-powered web browser | TechCrunchOpenAI is reportedly releasing an AI browser in the coming weeks | TechCrunchReplit Launches New Feature for its Agent, CEO Calls it ‘Deep Research for Coding’Cursor launches a web app to manage AI coding agentsCursor apologizes for unclear pricing changes that upset users | TechCrunch Applications & Business Lovable on track to raise $150M at $2B valuationAmazon built a massive AI supercluster for Anthropic called Project Rainier – here's what we know so farElon Musk confirms xAI is buying an overseas power plant and shipping the whole thing to the U.S. to power its new data center — 1 million AI GPUs and up to 2 Gigawatts of power under one roof, equivalent to powering 1.9 million homesMicrosoft's own AI chip delayed six months in major setback — in-house chip now reportedly expected in 2026, but won't hold a candle to Nvidia BlackwellIlya Sutskever becomes CEO of Safe Superintelligence after Meta poached Daniel GrossOpenAI’s Stock Compensation Reflect Steep Costs of Talent Wars Projects & Open Source Hugging Face Releases SmolLM3: A 3B Long-Context, Multilingual Reasoning Model - MarkTechPostKimi K2: Open Agentic IntelligenceKyutai Releases 2B Parameter Streaming Text-to-Speech TTS with 220ms Latency and 2.5M Hours of Training Research & Advancements Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM ReasoningMeasuring the Impact of Early-2025 AI on Experienced Open-Source Developer ProductivityMitigating Goal Misgeneralization with Minimax RegretCorrelated Errors in Large Language ModelsWhat skills does SWE-bench Verified evaluate? Policy & Safety Evaluating Frontier Models for Stealth and Situational AwarenessWhen Chain of Thought is Necessary, Language Models Struggle to Evade MonitorsWhy Do Some Language Models Fake Alignment While Others Don't?Positive review only': Researchers hide AI prompts in papersGoogle faces EU antitrust complaint over AI OverviewsThe transfer of user data by DeepSeek to China is unlawful': Germany calls for Google and Apple to remove the AI app from their storesVirology Capabilities Test (VCT): A Multimodal Virology Q&A Benchmark

Our 215th episode with a summary and discussion of last week's big AI news! Recorded on 07/04/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. In this episode: Timestamps + Links: Tools & Apps Cloudflare Introduces Default Blocking of A.I. Data ScrapersRunway is going to let people generate video games with AIGoogle embraces AI in the classroom with new Gemini tools for educators, chatbots for students, and moreNo one likes meetings. They’re sending their AI note takers instead.Google launches Doppl, a new app that lets you visualize how an outfit might look on youGoogle's Imagen 4 text-to-image model promises 'significantly improved' boring images Applications & Business Mark Zuckerberg announces his AI ‘superintelligence’ super-groupAnthropic Revenue Hits $4 Billion Annual Pace as Competition With Cursor IntensifiesAs job losses loom, Anthropic launches program to track AI’s economic falloutOpenAI says it has no plan to use Google's in-house chipNvidia stakes new startup that flips script on data center powerTSMC Arizona Chips Are Reportedly Being Flown Back to Taiwan For Packaging; U.S. Semiconductor Supply Chain Still Remains Dependent on Taiwan Projects & Open Source Baidu releases open source model family ERNIE 4.5Tencent Open Sources Hunyuan-A13B: A 13B Active Parameter MoE Model with Dual-Mode Reasoning and 256K ContextTogether AI Releases DeepSWE: A Fully Open-Source RL-Trained Coding Agent Based on Qwen3-32B and Achieves 59% on SWEBenchGLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement LearningDiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation Research & Advancements Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree SearchThe Automated LLM Speedrunning Benchmark: Reproducing NanoGPT ImprovementsClaude 4 Opus and Sonnet reach 50%-time-horizon point estimates of about 80 and 65 minutes, respectivelyPerformance Prediction for Large Systems via Text-to-Text RegressionDoes Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM ReasoningCorrelated Errors in Large Language Models Policy & Safety Forecasting Biosecurity Risks from LLMsAI Task Length Horizons in Offensive CybersecurityInside Tech's Risky Gamble to Kill State AI Regulations for a DecadeDenmark to tackle deepfakes by giving people copyright to their own features

Our 214th episode with a summary and discussion of last week's big AI news! Recorded on 06/27/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. In this episode: Timestamps + Links: Tools & Apps Google is bringing Gemini CLI to developers’ terminalsAnthropic now lets you make apps right from its Claude AI chatbot Applications & Business Sam Altman takes his ‘io’ trademark battle publicHuawei Matebook Contains Kirin X90, using SMIC 7nm (N+2) TechnologyAMD deploys its first Ultra Ethernet ready network card — Pensando Pollara provides up to 400 Gbps performanceAmazon joins the big nuclear party, buying 1.92 GW for AWSNvidia goes nuclear — company joins Bill Gates in backing TerraPower, a company building nuclear reactors for powering data centersMira Murati’s Thinking Machines Lab closes on $2B at $10B valuationMeta hires key OpenAI researcher to work on AI reasoning models Research & Advancements Google’s new AI will help researchers understand how our genes workDirect Reasoning Optimization: LLMs Can Reward And Refine Their Own Reasoning for Open-Ended TasksFarseer: A Refined Scaling Law in Large Language ModelsLLM-First Search: Self-Guided Exploration of the Solution Space Policy & Safety Unsupervised Elicitation of Language ModelsTaiwan Imposes Technology Export Controls on Huawei, SMICYour Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task Synthetic Media & Art Judge Rejects Authors’ Claim That Meta AI Training Violated CopyrightsGetty drops key copyright claims against Stability AI, but UK lawsuit continues

Our 213nd episode with a summary and discussion of last week's big AI news! Recorded on 06/21/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. In this episode: Timestamps + Links: Midjourney launches its first AI video generation model, V1Google’s Gemini AI family updated with stable 2.5 Pro, super-efficient 2.5 Flash-LiteGoogle’s AI Mode can now have back-and-forth voice conversationsYouTube to Add Google’s Veo 3 to Shorts in Move That Could Turbocharge AI on the Video Platform The ‘OpenAI Files’ will help you understand how Sam Altman’s company worksOpenAI drops Scale AI as a data provider following Meta dealAmazon’s Zoox opens its first major robotaxi production facility LiveCodeBench Pro: How Do Olympiad Medalists Judge LLMs in Competitive Programming?AbstentionBench: Reasoning LLMs Fail on Unanswerable QuestionsMiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention Scaling Laws of Motion Forecasting and Planning -- A Technical Report Universal Jailbreak Suffixes Are Strong Attention HijackersOpenAI found features in AI models that correspond to different ‘personas’OpenAI wins $200 million U.S. defense contract

Our 212th episode with a summary and discussion of last week's big AI news! Recorded on 06/33/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. In this episode: Timestamps + Links: OpenAI adds o3 Pro to ChatGPT and drops o3 price by 80 per cent, but open-source AI is delayedCursor AI editor hits 1.0 milestone, including BugBot and high-risk background agentsMistral releases a pair of AI reasoning modelsElevenlabs' Eleven v3 lets AI voices whisper, laugh and express emotions naturallyByteDance's Seedance 1.0 is trading blows with Google's Veo 3Google Reveals $20 AI Pro Plan With Veo 3 Fast Video Generator For Budget CreatorsOpenAI and DeepMind are losing engineers to Anthropic in a one-sided talent warOpenAI slams court order to save all ChatGPT logs, including deleted chatsNvidia’s Biggest Chinese Rival Huawei Struggles to Win at HomeHuawei Expected to Break Semiconductor Barriers with Development of High-End 3nm GAA Chips; Tape-Out by 2026TSMC’s 1.4nm Process, Also Called Angstrom, Will Make Even The Most Lucrative Clients Think Twice When Placing Orders, With An Estimate Claiming That Each Wafer Will Cost $45,000Mistral AI Launches Mistral Compute To Replace Cloud Providers from US, ChinaProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language ModelsKinetics: Rethinking Test-Time Scaling LawsThe Surprising Effectiveness of Negative Reinforcement in LLM ReasoningPredicting Empirical AI Research Outcomes with Language ModelsEXP-Bench: Can AI Conduct AI Research Experiments?Large Language Models Often Know When They Are Being EvaluatedBeyond Induction Heads: In-Context Meta Learning Induces Multi-Phase Circuit EmergenceExclusive: New Microsoft Copilot flaw signals broader risk of AI agents being hacked—‘I would be terrified’Claude Gov Models for U.S. National Security CustomersDisney And NBCUniversal Sue AI Company Midjourney For Copyright InfringementAMC Networks is teaming up with AI company Runway

Our 211th episode with a summary and discussion of last week's big AI news! Recorded on 05/31/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Join our Discord here! https://discord.gg/nTyezGSKwP In this episode: Timestamps + Links: Anthropic launches a voice mode for ClaudeBlack Forest Labs’ Kontext AI models can edit pics as well as generate themPerplexity’s new tool can generate spreadsheets, dashboards, and morexAI to pay Telegram $300M to integrate Grok into the chat appOpera’s new AI browser promises to write code while you sleepGoogle Photos debuts redesigned editor with new AI tools Top Chinese memory maker expected to abandon DDR4 manufacturing at the behest of BeijingOracle to Buy $40 Billion Worth of Nvidia Chips for First Stargate Data CenterUAE makes ChatGPT Plus subscription free for all residents as part of deal with OpenAINVIDIA Corporation (NVDA) to Launch Cheaper Blackwell AI Chip for China, Says ReportThe New York Times and Amazon ink AI licensing deal DeepSeek’s distilled new R1 AI model can run on a single GPUGoogle Unveils SignGemma, an AI Model That Can Translate Sign Language Into Spoken TextOpen-sourcing circuit tracing toolsHugging Face unveils two new humanoid robots PANGU PRO MOE: MIXTURE OF GROUPED EXPERTS FOR EFFICIENT SPARSITYDataRater: Meta-Learned Dataset CurationIncorrect Baseline Evaluations Call into Question Recent LLM-RL Claims Maximizing Confidence Alone Improves ReasoningGuided by Gut: Efficient Test-Time Scaling with Reinforced Intrinsic ConfidenceOne RL to See Them AllEfficient Reinforcement Finetuning via Adaptive Curriculum Learning Trump's 'Big Beautiful Bill' could ban states from regulating AI for a decadeResearchers claim ChatGPT o3 bypassed shutdown in controlled testAnthropic’s new AI model turns to blackmail when engineers try to take it offlineAnthropic Faces Backlash As Claude 4 Opus Can Autonomously Alert AuthoritiesClaude helps users make bioweaponsThe Claude 4 System Card is a Wild Read

Our 210th episode with a summary and discussion of last week's big AI news! Recorded on 05/23/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Join our Discord here! https://discord.gg/nTyezGSKwP In this episode: Timestamps + Links: Anthropic’s new Claude 4 AI models can reason over many steps Google Unveils A.I. Chatbot, Signaling a New Era for Search Google rolls out Project Mariner, its web-browsing AI agent Veo 3 can generate videos — and soundtracks to go along with them Imagen 4 is Google’s newest AI image generator Google Meet is getting real-time speech translation Google’s new Jules AI agent will help developers fix buggy code GitHub’s new AI coding agent can fix bugs for you Mistral’s new Devstral model was designed for codingOpenAI Unites With Jony Ive in $6.5 Billion Deal to Create A.I. Devices OpenAI’s planned data center in Abu Dhabi would be bigger than Monaco LM Arena, the organization behind popular AI leaderboards, lands $100M Nvidia CEO says next chip after H20 for China won't be from Hopper series Google’s Gemini AI app has 400M monthly active users AI Servers: End demand intact, but rising gap between upstream build and system production (2025.5.18) Meta Is Delaying the Rollout of Its Flagship AI ModelGemini Diffusion Chain-of-Model Learning for Language Model Seek in the Dark: Reasoning via Test-Time Instance-Level Policy Gradient in Latent Space Two Experts Are All You Need for Steering Thinking: Reinforcing Cognitive Effort in MoE Reasoning Models Without Additional Training Lessons from Defending Gemini Against Indirect Prompt Injections How Fast Can Algorithms Advance Capabilities? Reinforcement Learning Finetunes Small Subnetworks in Large Language ModelsExclusive: What OpenAI Told California's Attorney GeneralActivating AI Safety Level 3 Protections

Our 209th episode with a summary and discussion of last week's big AI news! Recorded on 05/16/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Join our Discord here! https://discord.gg/nTyezGSKwP In this episode: Timestamps + Links: OpenAI says non-profit will remain in control after backlashMicrosoft Moves to Protect Its Turf as OpenAI Turns Into RivalTSMC’s 2nm Process Said to Witness ‘Unprecedented’ Demand, Exceeding 3nm Due to Interest from Apple, NVIDIA, AMD, & Many OthersNVIDIA’s Global Headquarters Will Be In Taiwan, With CEO Huang Set To Announce Site Next Week, Says ReportCoreWeave in Talks for $1.5 Billion Debt Deal 6 Weeks After IPOThe Day Grok Told Everyone About ‘White Genocide’Figma releases new AI-powered tools for creating sites, app prototypes, and marketing assetsGoogle’s bringing Gemini to your car with Android AutoGoogle debuts an updated Gemini 2.5 Pro AI model ahead of I/OHugging Face releases a free Operator-like agentic AI toolStability AI releases an audio-generating model that can run on smartphonesFreepik releases an ‘open’ AI image generator trained on licensed dataAM-Thinking-v1: Advancing the Frontier of Reasoning at 32B ScaleBLIP3-o: A Family of Fully Open Unified Multimodal Models-Architecture, Training and DatasetDeepMind claims its newest AI tool is a whiz at math and science problemsAbsolute Zero: Reinforced Self-play Reasoning with Zero DataHow far can reasoning models scale?HealthBench: Evaluating Large Language Models Towards Improved Human HealthTrump administration officially rescinds Biden’s AI diffusion rulesTrump’s Mideast Visit Opens Floodgate of AI Deals Led by NvidiaScaling Laws For Scalable OversightOpenAI pledges to publish AI safety test results more often

Our 208th episode with a summary and discussion of last week's big AI news! Recorded on 05/02/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Join our Discord here! https://discord.gg/nTyezGSKwP In this episode: Timestamps + Links: Tools & Apps Anthropic lets users connect more apps to ClaudeOpenAI undoes its glaze-heavy ChatGPT update Baidu ERNIE X1 and 4.5 Turbo boast high performance at low cost Adobe adds more image generators to its growing AI family OpenAI makes its upgraded image generator available to developers xAI’s Grok chatbot can now ‘see’ the world around it Applications & Business: Thinking Machines Lab CEO Has Unusual Control in Andreessen-Led Deal Chip war heats up: Huawei 910C emerges as China’s answer to US export bans Huawei to Test New AI Chip ByteDance, Alibaba and Tencent stockpile billions worth of Nvidia chips Speculation mounts that Musk will raise tens of billions for AI supercomputer with 1 million GPUs: Report Projects & Open Source: Alibaba unveils Qwen 3, a family of ‘hybrid’ AI reasoning models Intellect-2 BitNet b1.58 2B4T Technical Report Meta AI Introduces Perception Encoder: A Large-Scale Vision Encoder that Excels Across Several Vision Tasks for Images and Video Research & Advancements: The Leaderboard Illusion Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? Reinforcement Learning for Reasoning in Large Language Models with One Training Example Sleep-time Compute: Beyond Inference Scaling at Test-time Policy & Safety: Every AI Datacenter Is Vulnerable to Chinese Espionage, Report Says OpenAI preparedness framework update Detecting and Countering Malicious Uses of Claude: March 2025 Chinese AI Will Match America's

Our 207th episode with a summary and discussion of last week's big AI news! Recorded on 04/14/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Join our Discord here! https://discord.gg/nTyezGSKwP In this episode: Timestamps + Links: OpenAI’s new GPT-4.1 AI models focus on codingChatGPT will now remember your old conversationsGoogle’s newest Gemini AI model focuses on efficiencyElon Musk’s AI company, xAI, launches an API for Grok 3Canva is now in the coding and spreadsheet businessMeta’s vanilla Maverick AI model ranks below rivals on a popular chat benchmarkIronwood: The first Google TPU for the age of inferenceAnthropic rolls out a $200-per-month Claude subscriptionOpenAI co-founder Ilya Sutskever’s Safe Superintelligence reportedly valued at $32BMira Murati’s AI startup gains prominent ex-OpenAI advisersHugging Face buys a humanoid robotics startupStargate developer Crusoe could spend $3.5 billion on a Texas data center. Most of it will be tax-free.OpenAI Open Sources BrowseComp: A New Benchmark for Measuring the Ability for AI Agents to Browse the WebSample, Don't Search: Rethinking Test-Time Alignment for Language ModelsConcise Reasoning via Reinforcement LearningGoing beyond open data – increasing transparency and trust in language models with OLMoTraceIndependent evaluations of Grok-3 and Grok-3 mini on our suite of benchmarksOpenAI countersues Elon Musk, calls for enjoinment from ‘further unlawful and unfair action’OpenAI slashes AI model safety testing timeEx-OpenAI staffers file amicus brief opposing the company’s for-profit transitionAccess to future AI models in OpenAI’s API may require a verified IDMeta whistleblower claims tech giant built $18 billion business by aiding China in AI race and undermining U.S. national security

Our 206th episode with a summary and discussion of last week's big AI news! Recorded on 04/07/2025 Try out the Astrocade demo here! https://www.astrocade.com/ Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Join our Discord here! https://discord.gg/nTyezGSKwP In this episode: Timestamps + Links: Tools & Apps Meta releases Llama 4, a new crop of flagship AI modelsAmazon unveils Nova Act, an AI agent that can control a web browserAlibaba Preparing for Flagship AI Model Release as Soon as AprilRunway releases an impressive new video-generating AI modelAdobe launches Premiere Pro’s generative AI video extenderOpenAI prepares reasoning slider and memory update for ChatGPT users Applications & Business Nvidia H20 Chips: $16 Billion Orders from ByteDance, Alibaba, and TencentElon Musk sells X for $33 billion to his own AI startup company xAISoftBank dethroned Microsoft as OpenAI's largest investor, pushing the ChatGPT maker's market cap to $300 billion — but reportedly buried itself in debtDeepMind is holding back release of AI research to give Google an edgeSMIC Is Rumored To Complete 5nm Chip Development By 2025; Costs Could Be Up To 50 Percent Higher Than TSMC’s Version Due To The Use Of Older-Generation EquipmentGoogle-backed Isomorphic Labs raises $600m to advance AI drug discovery Research & Advancements PaperBench: Evaluating AI's Ability to Replicate AI ResearchCrossing the Reward Bridge: Expanding RL with Verifiable Rewards Across Diverse DomainsInference-Time Scaling for Complex Tasks: Where We Stand and What Lies AheadOvertrained Language Models Are Harder to Fine-Tune Policy & Safety Taking a responsible path to AGIThis A.I. Forecast Predicts Storms AheadThe Secrets and Misdirection Behind Sam Altman’s Firing From OpenAI

Our 205th episode with a summary and discussion of last week's big AI news! Recorded on 03/28/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Join our Discord here! https://discord.gg/nTyezGSKwP In this episode: Timestamps + Links: Gemini 2.5: Our most intelligent AI modelOpenAI rolls out image generation powered by GPT-4o to ChatGPTIdeogram presents version 3.0 of its AI image generation systemNew Reve Image Generator Beats AI Art Heavyweights MidJourney and Flux at a Penny Per ImageAlibaba Releases Qwen2.5 Omni, Adds Voice and Video Modes to Qwen ChatThe official version of Tencent's Hunyuan Deep Thinking Model T1 is here, with fast articulation, instant responses, and a decoding speed increase of 2 times OpenAI Close to Finalizing $40 Billion SoftBank-Led FundingOpenAI reshuffles leadership as Sam Altman pivots to technical focusNvidia shows off Rubin Ultra with 600,000-Watt Kyber racks and infrastructure, coming in 2027China's SiCarrier emerges as challenger to ASML, other chip tool titansPony.ai wins first permit for fully driverless taxi operation in the center of China’s Silicon Valley A new, challenging AGI test stumps most AI modelsChallenging the Boundaries of Reasoning: An Olympiad-Level Math Benchmark for Large Language ModelsWan: Open and Advanced Large-Scale Video Generative ModelsDeepSeek V3-0324 tops non-reasoning AI models in open-source firstOpenAI adopts rival Anthropic’s standard for connecting AI models to data Anthropic can now track the bizarre inner workings of a large language modelChain-of-Tools: Utilizing Massive Unseen Tools in the CoT Reasoning of Frozen Language ModelsInside-Out: Hidden Factual Knowledge in LLMsSakana AI super-powers AI reasoning using Japan’s own Sudoku Puzzles Senator Wiener Introduces Legislation to Protect AI Whistleblowers & Boost Responsible AI DevelopmentNVIDIA & Other Tech Giants Demand Trump Administration To Reconsider “AI Diffusion” Policy Which Is Set To Be Effective By May 15U.S. blacklists over 50 Chinese companies in bid to curb Beijing's AI, chip capabilitiesNetflix’s Reed Hastings Gives $50 Million to Bowdoin for A.I. Program Judge allows 'New York Times' copyright case against OpenAI to go forwardJudge rules that AI can continue training on copyrighted lyrics, for now

Our 204th episode with a summary and discussion of last week's big AI news! Recorded on 03/21/2025 Hosted by Andrey Kurenkov and Jeremie Harris. Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai Read out our text newsletter and comment on the podcast at https://lastweekin.ai/. Join our Discord here! https://discord.gg/nTyezGSKwP In this episode: Timestamps + Links: Baidu launches two new versions of its AI model ErnieOpenAI Unveils New Audio Models to Make AI Agents Sound More Human Than EverOpenAI’s o1-pro is the company’s most expensive AI model yetGoogle brings a ‘canvas’ feature to Gemini, plus Audio OverviewAnthropic adds web search to its Claude chatbotxAI launches an API for generating images Nvidia announces Rubin GPUs in 2026, Rubin Ultra in 2027, Feynman also added to roadmapM3 Ultra Runs DeepSeek R1 With 671 Billion Parameters Using 448GB Of Unified Memory, Delivering High Bandwidth Performance At Under 200W Power Consumption, With No Need For A Multi-GPU SetupIntel reaches 'exciting milestone' for 18A 1.8nm-class wafers with first run at Arizona fabElon Musk’s AI company, xAI, acquires a generative AI video startupTencent Reportedly Makes Massive NVIDIA H20 Chip Purchase for WeChat’s DeepSeek Integration Anthropic’s Not-So-Secret Weapon That’s Giving Agents a BoostMistral AI drops new open-source model that outperforms GPT-4o Mini with fraction of parametersEXAONE Deep: Reasoning Enhanced Language Models Sample, Scrutinize and Scale: Effective Inference-Time Search by Scaling VerificationBlock Diffusion: Interpolating Between Autoregressive and Diffusion Language ModelsCommunication-Efficient Language Model Training Scales Reliably and Robustly: Scaling Laws for DiLoCoTransformers without NormalizationMeasuring AI Ability to Complete Long TasksHCAST: Human-Calibrated Autonomy Software Tasks Announcing Zochi, an Intology ProjectDeepSeek, a National Treasure in China, is Now Being Closely GuardedClaude Sonnet 3.7 (often) knows when it’s in alignment evaluations US appeals court rejects copyrights for AI-generated art lacking 'human' creatorTrump urged by Ben Stiller, Paul McCartney and hundreds of stars to protect AI copyright rules