#156 - OpenAI's Sora, Gemini 1.5, BioMistral, V-JEPA, AI Task Force, Fun!

Our 157th episode with a summary and discussion of last week's big AI news! Check out our sponsor, the SuperDataScience podcast. You can listen to SDS across all major podcasting platforms (e.g., Spotify, Apple Podcasts, Google Podcasts) plus there’s a video version on YouTube. Bonus plug: also check out this new book by Stanford AI expert, bestselling author, and Last Week in AI supporter Jerry Kaplan! Generative Artificial Intelligence: What Everyone Needs to Know Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai Timestamps + links: (00:00:00) Intro / Banter Tools & Apps (00:04:55) Google apologizes for ‘missing the mark’ after Gemini generated racially diverse Nazis (00:15:07) Stability announces Stable Diffusion 3, a next-gen AI image generator (00:19:04) Mistral AI releases new model to rival GPT-4 and its own chat assistant (00:24:56) Windows just got its own Magic Eraser to AI-modify your photos (00:25:47) Adobe previews new cutting-edge generative AI tools for crafting and editing custom audio (00:27:47) AI video wars heat up as Pika adds Lip Sync powered by ElevenLabs Applications & Business (00:30:09) Microsoft Strikes Deal with France’s Mistral AI (00:33:45) Figure Raises $675M at $2.6B Valuation and Signs Collaboration Agreement with OpenAI (00:37:05) Nvidia posts record revenue up 265% on booming AI business (00:39:54) MediaTek’s latest chipsets are now ‘optimized’ for Gemini Nano (00:41:28) Tumblr’s owner is striking deals with OpenAI and Midjourney for training data, says report (00:43:45) Mistral AI models coming soon to Amazon Bedrock Projects & Open Source (00:44:34) Generative AI Startup Mistral Releases Free ‘Open-Source’ 7.3B Parameter LLM (00:46:57) Google Delves Deeper Into Open Source with Launch of Gemma AI Model (00:51:10) Microsoft releases its internal generative AI red teaming tool to the public (00:54:25) Introducing Phind-70B – closing the code quality gap with GPT-4 Turbo while running 4x faster Research & Advancements (00:57:51) Genie: Generative Interactive Environments (01:07:08) Griffin: Mixing Gated Linear Recurrences with Local Attention for Efficient Language Models (01:15:16) Quantum Circuit Optimization with AlphaTensor (01:20:56) Back to Basics: Revisiting REINFORCE Style Optimization for Learning from Human Feedback in LLMs (01:22:10) Repetition Improves Language Model Embeddings Policy & Safety (01:25:45) AI Warfare Is Already Here (01:29:36) Man admits to paying magician $150 to create anti-Biden robocall (01:32:45) Google DeepMind forms a new org focused on AI safety (01:34:53)  Facebook whistleblower, AI godfather join hundreds calling for deepfake regulation US regulators investigate whether OpenAI investors were misled, say reports (01:36:51)  Users Say Microsoft's AI Has Alternate Personality as Godlike AGI That Demands to Be Worshipped Synthetic Media & Art (01:40:23) The Intercept, Raw Story, and AlterNet sue OpenAI and Microsoft (01:41:15) A viral photo of a guy smoking in McDonald's is completely fake — and of course made by AI Fun! (01:42:54) Impossible AI Food See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

3 Maalis 20241h 44min

Our 155th episode with a summary and discussion of last week's big AI news! Correction: Andrey said CLIP came out with DALL-E 2; it came out alongside the first DALL-E. Check out our sponsor, the SuperDataScience podcast. You can listen to SDS across all major podcasting platforms (e.g., Spotify, Apple Podcasts, Google Podcasts) plus there’s a video version on YouTube. Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai Timestamps + links: (00:00:00) Intro / Banter Tools & Apps (00:03:45) ChatGPT is getting ‘memory’ to remember who you are and what you like (00:08:20) Reka Flash: An Efficient and Capable Multimodal Language Model (00:14:22) Say What? Chat With RTX Brings Custom Chatbot to NVIDIA RTX AI PCs Applications & Business (00:17:49) Sam Altman Seeks Trillions of Dollars to Reshape Business of Chips and AI (00:26:12) Report: China’s SMIC to begin production of 5nm chips for Huawei (00:30:32) A crowd destroyed a driverless Waymo car in San Francisco (00:34:18) OpenAI reportedly developing two AI agents to automate entire work processes (00:37:23) Cruise names first chief safety officer following crash and controversy Projects & Open Source (00:39:15) Cohere for AI launches open source LLM for 101 languages (00:43:10 ) BUD-E: ENHANCING AI VOICE ASSISTANTS’ CONVERSATIONAL QUALITY, NATURALNESS AND EMPATHY (00:47:36) EVA-CLIP-18B: Scaling CLIP to 18 Billion Parameters (00:51:30) Introducing Stable Cascade Research & Advancements (00:53:14) SELF-DISCOVER: Large Language Models Self-Compose Reasoning Structures (01:01:30) BlackMamba: Mixture of Experts for State-Space Models (01:05:34) An Interactive Agent Foundation Model (01:09:00) Grandmaster-Level Chess Without Search (01:13:42) More Agents Is All You Need (01:16:04) MusicMagus: Zero-Shot Text-to-Music Editing via Diffusion Models Policy & Safety (01:16:41) Debating with More Persuasive LLMs Leads to More Truthful Answers (01:23:49) In Big Tech’s backyard, California lawmaker unveils landmark AI bill (01:28:06) AI Deployed Nukes 'to Have Peace in the World' in Tense War Simulation (01:31:32) Protesters gather outside OpenAI office, opposing military AI and AGI (01:35:43) AI safeguards can easily be broken, UK Safety Institute finds Synthetic Media & Art (01:39:35) Stability, Midjourney, Runway hit back in AI art lawsuit (01:42:25) AI Companies Take Hit as Judge Says Artists Have “Public Interest” In Pursuing Lawsuits (01:44:42) Outro See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

16 Helmi 20241h 45min

Our 154th episode with a summary and discussion of last week's big AI news! Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai Correction: Andrey mentioned "State space machines", he meant "State space models" Timestamps + links: (00:00:00) Intro / Banter Tools & Apps (00:02:06) Google Releases Gemini, an A.I.-Driven Chatbot and Voice Assistant (00:05:56) Copilot gets a big redesign and a new way to edit your AI-generated images (00:09:40) Arc Search's AI responses launched as an unfettered experience with no guardrails (00:12:40) Brilliant Labs’s Frame glasses serve as multimodal AI assistant (00:15:30) Stability AI launches SVD 1.1, a diffusion model for more consistent AI videos (00:16:18) OpenAI launches ChatGPT app for Apple Vision Pro Applications & Business (00:19:04) A Waymo robotaxi hit a cyclist in San Francisco – here’s what happened (00:23:30) Canon plans to disrupt chipmaking with low-cost “stamp” machine (00:27:33) US industry group calls for multilateral chip export controls to address disadvantage over Korea, other allies (00:30:10) U.S. blocks shipment of 24 Nvidia AI GPUs to China over concerns about self-driving truck company (00:32:24) Nvidia reportedly selects Intel Foundry Services for GPU packaging production — could produce over 300,000 H100 GPUs per month Projects & Open Source (00:37:23) Allen Institute for AI launches open and transparent OLMo large language model (00:42:46) Meet ‘Smaug-72B’: The new king of open-source AI (00:47:02) Introducing Qwen1.5 (00:50:54) Hugging Face launches open source AI assistant maker to rival OpenAI’s custom GPTs (00:53:20)  Apple releases ‘MGIE’, a revolutionary AI model for instruction-based image editing Research & Advancements (00:54:30) Learning Universal Predictors (01:01:00) Can Mamba Learn How to Learn? A Comparative Study on In-Context Learning Tasks (01:04:40) MusicRL: Aligning Music Generation to Human Preferences (01:05:47) FP6-LLM: Efficiently Serving Large Language Models Through FP6-CentricAlgorithm-System Co-Design (01:09:06) AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents (01:12:36) Specialized Language Models with Cheap Inference from Limited Domain Data Policy & Safety (01:13:22) EU’s AI Act passes last big hurdle on the way to adoption (01:17:04) Building an early warning system for LLM-aided biological threat creation (01:23:44) FCC votes to ban scam robocalls that use AI-generated voices (01:24:33) Biden administration names a director of the new AI Safety Institute (01:26:23) OpenAI's GPT-4 finally meets its match: Scots Gaelic smashes safety guardrails Synthetic Media & Art (01:28:04) AI poisoning tool Nightshade received 250,000 downloads in 5 days: ‘beyond anything we imagined’ (01:30:24) Labeling AI-Generated Images on Facebook, Instagram and Threads (01:33:06) OpenAI is adding new watermarks to DALL-E 3 (01:34:38)  Following lawsuit, rep admits “AI” George Carlin was human-written  (01:36:20) Outro See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

11 Helmi 20241h 37min

Our 153rd episode with a summary and discussion of last week's big AI news! Check out our sponsor, the SuperDataScience podcast. You can listen to SDS across all major podcasting platforms (e.g., Spotify, Apple Podcasts, Google Podcasts) plus there’s a video version on YouTube. Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai Timestamps + links: (00:00:00)  Intro / Banter Synthetic Media & Art (00:03:15) Trolls have flooded X with graphic Taylor Swift AI fakes (00:08:30) YouTube Deletes 1,000 Videos of Celebrity AI Scam Ads (00:10:20) Iceland Has Its Own AI George Carlin Moment, Considers Law Against Deepfaking the Dead (00:12:20) Guns N’ Roses share AI-generated video for ‘The General’ Tools & Apps (00:13:30) Microsoft Makes Swift Changes to AI Tool (00:16:38) OpenAI drops prices and fixes ‘lazy’ GPT-4 that refused to work (00:20:40) ChatGPT's new "@GPT" feature paves the way for OpenAI's vision of a universal assistant (00:23:47)  ChatGPT finally has competition — Google Bard with Gemini just matched it with a huge upgrade (00:26:37) Browser alternatives Brave, Arc, add new AI integrations (00:28:37) Baidu's Ernie AI chatbot to power Samsung's new Galaxy S24 smartphones Applications & Business (00:30:40) AI companies lose $190 billion in market cap after Alphabet and Microsoft report (00:34:25) Cruise wasn’t hiding the pedestrian-dragging video from regulators — it just had bad internet (00:40:32) Hugging Face teams up with Google to accelerate open AI development (00:43:33) Elon Musk's xAI eyes $6 billion funding to challenge ChatGPT maker OpenAI (00:47:22)  AI Chip Startup Rebellions Snags Funding to Challenge Nvidia Projects & Open Source (00:50:07) New open source AI coding tool surpasses its inspiration: Google DeepMind’s AlphaCode (00:53:00) Meta’s free Code Llama AI programming tool closes the gap with GPT-4 (00:54:30) DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence Lighting round Research & Advancements (00:55:50) Meta-Prompting: Enhancing Language Models with Task-Agnostic Scaffolding (01:05:15) EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty (01:09:20) Are Vision Transformers More Data Hungry Than Newborn Visual Systems? (01:15:52) Circuit Component Reuse Across Tasks in Transformer Language Models (01:20:30) A ‘Shocking’ Amount of the Web Is Already AI-Translated Trash, Scientists Determine Policy & Safety (01:22:45) Taylor Swift AI images prompt US bill to tackle nonconsensual, sexual deepfakes (01:27:50) OpenAI and Google will be required to notify the government about AI models (01:32:12) The campaign to take down the Biden AI executive order (01:33:06) Rep. Jeff Jackson Introduces Bipartisan CLOUD AI Act to Stop China from Remotely Using American Technology to Build AI Tools (01:39:00) AI Startup ElevenLabs Bans Account Blamed for Biden Audio Deepfake (01:40:06) How the West Can Match Russia in Drone Innovation (01:45:15) Outro See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

4 Helmi 20241h 46min

Our 152nd episode with a summary and discussion of last week's big AI news! Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ Email us your questions and feedback at contact@lastweekin.ai and/or jeremie@gladstone.ai Timestamps + links: (00:00:00) Intro / Banter Tools & Apps (00:02:15) Samsung’s latest Galaxy phones offer live translation over phone calls, texts (00:06:16) The Rabbit R1 will receive live info from Perplexity’s AI ‘answer engine’ (00:10:31) Google is using AI to organize and customize your Chrome browser (00:12:10) Adobe’s new AI-powered Premiere Pro features eradicate boring audio editing tasks (00:13:16) Waymo looks to launch full fleet of robotaxis in LA (00:15:16) Tesla finally releases FSD v12, its last hope for self-driving Applications & Business (00:19:45) OpenAI CEO Sam Altman is still chasing billions to build AI chips (00:25:00) Mark Zuckerberg’s new goal is creating artificial general intelligence (00:32:43) AI voice startup ElevenLabs lands $80M round, launches marketplace of cloned voices (00:35:03) Anthropic’s Margins Raise Questions on AI Startups’ Long-Term Profitability (00:38:00) Cohere in talks to raise as much as $1bn as AI arms race heats up (00:39:40) DOJ and FTC push to investigate Microsoft's OpenAI partnership- Politico (00:41:30)  Figure announces commercial agreement with BMW Manufacturing to bring general purpose robots into automotive production Projects & Open Source (00:43:28) Stability AI unveils smaller, more efficient 1.6B language model as part of ongoing innovation Research & Advancements (00:48:58) AlphaGeometry: An Olympiad-level AI system for geometry (00:54:02) Lumiere: A Space-Time Diffusion Model for Video Generation (00:56:33) ChatQA: Building GPT-4 Level Conversational QA Models (00:59:13) Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model (01:01:28) VMamba: Visual State Space Model (01:03:40) Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data Policy & Safety (01:04:40) OpenAI suspends developer behind Dean Phillips bot (01:06:49) Fake Joe Biden robocall tells New Hampshire Democrats not to vote Tuesday  (01:09:10)  Sharing Fake Nude Images Could Become a Federal Crime Under Proposed Law (01:11:00) San Francisco takes legal action over ‘unsafe,’ ‘disruptive’ self-driving cars (01:13:40) Slew of deepfake video adverts of Sunak on Facebook raises alarm over AI risk to election (01:15:55) AI is the buzz, the big opportunity and the risk to watch among the Davos glitterati Synthetic Media & Art (01:20:27) Test Yourself: Which Faces Were Made by A.I.? (01:23:48) AI models that don’t violate copyright are getting a new certification label See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

28 Tammi 20241h 27min

Our 151st episode with a summary and discussion of last week's big AI news! Check out our sponsor, the SuperDataScience podcast. You can listen to SDS across all major podcasting platforms (e.g., Spotify, Apple Podcasts, Google Podcasts) plus there’s a video version on YouTube. Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ Email us your questions and feedback at contact@lastweekin.ai Timestamps + links: (00:00:00) Intro / Banter Tools & Apps (00:02:55) Microsoft’s new Copilot Pro brings AI-powered Office features to the rest of us (00:04:43) Microsoft Copilot is now using the previously-paywalled GPT-4 Turbo, saving you $20 a month (00:07:02) Amazon launches generative AI tool to answer shoppers’ questions Applications & Business (00:08:07) DeepMind spin-off aims to halve drug discovery times following Big Pharma deals (00:13:16) China planning 1,600-core chips that use an entire wafer — similar to American company Cerebras 'wafer-scale' designs (00:17:12)  ChatGPT will have video functionality and more accuracy in future versions — Sam Altman says GPT-5 will be a big improvement (00:19:38) Huawei Teardown Shows 5nm Chip Made in Taiwan, Not China (00:22:15) OpenAI’s news publisher deals reportedly top out at $5 million a year (00:25:04) Inside AI Unicorn Anthropic’s Unusual $750 Million Fundraise (00:28:46) Microsoft executive Dee Templeton joins OpenAI board - Bloomberg News (00:31:33) OpenAI-backed 1X raises another $100M for the race to humanoid robots Projects & Open Source (00:34:25) Meet LLama.cpp: An Open-Source Machine Learning Library to Run the LLaMA Model Using 4-bit Integer Quantization on a MacBook (00:38:06) LLAMA PRO: Progressive LLaMA with Block Expansion (00:43:48) One of the world’s largest AI training datasets is about to get bigger and ‘substantially better’ Research & Advancements (00:47:52) Towards Conversational Diagnostic AI (00:51:50)  Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training (00:56:27) LLM Augmented LLMs: Expanding Capabilities through Composition (01:02:23) Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models (01:05:16) Clinical predictive models created by AI are accurate but study-specific, researchers find (01:09:01) Secure, Governable Chips Policy & Safety (01:14:01) US companies and Chinese experts engaged in secret diplomacy on AI safety (01:18:28) China Cyberspace Security Association releases first batch of Chinese basic corpus (01:22:33) OpenAI Bans Use of AI Tools for Campaigning, Voter Suppression (01:24:49) OpenAI Quietly Deletes Ban on Using ChatGPT for “Military and Warfare” Synthetic Media & Art (01:27:20) Musicians Set to Begin Contract Negotiations With Studios On AI, Streaming Priorities (01:29:26) Scammy AI-Generated Book Rewrites Are Flooding Amazon (01:33:15) Deepfaked Celebrity Ads Promoting Medicare Scams Run Rampant on YouTube (01:35:15) Outro See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

21 Tammi 20241h 36min

Our 150th episode with a summary and discussion of last week's big AI news! Read out our text newsletter and comment on the podcast at https://lastweekin.ai/ Email us your questions and feedback at contact@lastweekin.ai Timestamps + links: (00:00:00) Intro / Banter Tools & Apps (00:02:49) OpenAI’s custom GPT Store is now open for business (00:06:52) OpenAI debuts ChatGPT subscription aimed at small teams Lighting round (00:09:18) The Rabbit R1 is an AI-powered gadget that can use your apps for you (00:11:18) Amazon’s Alexa gets new generative AI-powered experiences (00:12:55) Google appears to be working on an ‘advanced’ version of Bard that you have to pay for Applications & Business (00:14:24) Nvidia's newest chips are designed to run AI at home as competition from Intel, AMD looms (00:16:52) Valve opens the door to more Steam games developed with AI Lighting round (00:19:16) AI-powered search engine Perplexity AI, now valued at $520M, raises $73.6M (00:21:33) Waymo will start testing robotaxis on Phoenix highways (00:23:24) Getty and Nvidia bring generative AI to stock photos Research & Advancements (00:25:10) DeepMind is trying to train robots for real-world activities (00:29:03) MoE-Mamba: Efficient Selective State Space Models with Mixture of Experts Lighting round (00:33:18) PIXART-δ: Fast and Controllable Image Generation with Latent Consistency Models (00:35:20)  InseRF: Text-Driven Generative Object Insertion in Neural 3D Scenes (00:37:06) The Impact of Reasoning Step Length on Large Language Models (00:39:33) Mixtral of Experts Policy & Safety (00:40:38) Meta and OpenAI have spawned a wave of AI sex companions—and some of them are children (00:46:30) ‘Impossible’ to create AI tools like ChatGPT without copyrighted material, OpenAI says Lighting round (00:49:52) Judges in England and Wales are given cautious approval to use AI in writing legal opinions (00:51:43) Hallucinating Law: Legal Mistakes with Large Language Models are Pervasive Synthetic Media & Art (00:55:09) A list going viral reveals famous artists whose work was used to train AI generator (00:58:28) YouTube is cracking down on AI-generated true crime deepfakes Lighting round (01:00:27) AI-Generated George Carlin Comedy Special Slammed by Comedian’s Daughter (01:01:56) SAG-AFTRA Signs Deal With Voiceover Studio for AI Use in Video Games (01:03:26) Early Mickey Mouse is now in the public domain—and AI is already on the case See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

14 Tammi 20241h 5min