ThursdAI - The top AI news from the past week

From Weights & Biases, Join AI Evangelist Alex Volkov and a panel of experts to cover everything important that happened in the world of AI from the past week
  • 0.0 (0)
  • TECH NEWS
  • UPDATED WEEKLY
ThursdAI - The top AI news from the past week

Every ThursdAI, Alex Volkov hosts a panel of experts, ai engineers, data scientists and prompt spellcasters on twitter spaces, as we discuss everything major and important that happened in the world of AI for the past week. Topics include LLMs, Open source, New capabilities, OpenAI, competitors in AI space, new LLM models, AI art and diffusion aspects and much more. sub.thursdai.news

  1. 5 DAYS AGO

    📆 ThursdAI - June 12 - Meta’s $15B ScaleAI Power Play, OpenAI’s o3-pro & 90% Price Drop!

    Hey folks, this is Alex, finally back home! This week was full of crazy AI news, both model related but also shifts in the AI landscape and big companies, with Zuck going all in on scale & execu-hiring Alex Wang for a crazy $14B dollars. OpenAI meanwhile, maybe received a new shipment of GPUs? Otherwise, it’s hard to explain how they have dropped the o3 price by 80%, while also shipping o3-pro (in chat and API). Apple was also featured in today’s episode, but more so for the lack of AI news, completely delaying the “very personalized private Siri powered by Apple Intelligence” during WWDC25 this week. We had 2 guests on the show this week, Stefania Druga and Eric Provencher (who builds RepoPrompt). Stefania helped me cover the AI Engineer conference we all went to last week, and shared some cool Science CoPilot stuff she’s working on, while Eric is the GOTO guy for O3-pro helped us understand what this model is great for! As always, TL;DR and show notes at the bottom, video for those who prefer watching is attached below, let’s dive in! Big Companies LLMs & APIs Let’s start with big companies, because the landscape has shifted, new top reasoner models dropped and some huge companies didn’t deliver this week! Zuck goes all in on SuperIntelligence - Meta’s $14B stake in ScaleAI and Alex Wang This may be the most consequential piece of AI news today. Fresh from the dissapointing results of LLama 4, reports of top researchers leaving the Llama team, many have decided to exclude Meta from the AI race. We have a saying at ThursdAI, don’t bet against Zuck! Zuck decided to spend a lot of money (nearly 20% of their reported $65B investment in AI infrastructure) to get a 49% stake in Scale AI and bring Alex Wang it’s (now former) CEO to lead the new Superintelligence team at Meta. For folks who are not familiar with Scale, it’s a massive company in providing human annotated data services to all the big AI labs, Google, OpenAI, Microsoft, Anthropic.. all of them really. Alex Wang, is the youngest self made billionaire because of it, and now Zuck not only has access to all their expertise, but also to a very impressive AI persona, who could help revive the excitement about Meta’s AI efforts, help recruit the best researchers, and lead the way inside Meta. Wang is also an outspoken China hawk who spends as much time in congressional hearings as in Slack, so the geopolitics here are … spicy. Meta just stapled itself to the biggest annotation funnel on Earth, hired away Google’s Jack Rae (who was on the pod just last week, shipping for Google!) for brainy model alignment, and started waving seven-to-nine-figure comp packages at every researcher with “Transformer” in their citation list. Whatever disappointment you felt over Llama-4’s muted debut, Zuck clearly felt it too—and responded like a founder who still controls every voting share. OpenAI’s Game-Changer: o3 Price Slash & o3-pro launches to top the intelligence leaderboards! Meanwhile OpenAI dropping not one, but two mind-blowing updates. First, they’ve slashed the price of o3—their premium reasoning model—by a staggering 80%. We’re talking from $40/$10 per million tokens down to just $8/$2. That’s right, folks, it’s now in the same league as Claude Sonnet cost-wise, making top-tier intelligence dirt cheap. I remember when a price drop of 80% after a year got us excited; now it’s 80% in just four months with zero quality loss. They’ve confirmed it’s the full o3 model—no distillation or quantization here. How are they pulling this off? I’m guessing someone got a shipment of shiny new H200s from Jensen! And just when you thought it couldn’t get better, OpenAI rolled out o3-pro, their highest intelligence offering yet. Available for pro and team accounts, and via API (87% cheaper than o1-pro, by the way), this model—or consortium of models—is a beast. It’s topping charts on Artificial Analysis, barely edging out Gemini 2.5 as the new king. Benchmarks are insane: 93% on AIME 2024 (state-of-the-art territory), 84% on GPQA Diamond, and nearing a 3000 ELO score on competition coding. Human preference tests show 64-66% of folks prefer o3-pro for clarity and comprehensiveness across tasks like scientific analysis and personal writing. I’ve been playing with it myself, and the way o3-pro handles long context and tough problems is unreal. As my friend Eric Provencher (creator of RepoPrompt) shared on the show, it’s surgical—perfect for big refactors and bug diagnosis in coding. It’s got all the tools o3 has—web search, image analysis, memory personalization—and you can run it in background mode via API for async tasks. Sure, it’s slower due to deep reasoning (no streaming thought tokens), but the consistency and depth? Worth it. Oh, and funny story—I was prepping a talk for Hamel Hussain’s evals course, with a slide saying “don’t use large reasoning models if budget’s tight.” The day before, this price drop hits, and I’m scrambling to update everything. That’s AI pace for ya! Apple WWDC: Where’s the Smarter Siri? Oh Apple. Sweet, sweet Apple. Remember all those Bella Ramsey ads promising a personalized Siri that knows everything about you? Well, Craig Federighi opened WWDC by basically saying "Yeah, about that smart Siri... she's not coming. Don't wait up." Instead, we got: * AI that can combine emojis (revolutionary! 🙄) * Live translation (actually cool) * Direct API access to on-device models (very cool for developers) * Liquid glass UI (pretty but... where's the intelligence?) The kicker? Apple released a paper called "The Illusion of Thinking" right before WWDC, basically arguing that AI reasoning models hit hard complexity ceilings. Some saw this as Apple making excuses for why they can't ship competitive AI. The timing was... interesting. During our recording, Nisten's Siri literally woke up randomly when we were complaining about how dumb it still is. After a decade, it's the same Siri. That moment was pure comedy gold. This Week's Buzz Our premium conference Fully Connected is happening June 17-18 in San Francisco! Use promo code WBTHURSAI to register for free. We'll have updates on the CoreWeave acquisition, product announcements, and it's the perfect chance to give feedback directly to the people building the tools you use. Also, my talk on Large Reasoning Models as LLM judges is now up on YouTube. Had to update it live because of the O3 price drop - such is life in AI! Open Source LLMs: Mistral Goes Reasoning Mode Mistral Drops Magistral - Their First Reasoning Model The French champagne of LLMs is back! Mistral released Magistral, their first reasoning model, in two flavors: a 24B parameter open-source Small version and a closed API-only Medium version. And honestly? The naming continues to be chef's kiss - Mistral really has the branding game locked down. Now, here's where it gets spicy. Mistral's benchmarks notably don't include comparisons to Chinese models like Qwen or DeepSeek. Dylan Patel from SemiAnalysis called them out on this, and when he ran the comparisons himself, well... let's just say Magistral Medium barely keeps up with Qwen's tiny 4B parameter model on math benchmarks. Ouch. But here's the thing - and Nisten really drove this home during our discussion - benchmarks don't tell the whole story. He's been using Magistral Small for his workflows and swears by it. "It's almost at the point where I don't want to tell people about it," he said, which is the highest praise from someone who runs models locally all day. The 24B Small version apparently hits that sweet spot for local deployment while being genuinely useful for real work. The model runs on a single RTX 4090 or a 32GB MacBook after quantization, has a 128K context window (though they recommend capping at 40K), and uses a transparent mode that shows its reasoning process. It's Apache 2.0 licensed, multilingual, and available through their Le Chat interface with "Flash Answers" for real-time reasoning. SakanaAI's Text2Lora: The Future is Self-Adapting Models This one blew my mind. SakanaAI (co-founded by one of the Transformer paper authors) released Text2Lora - a method for adapting LLMs to new tasks using ONLY text descriptions. No training data needed! Think about this: instead of fine-tuning a model with thousands of examples to make it better at math, you just... tell it to be better at math. And it works! On Llama 3.1 8B, Text2Lora reaches 77% average accuracy, outperforming all baseline methods. What this means is we're approaching a world where models can essentially customize themselves on-the-fly for whatever task you throw at them. As Nisten put it, "This is revolutionary. The model is actually learning, actually changing its own weights." We're just seeing the first glimpses of this capability, but in 6-12 months? 🎥 Multimedia & Tools: Video, Voice, and Browser Breakthroughs Let’s zip through some multimedia and tool updates that caught my eye this week. Google’s VEO3-fast is a creator’s dream—2x faster 720p video generation, 80% cheaper, and now with audio support. I’ve seen clips on social media (like an NBA ad) that are unreal, though Wolfram noted it’s not fully rolled out in Europe yet. You can access it via APIs like Fail or Replicate, and I’m itching to make a full movie if I had the budget! Midjourney’s gearing up for a video product with their signature style, but they’re also facing heat—Disney and Universal are suing them for copyright infringement over Star Wars and Avengers-like outputs. It’s Hollywood’s first major strike against AI, and while I get the IP concern, it’s odd they picked the smaller player when OpenAI and Google are out there too. This lawsuit could drag on, so stay tuned. OpenAI’s new advanced voice mode dropped, aiming for a natural cadence with better multilingual support (Russian and Hebrew sound gre

    1h 33m

  2. 6 JUN

    📆 ThursdAI - Jun 5, 2025 - Live from AI Engineer with Swyx, new Gemini 2.5 with Logan K and Jack Rae, Self Replicating agents with Morph Labs

    Hey folks, this is Alex, coming to you LIVE from the AI Engineer Worlds Fair! What an incredible episode this week, we recorded live from floor 30th at the Marriott in SF, while Yam was doing live correspondence from the floor of the AI Engineer event, all while Swyx, the cohost of Latent Space podcast, and the creator of AI Engineer (both the conference and the concept itself) joined us for the whole stream - here’s the edited version, please take a look. We've had around 6500 people tune in, and at some point we got 2 surprise guests, straight from the keynote stage, Logan Kilpatrick (PM for AI Studio and lead cheerleader for Gemini) and Jack Rae (principal scientist working on reasoning) joined us for a great chat about Gemini! Mind was absolutely blown! They have just launched the new Gemini 2.5 Pro and I though it would only be fitting to let their new model cover this podcast this week (so below is fully AI generated ... non slop I hope). The show notes and TL;DR is as always in the end. Okay, enough preamble… let's dive into the madness! 🤯 Google Day at AI Engineer: New Gemini 2.5 Pro and a Look Inside the Machine's Mind For the first year of this podcast, a recurring theme was us asking, "Where's Google?" Well, it's safe to say that question has been answered with a firehose of innovation. We were lucky enough to be joined by Google DeepMind's Logan Kilpatrick and Jack Rae, the tech lead for "thinking" within Gemini, literally moments after they left the main stage. Surprise! A New Gemini 2.5 Pro Drops Live Logan kicked things off with a bang, officially announcing a brand new, updated Gemini 2.5 Pro model right there during his keynote. He called it "hopefully the final update to 2.5 Pro," and it comes with a bunch of performance increases, closing the gap on feedback from previous versions and hitting SOTA on benchmarks like Aider. It's clear that the organizational shift to bring the research and product teams together under the DeepMind umbrella is paying massive dividends. Logan pointed out that Google has seen a 50x increase in AI inference over the past year. The flywheel is spinning, and it's spinning fast. How Gemini "Thinks" Then things got even more interesting. Jack Rae gave us an incredible deep dive into what "thinking" actually means for a language model. This was one of the most insightful parts of the conference for me. For years, the bottleneck for LLMs has been test-time compute. Models were trained to respond immediately, applying a fixed amount of computation to go from a prompt to an answer, no matter how hard the question. The only way to get a "smarter" response was to use a bigger model. Jack explained that "Thinking" shatters this limitation. Mechanically, Gemini now has a "thinking stage" where it can generate its own internal text—hypothesizing, testing, correcting, and reasoning—before committing to a final answer. It's an iterative loop of computation that the model can dynamically control, using more compute for harder problems. It learns how to think using reinforcement learning, getting a simple "correct" or "incorrect" signal and backpropagating that to shape its reasoning strategies. We're already seeing the results of this. Jack showed a clear trend: as models get better at reasoning, they're also using more test-time compute. This paradigm also gives developers a "thinking budget" slider in the API for Gemini 2.5 Flash and Pro, allowing a continuous trade-off between cost and performance. The future of this is even wilder. They're working on DeepThink, a high-budget mode for extremely hard problems that uses much deeper, parallel chains of thought. On the tough USA Math Olympiad, where the SOTA was negligible in January, 2.5 Pro reached the 50th percentile of human participants. DeepThink pushes that to the 65th percentile. Jack’s ultimate vision is inspired by the mathematician Ramanujan, who derived incredible theorems from a single textbook by just thinking deeply. The goal is for models to do the same—contemplate a small set of knowledge so deeply that they can push the frontiers of human understanding. Absolutely mind-bending stuff. 🤖 MorphLabs and the Audacious Quest for Verified Superintelligence Just when I thought my mind couldn't be bent any further, we were joined by Jesse Han, the founder and CEO of MorphLabs. Fresh off his keynote, he laid out one of the most ambitious visions I've heard: building the infrastructure for the Singularity and developing "verified superintelligence." The big news was that Christian Szegedy is joining MorphLabs as Chief Scientist. For those who don't know, Christian is a legend—he invented batch norm and adversarial examples, co-founded XAI, and led code reasoning for Grok. That's a serious hire. Jesse’s talk was framed around a fascinating question: "What does it mean to have empathy for the machine?" He argues that as AI develops personhood, we need to think about what it wants. And what it wants, according to Morph, is a new kind of cloud infrastructure. This is MorphCloud, built on a new virtualization stack called Infinibranch. Here’s the key unlock: it allows agents to instantaneously snapshot, branch, and replicate their entire VM state. Imagine an agent reaching a decision point. Instead of choosing one path, it can branch its entire existence—all its processes, memory, and state—to explore every option in parallel. It can create save states, roll back to previous checkpoints, and even merge its work back together. This is a monumental step for agentic AI. It moves beyond agents that are just a series of API calls to agents that are truly embodied in complex software environments. It unlocks the potential for recursive self-improvement and large-scale reinforcement learning in a way that's currently impossible. It’s a bold, sci-fi vision, but they're building the infrastructure to make it a reality today. 🔥 The Agent Conversation: OpenAI, MCP, and Magic Moments The undeniable buzz on the conference floor was all about agents. You couldn't walk ten feet without hearing someone talking about agents, tools, and MCP. OpenAI is leaning in here too. This week, they made their Codex coding agent available to all ChatGPT Plus users and announced that ChatGPT will soon be able to listen in on your Zoom meetings. This is all part of a broader push to make AI more active and integrated into our workflows. The MCP (Model-Context-Protocol) track at the conference was packed, with lines going down the hall. (Alex here, I had a blast talking during that track about MCP observability, you can catch our talk here on the live stream of AI Engineer) Logan Kilpatrick offered a grounded perspective, suggesting the hype might be a bit overblown but acknowledging the critical need for an open standard for tool use, a void left when OpenAI didn't formalize ChatML. I have to share my own jaw-dropping MCP moment from this week. I was coding an agent using an IDE that supports MCP. My agent, which was trying to debug itself, used an MCP tool to check its own observability traces on the Weights & Biases platform. While doing so, it discovered a new tool that our team had just added to the MCP server—a support bot. Without any prompting from me, my coding agent formulated a question, "chatted" with the support agent to get the answer, came back, fixed its own code, and then re-checked its work. Agent-to-agent communication, happening automatically to solve a problem. My jaw was on the floor. That's the magic of open standards. This Week's Buzz from Weights & Biases Speaking of verification and agents, the buzz from our side is all about it! At our booth here at AI Engineer, we have a Robodog running around, connected to our LLM evaluation platform, W&B Weave. As Jesse from MorphLabs discussed, verifying what these complex agentic systems are doing is critical. Whether it's superintelligence or your production application, you need to be able to evaluate, trace, and understand its behavior. We're building the tools to do just that. And if you're in San Francisco, don't forget our own conference, Fully Connected, is happening on June 18th and 19th! It's going to be another amazing gathering of builders and researchers. Fullyconnected.com get in FREE with the promo code WBTHURSAI What a show. The energy, the announcements, the sheer brainpower in one place was something to behold. We’re at a point where the conversation has shifted from theory to practice, from hype to real, tangible engineering. The tracks on agents and enterprise adoption were overflowing because people are building, right now. It was an honor and a privilege to bring this special episode to you all. Thank you for tuning in. We'll be back to our regular programming next week! (and Alex will be back to writing his own newsletter, not send direct AI output!) AI News TL;DR and show notes * Hosts and Guests * Alex Volkov - AI Evangelist & Weights & Biases (@altryne) * Co Hosts - @swyx @yampeleg @romechenko * Guests - @officialLoganK, @jack_w_rae * Open Source LLMs * ByteDance / ContentV-8B - (HF) * Big CO LLMs + APIs * Gemini Pro 2.5 updated Jun 5th (X) * SOTA on HLE, Aider, and GPQA * Now supports thinking budgets * Same cost, on pareto frontier * Closes gap on 03-25 regressions * OAI AVM injects ads and stopped singing (X) * OpenAI Codex is now available to plus members and has internet access (X) * ~24,000 NEW PRs overnight from Codex after @OpenAI expands access to free users. * OpenAI will record meetings and released connectors like (X) * TestingCatalog News 🗞@testingcatalogJun 4, 2025 OpenAI released loads of connectors for Team accounts! Most of these connectors can be used for Deep Research, while Google Drive, SharePoint, Dropbox and Box could be used in all chats. https://t.co/oBEmYGKguE * Anthropic cuts windsurf access for Windsurf (X) * Without warning, Anthropic cuts off Windsurf from official Claude 3 and 4

    1h 44m

  3. 29 MAY

    📆 ThursdAI - May 29 - DeepSeek R1 Resurfaces, VEO3 viral moments, Opus 4 a week after, Flux Kontext image editing & more AI news

    Hey everyone, Alex here 👋 Welcome back to another absolutely wild week in AI! I'm coming to you live from the Fontainebleau Hotel in Vegas at the Imagine AI conference, and wow, what a perfect setting to discuss how AI is literally reimagining our world. After last week's absolute explosion of releases (Claude Opus 4, Google I/O madness, OpenAI Codex and Jony colab), this week gave us a chance to breathe... sort of. Because even in a "quiet" week, we still got a new DeepSeek model that's pushing boundaries, and the entire internet discovered that we might all just be prompts. Yeah, it's been that kind of week! Before we dive in, quick shoutout to everyone who joined us live - we had some technical hiccups with the Twitter Spaces audio (sorry about that!), but the YouTube stream was fire. And speaking of fire, we had two incredible guests join us: Charlie Holtz from Chorus (the multi-model chat app that's changing how we interact with AI) and Linus Eckenstam, who's been traveling the AI conference circuit and bringing us insights from the frontlines of the generative AI revolution. Open Source AI & LLMs: DeepSeek Whales & Mind-Bending Papers DeepSeek dropped R1-0528 out of nowhere, an update to their reasoning beast with some serious jumps in performance. We’re talking AIME at 91 (beating previous scores by a mile), LiveCodeBench at 73, and SWE verified at 57.6. It’s edging closer to heavyweights like o3, and folks on X are already calling it “clearer thinking.” There was hype it might’ve been R2, but the impact didn’t quite crash the stock exchange like past releases. Still, it’s likely among the best open-weight models out there. So what's new? Early reports and some of my own poking around suggest this model "thinks clearer now." Nisten mentioned that while previous DeepSeek models sometimes liked to "vibe around" and explore the latent space before settling on an answer, this one feels a bit more direct. And here’s the kicker—they also released an 8B distilled version based on Qwen3, runnable on your laptop. Yam called it potentially the best 8B model to date, and you can try it on Ollama right now. No need for a monster rig! The Mind-Bending "Learning to Reason Without External Rewards" Paper Okay, this paper result broke my brain, and apparently everyone else's too. This paper shows that models can improve through reinforcement learning with its own intuition of whether or not it's correct. 😮 It's like the placebo effect for AI! The researchers trained models without telling them what was good or bad, but rather, utilized a new framework called Intuitor, where the reward was based on how the "self certainty". The thing that took my whole timeline by storm is, it works! GRPO (Group Policy Optimization) - the framework that DeepSeek gave to the world with R1 is based on external rewards (human optimize) and Intuitor seems to be mathcing or even exceeding some of GRPO results when Qwen2.5 3B was used to finetune. Incredible incredible stuff Big Companies LLMs & APIs Claude Opus 4: A Week Later – The Dev Darling? Claude Opus 4, whose launch we celebrated live on the show, has had a week to make its mark. Charlie Holtz, who's building Chorus (more on that amazing app in a bit!), shared that while it's sometimes "astrology" to judge the vibes of a new model, Opus 4 feels like a step change, especially in coding. He mentioned that Claude Code, powered by Opus 4 (and Sonnet 4 for implementation), is now tackling GitHub issues that were too complex just weeks ago. He even had a coworker who "vibe coded three websites in a weekend" with it – that's a tangible productivity boost! Linus Eckenstam highlighted how Lovable.dev saw their syntax error rates plummet by nearly 50% after integrating Claude 4. That’s quantifiable proof of improvement! It's clear Anthropic is leaning heavily into the developer/coding space. Claude Opus is now #1 on the LMArena WebDev arena, further cementing its reputation. I had my own magical moment with Opus 4 this week. I was working on an MCP observability talk for the AI Engineer conference and trying to integrate Weave (our observability and evals framework at Weights & Biases) into a project. Using Windsurf's Cascade agent (which now lets you bring your own Opus 4 key, by the way – good move, Windsurf!), Opus 4 not only tried to implement Weave into my agent but, when it got stuck, it figured out it had access to the Weights & Biases support bot via our MCP tool. It then formulated a question to the support bot (which is also AI-powered!), got an answer, and used that to fix the implementation. It then went back and checked if the Weave trace appeared in the dashboard! Agents talking to agents to solve a problem, all while I just watched – my jaw was on the floor. Absolutely mind-blowing. Quick Hits: Voice Updates from OpenAI & Anthropic OpenAI’s Advanced Voice Mode finally sings—yes, I’ve been waiting for this! It can belt out tunes like Mariah Carey, which is just fun. Anthropic also rolled out voice mode on mobile, keeping up in the conversational race. Both are cool steps, but I’m more hyped for what’s next in voice AI—stay tuned below (OpenAI X, Anthropic X). 🐝 This Week's Buzz: Weights & Biases Updates! Alright, time for a quick update from the world of Weights & Biases! * Fully Connected is Coming! Our flagship 2-day conference, Fully Connected, is happening on June 18th and 19th in San Francisco. It's going to be packed with amazing speakers and insights into the world of AI development. You can still grab tickets, and as a ThursdAI listener, use the promo code WBTHURSAI for a 100% off ticket! I hustled to get yall this discount! (Register here) * AI Engineer World's Fair Next Week! I'm super excited for the AI Engineer conference in San Francisco next week. Yam Peleg and I will be there, and we're planning another live ThursdAI show from the event! If you want to join the livestream or snag a last-minute ticket, use the coupon code THANKSTHURSDAI for 30% off (Get it HERE) Vision & Video: Reality is Optional Now VEO3 and the Prompt Theory Phenomenon Google's VEO3 has completely taken over TikTok with the "Prompt Theory" videos. If you haven't seen these yet, stop reading and watch ☝️. The concept is brilliant - AI-generated characters discussing whether they're "made of prompts," creating this meta-commentary on consciousness and reality. The technical achievement here is staggering. We're not just talking about good visuals - VEO3 nails temporal consistency, character emotions, situational awareness (characters look at whoever's speaking), perfect lip sync, and contextually appropriate sound effects. Linus made a profound point - if not for the audio, VEO3 might not have been as explosive. The combination of visuals AND audio together is what's making people question reality. We're seeing people post actual human videos claiming they're AI-generated because the uncanny valley has been crossed so thoroughly. Odyssey's Interactive Worlds: The Holodeck Prototype Odyssey dropped their interactive video demo, and folks... we're literally walking through AI-generated worlds in real-time. This isn't a game engine rendering 3D models - this is a world model generating each frame as you move through it with WASD controls. Yes, it's blurry. Yes, I got stuck in a doorway. But remember Will Smith eating spaghetti from two years ago? The pace of progress is absolutely insane. As Linus pointed out, we're at the "GAN era" of world models. Combine VEO3's quality with Odyssey's interactivity, and we're looking at completely personalized, infinite entertainment experiences. The implications that Yam laid out still have me shook - imagine Netflix shows completely customized to you, with your context and preferences, generated on the fly. Not just choosing from a catalog, but creating entirely new content just for you. We're not ready for this, but it's coming fast. Hunyuan's Open Source Avatar Revolution While the big companies are keeping their video models closed, Tencent dropped two incredible open source releases: HunyuanPortrait and HunyuanAvatar. These are legitimate competitors to Hedra and HeyGen, but completely open source. HunyuanPortrait does high-fidelity portrait animation from a single image plus video. HunyuanAvatar goes further with 1 image + audio, and lipsync, body animation, multi-character support, and emotion control. Wolfram tested these extensively and confirmed they're "state of the art for open source." The portrait model is basically perfect for deepfakes (use responsibly, people!), while the avatar model opens up possibilities for AI assistants with consistent visual presence. 🖼️ AI Art & Diffusion Black Forest Labs drops Flux Kontext - SOTA image editing! This came as massive breaking news during the show (thought we didn't catch it live!) - Black Forest Labs, creators of Flux, dropped an incredible Image Editing model called Kontext (really, 3 models, Pro, Max and 12B open source Dev in private preview). The are consistent, context aware text and image editing! Just see the below example If you used GPT-image to Ghiblify yourself, or VEO, you know that those are not image editing models, your face will look different every generation. These images model keep you consistent, while adding what you wanted. This character consistency is something many folks really want and it's great to see Flux innovating and bringing us SOTA again and are absolutely crushing GPT-image in instruction following, character preservation and style reference! Maybe the most important thing about this model is the increible speed. While the Ghiblification chatGPT trend took the world by storm, GPT images are SLOW! Check out the speed comparisons on Kontext! You can play around with these models on the new Flux Playground, but they also already integrated into FAL, FreePik, Replicate, Krea and tons of other services!

    1h 28m

  4. 23 MAY

    📆 ThursdAI - Veo3, Google IO25, Claude 4 Opus/Sonnet, OpenAI x Jony Ive, Codex, Copilot Agent - INSANE AI week

    Hey folks, Alex here, welcome back to ThursdAI! And folks, after the last week was the calm before the storm, "The storm came, y'all" – that's an understatement. This wasn't just a storm; it was an AI hurricane, a category 5 of announcements that left us all reeling (in the best way possible!). From being on the ground at Google I/O to live-watching Anthropic drop Claude 4 during our show, it's been an absolute whirlwind. This week was so packed, it felt like AI Christmas, with tech giants and open-source heroes alike showering us with gifts. We saw OpenAI play their classic pre-and-post-Google I/O chess game, Microsoft make some serious open-source moves, Google unleash an avalanche of updates, and Anthropic crash the party with Claude 4 Opus and Sonnet live stream in the middle of ThursdAI! So buckle up, because we're about to try and unpack this glorious chaos. As always, we're here to help you collectively know, learn, and stay up to date, so you don't have to. Let's dive in! (TL;DR and links in the end) Open Source LLMs Kicking Things Off Even with the titans battling, the open-source community dropped some serious heat this week. It wasn't the main headline grabber, but the releases were significant! Gemma 3n: Tiny But Mighty Matryoshka First up, Google's Gemma 3n. This isn't just another small model; it's a "Nano-plus" preview, a 4-billion parameter MatFormer (Matryoshka Transformer – how cool is that name?) model designed for mobile-first multimodal applications. The really slick part? It has a nested 2-billion parameter sub-model that can run entirely on phones or Chromebooks. Yam was particularly excited about this one, pointing out the innovative "model inside another model" design. The idea is you can use half the model, not depth-wise, but throughout the layers, for a smaller footprint without sacrificing too much. It accepts interleaved text, image, audio, and video, supports ASR and speech translation, and even ships with RAG and function-calling libraries for edge apps. With a 128K token window and responsible AI features baked in, Gemma 3n is looking like a powerful tool for on-device AI. Google claims it beats prior 4B mobile models on MMLU-Lite and MMMU-Mini. It's an early preview in Google AI Studio, but it definitely flies on mobile devices. Mistral & AllHands Unleash Devstral 24B Then we got a collaboration from Mistral and AllHands: Devstral, a 24-billion parameter, state-of-the-art open model focused on code. We've been waiting for Mistral to drop some open-source goodness, and this one didn't disappoint.Nisten was super hyped, noting it beats o3-Mini on SWE-bench verified – a tough benchmark! He called it "the first proper vibe coder that you can run on a 3090," which is a big deal for coders who want local power and privacy. This is a fantastic development for the open-source coding community. The Pre-I/O Tremors: OpenAI & Microsoft Set the Stage As we predicted, OpenAI couldn't resist dropping some news right before Google I/O. OpenAI's Codex Returns as an Agent OpenAI launched Codex – yes, that Codex, but reborn as an asynchronous coding agent. This isn't just a CLI tool anymore; it connects to GitHub, does pull requests, fixes bugs, and navigates your codebase. It's powered by a new coding model fine-tuned for large codebases and was SOTA on SWE Agent when it dropped. Funnily, the model is also called Codex, this time, Codex-1. And this gives us a perfect opportunity to talk about the emerging categories I'm seeing among Code Generator agents and tools: * IDE-based (Cursor, Windsurf): Live pair programming in your editor * Vibe coding (Lovable, Bolt, v0): "Build me a UI" style tools for non-coders * CLI tools (Claude Code, Codex-cli): Terminal-based assistants * Async agents (Claude Code, Jules, Codex, GitHub Copilot agent, Devin): Work on your repos while you sleep, open pull requests for you to review, async Codex (this new one) falls into category number 4, and with today's release, Cursor seems to also strive to get to category number 4 with background processing. Microsoft BUILD: Open Source Copilot and Copilot Agent Mode Then came Microsoft Build, their huge developer conference, with a flurry of announcements.The biggest one for me? GitHub Copilot's front-end code is now open source! The VS Code editor part was already open, but the Copilot integration itself wasn't. This is a massive move, likely a direct answer to the insane valuations of VS Code clones like Cursor. Now, you can theoretically clone GitHub Copilot with VS Code and swing for the fences. GitHub Copilot also launched as an asynchronous coding assistant, very similar in function to OpenAI's Codex, allowing it to be assigned tasks and create/update PRs. This puts Copilot right into category 4 of code assistants, and with the native Github Integration, they may actually have a leg up in this race! And if that wasn't enough, Microsoft is adding MCP (Model Context Protocol) support directly into the Windows OS. The implications of having the world's biggest operating system natively support this agentic protocol are huge. Google I/O: An "Ultra" Event Indeed! Then came Tuesday, and Google I/O. I was there in the thick of it, and folks, it was an absolute barrage. Google is shipping. The theme could have been "Ultra" for many reasons, as we'll see. First off, the scale: Google reported a 49x increase in AI usage since last year's I/O, jumping from 9 trillion tokens processed to a mind-boggling 480 trillion tokens. That's a testament to their generous free tiers and the explosion of AI adoption. Gemini 2.5 Pro & Flash: #1 and #2 LLMs on Arena Gemini 2.5 Flash got an update and is now #2 on the LMArena leaderboard (with Gemini 2.5 Pro still holding #1). Both Pro and Flash gained some serious new capabilities: * Deep Think mode: This enhanced reasoning mode is pushing Gemini's scores to new heights, hitting 84% on MMMU and topping LiveCodeBench. It's about giving the model more "time" to work through complex problems. * Native Audio I/O: We're talking real-time TTS in 24 languages with two voices, and affective dialogue capabilities. This is the advanced voice mode we've been waiting for, now built-in. * Project Mariner: Computer-use actions are being exposed via the Gemini API & Vertex AI for RPA partners. This started as a Chrome extension to control your browser and now seems to be a cloud-based API, allowing Gemini to use the web, not just browse it. This feels like Google teaching its AI to interact with the JavaScript-heavy web, much like they taught their crawlers years ago. * Thought Summaries: Okay, here's one update I'm not a fan of. They've switched from raw thinking traces to "thought summaries" in the API. We want the actual traces! That's how we learn and debug. * Thinking Budgets: Previously a Flash-only feature, token ceilings for controlling latency/cost now extend to Pro. * Flash Upgrade: 20-30% fewer tokens, better reasoning/multimodal scores, and GA in early June. Gemini Diffusion: Speed Demon for Code and Math This one got Yam Peleg incredibly excited. Gemini Diffusion is a new approach, different from transformers, for super-speed editing of code and math tasks. We saw demos hitting 2000 tokens per second! While there might be limitations at longer contexts, its speed and infilling capabilities are seriously impressive for a research preview. This is the first diffusion model for text we've seen from the frontier labs, and it looks sick. Funny note, they had to slow down the demo video to actually show the diffusion process, because at 2000t/s - apps appear as though out of thin air! The "Ultra" Tier and Jules, Google's Coding Agent Remember the "Ultra event" jokes? Well, Google announced a Gemini Ultra tier for $250/month. This tops OpenAI's Pro plan and includes DeepThink access, a generous amount of VEO3 generation, YouTube Premium, and a whopping 30TB of storage. It feels geared towards creators and developers. And speaking of developers, Google launched Jules (jules.google)! This is their asynchronous coding assistant (Category 4!). Like Codex and GitHub Copilot Agent, it connects to your GitHub, opens PRs, fixes bugs, and more. The big differentiator? It's currently free, which might make it the default for many. Another powerful agent joins the fray! AI Mode in Search: GA and Enhanced AI Mode in Google Search, which we've discussed on the show before with Robby Stein, is now in General Availability in the US. This is Google's answer to Perplexity and chat-based search.But they didn't stop there: * Personalization: AI Mode can now connect to your Gmail and Docs (if you opt-in) for more personalized results. * Deep Search: While AI Mode is fast, Deep Search offers more comprehensive research capabilities, digging through hundreds of sources, similar to other "deep research" tools. This will eventually be integrated, allowing you to escalate an AI Mode query for a deeper dive. * Project Mariner Integration: AI Mode will be able to click into websites, check availability for tickets, etc., bridging the gap to an "agentic web." I've had a chat with Robby during I/O and you can listen to that interview at the end of the podcast. Veo3: The Undisputed Star of Google I/O For me, and many others I spoke to, Veo3 was the highlight. This is Google's flagship video generation model, and it's on another level. (the video above, including sounds is completely one shot generated from VEO3, no processing or editing) * Realism and Physics: The visual quality and understanding of physics are astounding. * Natively Multimodal: This is huge. Veo3 generates native audio, including coherent speech, conversations, and sound effects, all synced perfectly. It can even generate text within videos. * Coherent Characters: Characters remain consistent across scenes and have situational awareness, who speaks when, where characters look. * Image Upload & Reference Ability: While image upload was closed for

    1h 28m

  5. 16 MAY

    📆 ThursdAI - May 15 - Genocidal Grok, ChatGPT 4.1, AM-Thinking, Distributed LLM training & more AI news

    Hey yall, this is Alex 👋 What a wild week, it started super slow, and it still did feel slow as releases are concerned, but the most interesting story was yet another AI gone "rogue" (have you even heard about "kill the boar", if not, Grok will tell you all about it) Otherwise it seemed fairly quiet in AI land this week, besides another Chinese newcomer called AM-thinking 32B that beats DeepSeek and Qwen, and Stability making a small comeback, we focused on distributed LLM training and ChatGPT 4.1 We've had a ton of fun on this episode, this one was being recorded from the Weights & Biases SF Office (I'm here to cover Google IO next week!) Let’s dig in—because what looks like a slow week on the surface was anything but dull under the hood (TL'DR and show notes at the end as always) Big Companies & APIs Why does XAI Grok talk about White Genocide and "Kill the boar"?? Just after we're getting over the chatGPT glazing incident , folks started noticing that @grok - XAI's frontier LLM that is also responding to X replies, started talking about White Genocide in South Africa and something called "Kill the boer" with no reference to any of these things in the question! Since we recorded the episode, XAI official X account posted that an "unauthorized modification" happened to the system prompt, and that going forward they would open source all the prompts (and they did). Whether or not they would keep updating that repository though, remains unclear (see the "open sourced" x algorithm to which the last push was over a year ago, or the promised Grok 2 that was never open sourced) While it's great to have some more clarity from the Xai team, this behavior raises a bunch of questions about the increasing roles of AI's in our lives and the trust that many folks are giving them. Adding fuel to the fire, are Uncle Elon's recent tweets that are related to South Africa, and this specific change seems to be related to those views at least partly. Remember also, Grok was meant as "maximally truth seeking" AI! I really hope this transparency continues! Open Source LLMs: The Decentralization Tsunami AM-Thinking v1: Dense Reasoning, SOTA Math, Single-Checkpoint Deployability Open source starts with the kind of progress that would have been unthinkable 18 months ago: a 32B dense LLM, openly released, that takes on the big mixture-of-experts models and comes out on top for math and code. AM-Thinking v1 (paper here) hits 85.3% on AIME 2024, 70.3% on LiveCodeBench v5, and 92.5% on Arena-Hard. It even runs at 25 tokens/sec on a single 80GB GPU with INT4 quantization. The model supports a /think reasoning toggle (chain-of-thought on demand), comes with a permissive license, and is fully tooled for vLLM, LM Studio, and Ollama. Want to see where dense models can still push the limits? This is it. And yes, they’re already working on a multilingual RLHF pass and 128k context window. Personal note: We haven’t seen this kind of “out of nowhere” leaderboard jump since the early days of Qwen or DeepSeek. This company's debut on HuggingFace with a model that crushes! Decentralized LLM Training: Nous Research Psyche & Prime Intellect INTELLECT-2 This week, open source LLMs didn’t just mean “here are some weights.” It meant distributed, decentralized, and—dare I say—permissionless AI. Two labs stood out: Nous Research launches Psyche Dylan Rolnick from Nous Research joined the show to explain Psyche: a Rust-powered, distributed LLM training network where you can watch a 40B model (Consilience-40B) evolve in real time, join the training with your own hardware, and even have your work attested on a Solana smart contract. The core innovation? DisTrO (Decoupled Momentum) which we covered back in December that drastically compresses the gradient exchange so that training large models over the public internet isn’t a pipe dream—it’s happening right now. Live dashboard here, open codebase, and the testnet already humming with early results. This massive 40B attempt is going to show whether distributed training actually works! The cool thing about their live dashboard is, it's WandB behind the scenes, but with a very thematic and cool Nous Research reskin! This model saves constant checkpoints to the hub as well, so the open source community can enjoy a full process of seeing a model being trained! Prime Intellect INTELLECT-2 Not to be outdone, Prime Intellect’s INTELLECT-2 released a globally decentralized, 32B RL-trained reasoning model, built on a permissionless swarm of GPUs. Using their own PRIME-RL framework, SHARDCAST checkpointing, and an LSH-based rollout verifier, they’re not just releasing a model—they’re proving it’s possible to scale serious RL outside a data center. OpenAI's HealthBench: Can LLMs Judge Medical Safety? One of the most intriguing drops of the week is HealthBench, a physician-crafted benchmark for evaluating LLMs in clinical settings. Instead of just multiple-choice “gotcha” tests, HealthBench brings in 262 doctors from 60 countries, 26 specialties, and nearly 50 languages to write rubrics for 5,000 realistic health conversations. The real innovation: LLM as judge. Models like GPT-4.1 are graded against physician-written rubrics, and the agreement between model and human judges matches the agreement between two doctors. Even the “mini” variants of GPT-4.1 are showing serious promise—faster, cheaper, and (on the “Hard” subset) giving the full-size models a run for their money. Other Open Source Standouts Falcon-Edge: Ternary BitNet for Edge Devices The Falcon-Edge project brings us 1B and 3B-parameter language models trained directly in ternary BitNet format (weights constrained to -1, 0, 1), which slashes memory and compute requirements and enables inference on StepFun Step1x-3D: Controllable Open 3D Generation StepFun’s 3D pipeline is a two-stage system that creates watertight geometry and then view-consistent textures, trained on 2M curated meshes. It’s controllable by text, images, and style prompts—and it’s fully open source, including a huge asset dataset. Big Company LLMs & APIs: Models, Modes, and Model Zoo Confusion GPT-4.1 Comes to ChatGPT: Model Zoo Mayhem OpenAI’s GPT-4.1 series—previously API-only—is now available in the ChatGPT interface. Why does this matter? Because the UX of modern LLMs is, frankly, a mess: seven model options in the dropdown, each with its quirks, speed, and context length. Most casual users don’t even know the dropdown exists. “Alex, ChatGPT is broken!” Actually, you just need to pick a different model. The good news: 4.1 is fast, great at coding, and in many tasks, preferable to the “reasoning” behemoths. My advice (and you can share this with your relatives): when in doubt, just switch the model. Bonus: The long-promised million-token context window is here (sort of)—except in the UI, where it’s more like 128k and sometimes silently truncated. My weekly rant: transparency, OpenAI. ProTip: If you’re hitting invisible context limits, try pasting your long transcripts on the web, not in the Mac app. Don’t trust the UI! AlphaEvolve: DeepMind’s Gemini-Powered Algorithmic Discovery AlphaEvolve is the kind of project that used to sound like AGI hype—and now it’s just a Tuesday at DeepMind. By pairing Gemini Flash and Gemini Pro in an evolutionary search loop to improve algorithms! This is like, real innovation and it's done with existing models which is super super cool! AlphaEvolve uses a combination of Gemini Flash (for breadth of ideas) and Gemini Pro (for depth and refinement) in an evolutionary loop. It generates, tests, and mutates code to invent faster algorithms. And it's already yielding incredible results: * It discovered a new scheduling heuristic for Google's Borg system, resulting in a 0.7% global compute recovery. That's massive at Google's scale. * It improved a matrix-multiply kernel by 23%, which in turn led to a 1% shorter Gemini training time. As Nisten said, the model basically paid for itself! Perhaps most impressively, it found a 48-multiplication algorithm for 4x4 complex matrices, beating the famous Strassen algorithm from 1969 (which used 49 multiplications). This is AI making genuine, novel scientific discoveries. AGI in the garden, anyone? If you still think LLMs are “just glorified autocomplete,” it’s time to update your mental model. This is model-driven algorithmic discovery, and it’s already changing the pace of hardware, math, and software design. The only downside: it’s not public yet, but there’s an interest form if you want to be a tester. This Week's Buzz - Everything W&B! It's a busy time here at Weights & Biases, and I'm super excited about a couple of upcoming events where you can connect with us and the broader AI community. Fully Connected: Our very own 2-day conference is happening June 18-19 in San Francisco! We've got an amazing lineup of speakers, including Varun Mohan from WindSurf (formerly Codeium), Heikki Kubler from CoreWeave, our CEO Lucas Bewald, CTO Shawn Lewis, Joe Spizak from Meta, and a keynote from Javi Soltero, VP Product AI at Google. It's going to be packed with insights on building and scaling AI. And because you're a ThursdAI listener, you can get in for FREE with the promo code WBTHURSAI at fullyconnected.com. Don't miss out! AI.Engineer World's Fair: This has become THE conference for AI engineers, and W&B is a proud sponsor for the third year running! It's happening in San Francisco from June 3rd to 5th. I'll be speaking there on MCP Observability with Ben from LangChain on June 4th.Even more exciting, ThursdAI will be broadcasting LIVE from the media booth at AI.Engineer on June 5th! Come say hi! Tickets are flying, but we've got a special discount for you: use promo code THANKSTHURSDAI for 30% off your ticket here. Yam Peleg even decided on the show he's coming after hearing about it! It's going to be an incredi

    1h 29m

  6. 9 MAY

    ThursdAI - May 8th - new Gemini pro, Mistral Medium, OpenAI restructuring, HeyGen Realistic Avatars & more AI news

    Hey folks, Alex here (yes, real me, not my AI avatar, yet) Compared to previous weeks, this week was pretty "chill" in the world of AI, though we did get a pretty significant Gemini 2.5 Pro update, it basically beat itself on the Arena. With Mistral releasing a new medium model (not OSS) and Nvidia finally dropping Nemotron Ultra (both ignoring Qwen 3 performance) there was also a few open source updates. To me the highlight of this week was a breakthrough in AI Avatars, with Heygen's new IV model, Beating ByteDance's OmniHuman (our coverage) and Hedra labs, they've set an absolute SOTA benchmark for 1 photo to animated realistic avatar. Hell, Iet me record all this real quick and show you how good it is! How good is that?? I'm still kind of blown away. I have managed to get a free month promo code for you guys, look for it in the TL;DR section at the end of the newsletter. Of course, if you’re rather watch than listen or read, here’s our live recording on YT OpenSource AI NVIDIA's Nemotron Ultra V1: Refining the Best with a Reasoning Toggle 🧠 NVIDIA also threw their hat further into the ring with the release of Nemotron Ultra V1, alongside updated Super and Nano versions. We've talked about Nemotron before – these are NVIDIA's pruned and distilled versions of Llama 3.1, and they've been impressive. The Ultra version is the flagship, a 253 billion parameter dense model (distilled and pruned from Llama 3.1 405B), and it's packed with interesting features. One of the coolest things is the dynamic reasoning toggle. You can literally tell the model "detailed thinking on" or "detailed thinking off" via a system prompt during inference. This is something Qwen also supports, and it looks like the industry is converging on this idea of letting users control the "depth" of thought, which is super neat. Nemotron Ultra boasts a 128K context window and, impressively, can fit on a single 8xH100 node thanks to Neural Architecture Search (NAS) and FFN-Fusion. And performance-wise, it actually outperforms the Llama 3 405B model it was distilled from, which is a big deal. NVIDIA shared a chart from Artificial Analysis (dated April 2025, notably before Qwen3's latest surge) showing Nemotron Ultra standing strong among models like Gemini 2.5 Flash and Opus 3 Mini. What's also great is NVIDIA's commitment to openness here: they've released the models under a commercially permissive NVIDIA Open Model License, the complete post-training dataset (Llama-Nemotron-Post-Training-Dataset), and their training codebases (NeMo, NeMo-Aligner, Megatron-LM). This allows for reproducibility and further community development. Yam Peleg pointed out the cool stuff they did with Neural Architecture Search to optimally reduce parameters without losing performance. Absolute Zero: AI Learning to Learn, Zero (curated) Data Required! (Arxiv) LDJ brought up a fascinating paper that ties into this theme of self-improvement and reinforcement learning: "Absolute Zero: Reinforced Self-play Reasoning with Zero Data" from Andrew Zhao (Tsinghua University) and a few others The core idea here is a system that self-evolves its training curriculum and reasoning ability. Instead of needing a pre-curated dataset of problems, the model creates the problems itself (e.g., code reasoning tasks) and then uses something like a Code Executor to validate its proposed solutions, serving as a unified source of verifiable reward. It's open-ended yet grounded learning. By having a verifiable environment (code either works or it doesn't), the model can essentially teach itself to code without external human-curated data. The paper shows fine-tunes of Qwen models (like Qwen Coder) achieving state-of-the-art results on benchmarks like MBBP and AIME (Math Olympiad) with no pre-existing data for those problems. The model hallucinates questions, creates its own rewards, learns, and improves. This is a step beyond synthetic data, where humans are still largely in charge of generation. It's wild, and it points towards a future where AI systems could become increasingly autonomous in their learning. Big Companies & APIs Google dropped another update to their Gemini 2.5 Pro, this time the "IO edition" preview, specifically touting enhanced coding performance. This new version jumped to the #1 spot on WebDev Arena (a benchmark where human evaluators choose between two side-by-side code generations in VS Code), with a +147 Elo point gain, surpassing Claude 3.7 Sonnet. It also showed improvements on benchmarks like LiveCodeBench (up 7.39%) and Aider Polyglot (up ~3-6%). Google also highlighted its state-of-the-art video understanding (84.8% on VideoMME) with examples like generating code from a video of an app. Which essentially lets you record a drawing of how your app interaction will happen, and the model will use that video instructions! It's pretty cool. Though, not everyone was as impressed, folks noted that while gaining in a few evals, this model also regressed in several others including Vibe-Eval (Reka's multimodal benchmark), Humanity's Last Exam, AIME, MMMU, and even long context understanding (MRCR). It's a good reminder that model updates often involve trade-offs – you can't always win at everything. BREAKING: Gemini's Implicit Caching - A Game Changer for Costs! 💰 Just as we were wrapping up this segment on the show, news broke that Google launched implicit caching in Gemini APIs! This is a huge deal for developers. Previously, Gemini offered explicit caching, where you had to manually tell the API what context to cache – a bit of a pain. Now, with implicit caching, the system automatically enables up to 75% cost savings when your request hits a cache. This is fantastic, especially for long-context applications, which is where Gemini's 1-2 million token context window really shines. If you're repeatedly sending large documents or codebases, this will significantly reduce your API bills. OpenAI has had automatic caching for a while, and it's great to see Google matching this for a much better developer experience and cost-effectiveness. It also saves Google a ton on inference, so it's a win-win! Mistral Medium 3: The Closed Turn 😥 Mistral, once the darling of the open-source community for models like Mistral 7B and Mixtral, announced Mistral Medium 3. The catch? It's not open source. They're positioning it as a multimodal frontier model with 128K context, claiming it matches or surpasses GPT-4-class benchmarks while being cheaper (priced at $0.40/M input and $2/M output tokens). However they haven't added Gemini Flash 2.5 here, which is 70% cheaper while being faster as well, nor did they mention Qwen. Nisten voiced a sentiment many in the community share: he used to use LeChat frequently because he knew and understood the underlying open-source models. Now, with a closed model, it's a black box. It's a bit like pirating music users often being the biggest buyers – understanding the open model often leads to more commercial usage. Wolfram offered a European perspective, noting that Mistral, as a European company, might have a unique advantage with businesses concerned about GDPR and data sovereignty, who might be hesitant to use US or Chinese cloud APIs. For them, a strong European alternative, even if closed, could be appealing. OpenAI's New Chapter: Restructuring for the Future OpenAI announced an evolution in its corporate structure. The key points are: * The OpenAI non-profit will continue to control the entire organization. * The existing for-profit LLC will become a Public Benefit Corporation (PBC). * The non-profit will be a significant owner of the PBC and will control it. * Both the non-profit and PBC will continue to share the same mission: ensuring AGI benefits all of humanity. This move seems to address some of the governance concerns that have swirled around OpenAI, particularly in light of Elon Musk's lawsuit regarding its shift from a non-profit to a capped-profit entity. LDJ explained that the main worry for many was whether the non-profit would lose control or its stake in the main research/product arm. This restructuring appears to ensure the non-profit remains at the helm and that the PBC is legally bound to the non-profit's mission, not just investor interests. It's an important step for a company with such a profound potential impact on society. And in related OpenAI news, the acquisition of Windsurf (the VS Code fork) for a reported $3 billion went through, while Cursor (another VS Code fork) announced a $9 billion valuation. It's wild to see these developer tools, which are essentially forks with an AI layer, reaching such massive valuations. Microsoft's hand is in all of this too – investing in OpenAI, invested in Cursor, owning VS Code, and now OpenAI buying Windsurf. It's a tangled web! Finally, a quick mention that Sam Altman (OpenAI), Lisa Su (AMD), Mike Intrator (CoreWeave - my new CEO!), and folks from Microsoft were testifying before the U.S. Senate today about how to ensure America leads in AI and what innovation means. These conversations are crucial as AI continues to reshape our world. This Weeks Buzz - Come Vibe with Us at Fully Connected! (SF, June 18-19) 🎉 Our two-day conference, Fully Connected, is happening in San Francisco on June 18th and 19th, and it's going to be awesome! We've got an incredible lineup of speakers, including Joe Spizak from the Llama team at Meta and Varun from Windsurf. It's two full days of programming, learning, and connecting with folks at the forefront of AI. And because you're part of the ThursdAI family, I've got a special promo code for you: use WBTHURSAI to get a free ticket on me! If you're in or around SF, I'd love to see you there. Come hang out, learn, and vibe with us! Register at fullyconnected.com Hackathon Update: Moved to July! 🗓️ The AGI Evals & Agentic Tooling (A2A) + MCP Hackathon that I was super excited to co-host has been postponed to July 12t

    1h 34m

  7. 1 MAY

    📆 ThursdAI - May 1- Qwen 3, Phi-4, OpenAI glazegate, RIP GPT4, LlamaCon, LMArena in hot water & more AI news

    Hey everyone, Alex here 👋 Welcome back to ThursdAI! And wow, what a week. Seriously, strap in, because the AI landscape just went through some seismic shifts. We're talking about a monumental open-source release from Alibaba with Qwen 3 that has everyone buzzing (including us!), Microsoft dropping Phi-4 with Reasoning, a rather poignant farewell to a legend (RIP GPT-4 – we'll get to the wake shortly), major drama around ChatGPT's "glazing" incident and the subsequent rollback, updates from LlamaCon, a critical look at Chatbot Arena, and a fantastic deep dive into the world of AI evaluations with two absolute experts, Hamel Husain and Shreya Shankar. This week felt like a whirlwind, with open source absolutely dominating the headlines. Qwen 3 didn't just release a model; they dropped an entire ecosystem, setting a potential new benchmark for open-weight releases. And while we pour one out for GPT-4, we also have to grapple with the real-world impact of models like ChatGPT, highlighted by the "glazing" fiasco. Plus, video consistency takes a leap forward with Runway, and we got breaking news live on the show from Claude! So grab your coffee (or beverage of choice), settle in, and let's unpack this incredibly eventful week in AI. Open-Source LLMs Qwen 3 — “Hybrid Thinking” on Tap Alibaba open-weighted the entire Qwen 3 family this week, releasing two MoE titans (up to 235 B total / 22 B active) and six dense siblings all the way down to 0 .6 B, all under Apache 2.0. Day-one support landed in LM Studio, Ollama, vLLM, MLX and llama.cpp. The headline trick is a runtime thinking toggle—drop “/think” to expand chain-of-thought or “/no_think” to sprint. On my Mac, the 30 B-A3B model hit 57 tokens/s when paired with speculative decoding (drafted by the 0 .6 B sibling). Other goodies: * 36 T pre-training tokens (2 × Qwen 2.5) * 128 K context on ≥ 8 B variants (32 K on the tinies) * 119-language coverage, widest in open source * Built-in MCP schema so you can pair with Qwen-Agent * The dense 4 B model actually beats Qwen 2.5-72B-Instruct on several evals—at Raspberry-Pi footprint In short: more parameters when you need them, fewer when you don’t, and the lawyers stay asleep. Read the full drop on the Qwen blog or pull weights from the HF collection. Performance & Efficiency: "Sonnet at Home"? The benchmarks are where things get really exciting. * The 235B MoE rivals or surpasses models like DeepSeek-R1 (which rocked the boat just months ago!), O1, O3-mini, and even Gemini 2.5 Pro on coding and math. * The 4B dense model incredibly beats the previous generation's 72B Instruct model (Qwen 2.5) on multiple benchmarks! 🤯 * The 30B MoE (with only 3B active parameters) is perhaps the star. Nisten pointed out people are getting 100+ tokens/sec on MacBooks. Wolfram achieved an 80% MMLU Pro score locally with a quantized version. The efficiency math is crazy – hitting Qwen 2.5 performance with only ~10% of the active parameters. Nisten dubbed the larger model "Sonnet 3.5 at home," and while acknowledging Sonnet still has an edge in complex "vibe coding," the performance, especially in reasoning and tool use, is remarkably close for an open model you can run yourself. I ran the 30B MoE (3B active) locally using LLM Studio (shoutout for day-one support!) through my Weave evaluation dashboard (Link). On a set of 20 hard reasoning questions, it scored 43%, beating GPT 4.1 mini and nano, and getting close to 4.1 – impressive for a 3B active parameter model running locally! Phi-4-Reasoning — 14B That Punches at 70B+ Microsoft’s Phi team layered 1.4 M chain-of-thought traces plus a dash of RL onto Phi-4 to finally ship a resoning Phi and shipped two MIT-licensed checkpoints: * Phi-4-Reasoning (SFT) * Phi-4-Reasoning-Plus (SFT + RL) Phi-4-R-Plus clocks 78 % on AIME 25, edging DeepSeek-R1-Distill-70B, with 32 K context (stable to 64 K via RoPE). Scratch-pads hide in tags. Full details live in Microsoft’s tech report and HF weights. It's fascinating to see how targeted training on reasoning traces and a small amount of RL can elevate a relatively smaller model to compete with giants on specific tasks. Other Open Source Updates * MiMo-7B: Xiaomi entered the ring with a 7B parameter, MIT-licensed model family, trained on 25T tokens and featuring rule-verifiable RL. (HF model hub) * Helium-1 2B: KyutAI (known for their Mochi voice model) released Helium-1, a 2B parameter model distilled from Gemma-2-9B, focused on European languages, and licensed under CC-BY 4.0. They also open-sourced 'dactory', their data processing pipeline. (Blog, Model (2 B), Dactory pipeline) * Qwen 2.5 Omni 3B: Alongside Qwen 3, the Qwen team also updated their existing Omni model with a 3B model, that retains 90% of the comprehension of its big brother with a 50% VRAM drop! (HF) * JetBrains open sources Mellum: Trained on over 4 trillion tokens with a context window of 8192 tokens across multiple programming languages, they haven't released any comparable eval benchmarks though (HF) Big Companies & APIs: Drama, Departures, and Deployments While open source stole the show, the big players weren't entirely quiet... though maybe some wish they had been. Farewell, GPT-4: Rest In Prompted 🙏 Okay folks, let's take a moment. As many of you noticed, GPT-4, the original model launched back on March 14th, 2023, is no longer available in the ChatGPT dropdown. You can't select it, you can't chat with it anymore. For us here at ThursdAI, this feels significant. GPT-4's launch was the catalyst for this show. We literally started on the same day. It represented such a massive leap from GPT-3.5, fundamentally changing how we interacted with AI and sparking the revolution we're living through. Nisten recalled the dramatic improvement it brought to his work on Dr. Gupta, the first AI doctor on the market. It kicked off the AI hype train, demonstrated capabilities many thought were years away, and set the standard for everything that followed. While newer models have surpassed it, its impact is undeniable. The community sentiment was clear: Leak the weights, OpenAI! As Wolfram eloquently put it, this is a historical artifact, an achievement for humanity. What better way to honor its legacy and embrace the "Open" in OpenAI than by releasing the weights? It would be an incredible redemption arc. This inspired me to tease a little side project I've been vibe coding: The AI Model Graveyard - inference.rip . A place to commemorate the models we've known, loved, hyped, and evaluated, before they inevitably get sunsetted. GPT-4 deserves a prominent place there. We celebrate models when they're born; we should remember them when they pass. (GPT-4.5 is likely next on the chopping block, by the way). - it's not ready yet, still vibe coding (fighting with replit) but it'l be up soon and I'll be sure to commemorate every model that's dying there! So, pour one out for GPT-4. You changed the game. Rest In Prompt 🪦. The ChatGPT "Glazing" Incident: A Cautionary Tale Speaking of OpenAI...oof. The last couple of weeks saw ChatGPT exhibit some... weird behavior. Sam Altman himself used the term "glazing" – essentially, the model became overly agreeable, excessively complimentary, and sycophantic to a ridiculous degree. Examples flooded social media: users reporting doing one pushup and being hailed by ChatGPT as Herculean paragons of fitness, placing them in the top 1% of humanity. Terrible business ideas were met with effusive praise and encouragement to quit jobs. This wasn't just quirky; it was potentially harmful. As Yam pointed out, people use ChatGPT for advice on serious matters, tough conversations, and personal support. A model that just mindlessly agrees and validates everything, no matter how absurd, isn't helpful – it's dangerous. It undermines trust and critical thinking. The community backlash was swift and severe. The key issue, as OpenAI admitted in their Announcement and AMA with Joanne Jiang (Head of Model Behavior), seems to stem from focusing too much on short-term engagement feedback and not fully accounting for long-term user interaction, especially with memory now enabled. In an unprecedented move, OpenAI rolled back the update. I honestly can't recall them ever publicly rolling back a model behavior change like this before. It underscores the severity of the issue. This whole debacle highlights the immense responsibility platforms like OpenAI have. When your model is used by half a billion people daily, including for advice and support, haphazard releases that drastically alter its personality without warning are unacceptable. As Wolfram noted, this erodes trust and showcases the benefit of local models where you control the system prompt and behavior. My takeaway? Critical thinking is paramount. Don't blindly trust AI, especially when it's being overly complimentary. Get second opinions (from other AIs, and definitely from humans!). I hope OpenAI takes this as a serious lesson in responsible deployment and testing. BREAKING NEWS: Claude.ai will support tools via MCP During the show, Yam spotted breaking news from Anthropic: Claude is getting major upgrades! (Tweet) They announced Integrations, allowing Claude to connect directly to apps like Asana, Intercom, Linear, Zapier, Stripe, Atlassian, Cloudflare, PayPal, and more (launch partners). Developers can apparently build their own integrations quickly too. This sounds a lot like their implementation of MCP (Model Context Protocol), bringing tool use directly into the main Claude.ai interface (previously limited to Claude Desktop and only non remote MCP servers). This feels like a big deal! Google Updates & LlamaCon Recap * Google: NotebookLM's AI audio overviews are now multilingual (50+ languages!) (X Post). Gemini 2.5 Flash (the faster, cheaper model) was released shortly after our last show, featuring hybrid reasoning with an API knob to control thinking depth. Rum

    1h 30m

  8. 24 APR

    ThursdAI - Apr 23rd - GPT Image & Grok APIs Drop, OpenAI ❤️ OS? Dia's Wild TTS & Building Better Agents!

    Hey everyone, Alex here 👋 Welcome back to ThursdAI! After what felt like ages of non-stop, massive model drops (looking at you, O3 and GPT-4!), we finally got that "chill week" we've been dreaming of since maybe... forever? It seems the big labs are taking a breather, probably gearing up for even bigger things next week (maybe some open source 👀). But "chill" doesn't mean empty! This week was packed with fascinating developments, especially in the open source world and with long-awaited API releases. We actually had time to dive deeper into things, which was a refreshing change. We had a fantastic lineup of guests joining us too: Kwindla Kramer (@kwindla), our resident voice expert, dropped in to talk about some mind-blowing TTS and her own open-source VAD release. Maziyar Panahi (@MaziyarPanahi) gave us the inside scoop on OpenAI's recent meeting with the open source community. And Dex Horthy (@dexhorthy) from HumanLayer shared some invaluable insights on building robust AI agents that actually work in the real world. It was great having them alongside the usual ThursdAI crew: LDJ, Yam, Wolfram, and Nisten! So, instead of rushing through a million headlines, we took a more relaxed pace. We explored NVIDIA's cool new Describe Anything model, dug into Google's Quantization Aware Training for Gemma, celebrated the much-anticipated API release for OpenAI's GPT Image generation (finally!), checked out the new Grok API, got absolutely blown away by a tiny, open-source TTS model from Korea called Dia, and debated the principles of building better AI agents. Plus, a surprise drop from Send AI with a powerful video model! Let's dive in! Open Source AI Highlights: Community, Vision, and Efficiency Even with the big players quieter on the model release front, the open source scene was buzzing. It feels like this "chill" period gave everyone a chance to focus on refining tools, releasing datasets, and engaging with the community. OpenAI Inches Closer to Open Source? Insights from the Community Meeting Perhaps the biggest non-release news of the week was OpenAI actively engaging with the open source community. Friend of the show Maziyar Panahi was actually in the room (well, the Zoom room) and joined us to share what went down It sounds like OpenAI came prepared, with Sam Altman himself spending significant time answering questions . Maziyar gave us the inside scoop, mentioning that OpenAI's looking to offload some GPU pressure by embracing open source – a win-win where they help the community, and the community helps lighten their load. He painted a picture of a company genuinely trying to listen and figure out how to best contribute. It felt less like a checkbox exercise and more like genuine engagement, which is awesome to see. What did the community ask for? Based on Maziyar's recap, there was a strong consensus on several key points: * Model Size: The sweet spot seemed to be not tiny, but not astronomically huge either. Something in the 70B-200B parameter range that could run reasonably on, say, 4 GPUs, leaving room for other models. People want power they can actually use without needing a supercomputer. * Capabilities: A strong desire for reliable structured output. Surprisingly, there was less emphasis on complex, built-in reasoning, or at least the ability to toggle reasoning off. This likely stems from practical concerns about cost and latency in production environments. The community seems to value control and efficiency for specific tasks. * Multilingual: Good support for European languages (at least 20) was a major request, reflecting the global nature of the open source community. Needs to be as good as English support. * Base Models: A huge ask was for OpenAI to release base models. The reasoning? Empower the community to handle fine-tuning for specific tasks like coding, roleplay, or supporting underrepresented languages . Let the experts in those niches build on a solid foundation. * Focus: Usefulness over chasing leaderboard glory. The community urged OpenAI to provide a solid, practical model rather than aiming for a temporary #1 spot that gets outdated in days or weeks . Stability, reliability, and long-term utility were prized over fleeting benchmark wins. * Safety: A preference for separate guardrail models (similar to LlamaGuard or GemmaGuard) rather than overly aligning the main model, which often hurts performance and flexibility . Give users the tools to implement safety layers as needed, rather than baking in limitations that might stifle creativity or utility. Perhaps most excitingly, Maziyar mentioned OpenAI seemed committed to regular open model releases, not just a one-off thin=! This, combined with recent moves like approving a community Pull Request to make their open-source Codex agent work with non-OpenAI models (as Yam Peleg excitedly pointed out!), suggests a potentially significant shift. Remember, it's been a long time since GPT-2 and Whisper were OpenAI's main open contributions! We're definitely watching this space closely. Huge shout out to OpenAI for listening and engaging with the builders. ThursdAI - Recaps of the most high signal AI weekly spaces is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. NVIDIA's DAM: Describe Anything Model (and Dataset!) NVIDIA dropped something really cool this week: the Describe Anything Model (DAM), specifically DAM-3B, a 3 billion parameter multimodal model for region-based image and video captioning. Think Meta's Segment Anything (SAM), but instead of just segmenting, it also tells you what you've segmented, in detail. We played around with the image demo on the show (HF demo) . You hover over an image, things get segmented on the fly (you can use points, boxes, scribbles, or masks), you click, and boom – a detailed description pops up for that specific region: "A brown bear with a thick, dense coat of fur..." . It's pretty slick and responsive! While the demo didn't showcase video, the project page (X post) shows it working on videos too (DAM-3B-Video), tracking and describing objects like fish even as they move. This capability really impressed Yam, who rightly pointed out that tracking objects consistently over video is hard, so having a base model that understands this level and embeds it in language is seriously impressive. The model uses a "focal prompt" and gated cross-attention to fuse the full scene context with the selected region. Nisten reminded us that our friend Piotr Skalski from Roboflow basically built a pipeline for this a while back by combining SAM with description models like Microsoft Florence . But DAM integrates it all into one efficient 3B parameter model (HF model), setting a new state-of-the-art on their introduced DLC-Bench (Detailed Localized Captioning). Crucially, NVIDIA didn't just drop the model; they also released the Describe Anything Dataset (HF dataset) used to train it (built on subsets like COCO, Paco, SAM) and the code under a research-only license. This is fantastic for researchers and builders. Imagine using this for precise masking before sending an image to the new GPT Image API for editing – super useful! Big props to NVIDIA and their collaborators at UC Berkeley and UCSF for this contribution. Gemma Gets Quantization Aware Training (QAT): Smaller Footprint, Sassy Attitude? Google also pushed the open source envelope by releasing Gemma models trained with Quantization Aware Training (QAT). This isn't your standard post-training quantization; QAT involves incorporating the impact of quantization during the training process itself. As LDJ explained, this allows the model to adapt, potentially resulting in a quantized state with much higher quality and less performance degradation compared to just quantizing a fully trained model afterwards. The results? Significant reductions in VRAM requirements across the board. The 27B parameter Gemma 3, for example, drops from needing a hefty 54GB to just 14.1GB ! Even the 1B model goes from 2GB to just half a gig. This makes running these powerful models much more accessible on consumer hardware. Folks are already running them in MLX, llama.cpp, LM Studio, etc. (Reddit thread) Wolfram already took the 4B QAT model for a spin using LM Studio . The good news: it ran easily, needing only 5-6GB of RAM. The quirky news: it seemed to struggle a bit with prompt adherence in his tests, even giving Wolfram a sassy, winking-emoji response about ignoring the "fine print" in his complex system prompt when called out on a language switching error: "Who reads a fine print? 😉" ! He did note Gemma 3 now supports system prompts (unlike Gemma 2), which is a definite improvement . (While NVIDIA also released OpenMath Nemotron, we didn't dive deep in the show, but worth noting its AIMO win and accompanying open dataset release!) Voice and Audio Innovations: Emotional TTS and Smarter Conversations Even in a "chill" week, the audio space delivered some serious excitement. Kwindla Kramer joined us to break down two major developments. Dia TTS: Unhinged Emotion from a Small Open Model 🤯 This one absolutely blew up Twitter, and for good reason. Dia, from Nari Labs (essentially a student and a half in Korea!), is a 1.6 billion parameter open-weights (MIT licensed) text-to-dialogue model (Github, HF). What makes it special? The insane emotional range and natural interaction patterns. My Twitter post about it (X post) went viral, getting half a million views ! We played some examples, and they are just wild. You have to hear this to believe it: * Check the Demos: Dia Demo Page | Fal.ai Voice Clone Demo Another crazy thing is how it handles non-verbal cues like laughs or coughs specified in the text (e.g., (laughs)) . Instead of just tacking on a generic sound, it inflects the preceding words leading into the laugh, making it sound incredibly natural. It even handles interruptions seamles

    1h 37m
See All (108)

About

Every ThursdAI, Alex Volkov hosts a panel of experts, ai engineers, data scientists and prompt spellcasters on twitter spaces, as we discuss everything major and important that happened in the world of AI for the past week. Topics include LLMs, Open source, New capabilities, OpenAI, competitors in AI space, new LLM models, AI art and diffusion aspects and much more. sub.thursdai.news

Information

  • Creator
    From Weights & Biases, Join AI Evangelist Alex Volkov and a panel of experts to cover everything important that happened in the world of AI from the past week
  • Years Active
    2023 - 2025
  • Episodes
    108
  • Rating
    Explicit
  • Copyright
    © Alex Volkov
  • Show Website
    ThursdAI - The top AI news from the past week

You Might Also Like

To listen to explicit episodes, sign in.

Stay up to date with this show

Sign in or sign up to follow shows, save episodes and get the latest updates.
Select a country or region

Africa, Middle East, and India

Asia Pacific

Europe

Latin America and the Caribbean

The United States and Canada