0:00 Alibaba just led a two-hundred-and-ninety-million-dollar funding round for a Chinese startup named Shengshu.0:08 It’s a massive bet, and it’s not for another chatbot.0:12 The company is building what it calls “general world models.” This isn’t about generating better text or images.0:21 It’s about enabling machines—robots, autonomous cars, factory arms—to understand the physical world.0:29 This investment signals a fundamental pivot.0:32 The consensus among the giants is forming: the race to scale up large language models is hitting a point of diminishing returns.0:42 The next frontier isn't about language.0:45 It's about physics.0:47 That bet on the physical world is the most important signal today, but it's not the only one.0:54 The entire landscape is shifting away from the models that defined the last few years.1:01 Meta just confirmed it's replacing its open-source Llama line.1:05 The replacement is a new model called Muse Spark.1:09 Unlike Llama, Muse Spark is closed.1:12 It’s designed for agent-based systems that can handle multi-step reasoning.1:18 It processes text and images, but its purpose is not to be a standalone tool for developers.1:25 Its purpose is to be integrated directly into Meta's products to perform actions on a user's behalf.1:33 The open-source chapter at Meta appears to be closing.1:37 The new strategy is building agents that live inside the walled garden.1:43 Meanwhile, in biotech, a new company called Xaira Therapeutics just launched with over one billion dollars in funding.1:52 It was founded by Marc Tessier-Lavigne, the former chief scientific officer of Genentech, a giant in the space.2:01 Xaira is designed from the ground up to use AI for drug discovery.2:06 What makes this billion-dollar figure so significant is the context.2:11 Global venture funding for health-related AI peaked at twenty-two billion dollars in 2021.2:18 By 2024, that number was cut by more than half, down to ten-point-five billion.2:24 Investors are no longer writing checks based on hype.2:28 They are demanding evidence.2:31 Xaira’s ability to raise this much capital shows that while the bar is now much, much higher...2:38 the conviction about AI's potential to remake pharma is still there for teams with the right pedigree.2:46 These moves at the frontier are happening alongside a more complicated reality on the ground.2:53 A new analysis of the "AI productivity paradox" is gaining traction.2:59 It finds that the time employees spend supervising AI and fixing its errors can completely offset productivity gains.3:08 The author, Rubie Ullrich, puts it plainly: if you measure productivity by lines of code or pages produced, AI looks like a miracle.3:18 But if you measure outcomes—revenue, decision quality, risk reduction—the picture becomes far more nuanced.3:26 Human oversight remains the critical, and expensive, component.3:31 This friction is also appearing in the public sector.3:35 AI tools are becoming the default "front door" for citizens seeking government services.3:42 The problem is, most official guidance on taxes, unemployment, or food assistance isn't formatted in a way that AIs can reliably read.3:53 So the chatbots give confident, but often incomplete or simply wrong, answers.3:59 This isn't a technical glitch.4:01 It's a systemic failure to prepare the foundation before building the house.4:07 So let's connect the two most important moves today.4:11 Alibaba's two-hundred-and-ninety-million-dollar bet on world models, and Meta's pivot to the closed, agentic Muse Spark model.4:21 These are not isolated events.4:23 They are two sides of the same coin, and that coin represents the end of an era.4:30 For the last several years, the dominant idea in AI was scaling.4:35 Make the models bigger, feed them more data, and emergent capabilities will appear.4:41 It worked.4:42 But industry leaders, from OpenAI's Sam Altman on down, now admit the returns from pure scaling are diminishing.4:51 Large language models can write a decent email, but they are spectacularly bad at tasks requiring an understanding of cause and effect.5:01 They have no concept of physics, spatial reasoning, or how objects interact.5:07 You can’t build a self-driving car or a warehouse robot on a system that just predicts the next word.5:15 This is the turn.5:16 The problem everyone was trying to solve was human language.5:21 The solution everyone was building—the LLM—is now being reframed.5:26 It’s no longer the destination.5:29 It’s a component.5:30 Alibaba's investment in Shengshu is a direct pursuit of the next paradigm.5:36 A "world model" is an attempt to teach an AI the underlying rules of reality.5:42 It's a simulation of physics, not just a statistical map of text.5:47 The goal, which McKinsey estimates could be a four trillion dollar market, is to give machines common sense about the physical world.5:57 This is the foundation for robotics and true automation.6:02 Meta's move with Muse Spark is the consumer-facing version of the same strategic shift.6:08 By creating a closed, agentic model, they are building a system designed to do things.6:15 An agent doesn't just answer your question.6:19 It coordinates a multi-step task: it finds the flight, compares prices from three vendors, checks your calendar, and then asks for permission to book it.6:30 The language model is just one piece of that chain—the part that understands your initial request.6:38 The real work is done by the reasoning and action components that Meta is now building in-house and keeping proprietary.6:47 What we are witnessing is the great refactoring of the AI stack.6:53 The LLM is being demoted from the brain of the operation to the user interface.6:59 This is the difference between a calculator and a spreadsheet.7:04 One gives you an answer.7:05 The other gives you a system to work within.7:09 The entire industry is now trying to build the spreadsheet.7:13 This brings us back to the friction on the ground.7:17 The ambition at the frontier of AI has never been greater.7:22 We are seeing billion-dollar bets on creating new medicines and quarter-billion-dollar bets on creating models that understand reality itself.7:33 These are attempts to solve fundamental challenges in science and engineering.7:39 At the very same time, the AI we already have is struggling with much simpler tasks.7:46 The productivity paradox shows us that using AI to write reports isn't a simple win.7:52 It creates a new category of work: AI supervision.7:56 The most important skill is no longer writing the first draft, but having the discernment to know when the AI's plausible-sounding output is subtly wrong or completely detached from the constraints of the real world.8:13 That same detachment is what's causing AI chatbots to fail as a front door to government.8:20 The model doesn't know what it doesn't know.8:23 It can't tell the difference between official, machine-readable data and a random webpage that happens to mention unemployment benefits.8:34 It just produces a confident paragraph.8:37 The system offers enormous upside, but only if humans retain the ability to steer and refine it.8:45 The landscape today is defined by this split.8:48 The most advanced research and investment capital is focused on giving AI agency and a grasp of the physical world.8:57 But the practical application of yesterday's models is teaching us a lesson in humility.9:04 AI accelerates parts of a workflow, but it cannot replace the accountability required to produce good work.9:12 The future is being built on a foundation that is still proving to be unstable.