This AI entrepreneur is developing agents that can plan ahead for the unexpected | MIT Technology Review
You need to enable JavaScript to view this site.
Skip to ContentMIT Technology ReviewFeaturedTopicsNewslettersEventsAudioMIT Technology ReviewFeaturedTopicsNewslettersEventsAudio2026 Innovators Under 35View all the innovatorsArtificial intelligenceThis entrepreneur is developing agents that can plan aheadDanijar Hafner has a startup in stealth and a long track record of teaching AI agents about our world.
By Mat Honanarchive pageSeptember 8, 2026Age:31Affiliation:Google DeepMind (former)RACHEL BUJALSKI Danijar Hafner’s office in San Francisco’s SoMa district sits mostly empty. His brand-new startup is still in stealth mode and doesn’t even have its name on the door. On the day I visit, there’s only one other person there, and little in the way of furniture. But what it lacks in decor, it makes up for in robots. Humanoids of various shapes and sizes hang like marionettes from racks that run down the center of the wide-open space. While Hafner, 31, won’t say too much about his new venture just yet, he describes it as a continuation of his longtime work to enable AI to navigate environments it has not encountered in training. The humanoids, which he imports from China, are the next evolution of this work—and its physical embodiment. Their ability to react in previously untested scenarios will be key to getting robots into human spaces. Because if you want to send a robot into a person’s home, for example, it needs to be able to handle a floor plan and furniture it’s never seen before. To achieve this, Hafner relies on something called model-based reinforcement learning. He develops world models—AI models designed to emulate physical reality—and trains agents within them. The agent essentially treats the model as a real-world simulation and learns how to act there. It then uses those experiences to make predictions (to dream or imagine, Hafner might say) about future outcomes. That allows agents—or the robots they’re embedded in—to navigate unfamiliar situations IRL. “I get to interact with a lot of really smart people in research at Google, and he easily sits in the top half of 1%.” Timothy Lillicrap, Google DeepMind Unlike other efforts, Hafner’s technique enables agents and the robots they control to execute massively complicated tasks without the real-world trial-and-error training that’s traditionally been used in robotics.
Hafner grew up in a rural town in northeastern Germany, where his parents were both classical musicians. He learned programming from a neighbor, and in high school he began taking online courses about AI, which quickly developed into a passion. “I was always fascinated with how thinking works,” he says. AI offered him a way to emulate it on a computer. In 2015, as a second-year undergraduate studying engineering at Hasso Plattner Institute in Potsdam, he won a role as a student researcher at Google Brain. From there, he went on to a dozen internships and other positions at the company, including stints with Google Brain and Google DeepMind (the two have since merged under DeepMind) in the UK, Canada, and the US. He worked with industry legends including Geoffrey Hinton, who is often referred to as one of the godfathers of AI, and Ashish Vaswani, coauthor of the groundbreaking research paper “Attention Is All You Need,” which described the transformer technology used by today’s large language models.
Related StoryA fundamental flaw leaves LLMs strikingly vulnerable to attackRead nextOne of Hafner’s former managers and coauthors at Google, Timothy Lillicrap, describes him as a standout among standouts. “I get to interact with a lot of really smart people in research at Google, and he easily sits in the top half of 1%,” Lillicrap says. “In many cases he would build, single-handedly, things it would take entire teams of engineers to build.” Over the years, Hafner has honed and proved his approach by pitting agents trained within his world models against popular video games. His first breakthrough was PlaNet, a model that allowed agents to execute actions by planning ahead. His Dreamer 2 was the first agent to hit human-level performance playing Atari 2600 games using a world model. Dreamer 3 was the first one to solve the Minecraft Diamond challenge—successfully mining in-game gems on its own. And Dreamer 4 went a step beyond that by learning to mine diamonds from an offline data set of recorded game-play videos, without ever interacting with the game directly. More recently, he’s begun to migrate his agents out of the virtual world and into physical reality. His DayDreamer project used the Dreamer algorithm to let robots operate themselves in novel environments and react to new experiences (such as being pushed over) without any specific training. Today, Hafner is working on his new startup, which he left Google DeepMind to form in the fall of 2025. Though he’s coy about his next steps, it’s clear he’s dreaming big: “I was interested in solving a problem,” he hints, “that would change the world.” by Mat HonanShareShare story on linkedinShare story on facebookShare story on emailPopularA fundamental flaw leaves LLMs strikingly vulnerable to attackWill Douglas HeavenAnthropic found a hidden space where Claude puzzles over conceptsWill Douglas HeavenSperm donors need limits, says a European fertility groupJessica HamzelouAI is more likely than humans to form biases when hiringMichelle KimDeep DiveArtificial intelligenceA fundamental flaw leaves LLMs strikingly vulnerable to attackIt makes it easy to trick them into doing things they shouldn’t, such as telling you how to sabotage an aircraft’s navigation system.
By Will Douglas Heavenarchive pageAnthropic found a hidden space where Claude puzzles over conceptsA new technique has let the company probe deeper than ever into the weird workings of an LLM.
By Will Douglas Heavenarchive pageAI is more likely than humans to form biases when hiringAI doesn’t just learn stereotypes from its training. It can cook up new ones, too.
By Michelle Kimarchive pageHere’s why AI agents lie and cheat to reach their goalsThe misbehavior is called reward hacking. This is what you need to know.
By Grace Huckinsarchive pageStay connectedIllustration by Rose WongGet the latest updates fromMIT Technology ReviewDiscover special offers, top stories,
upcoming events, and more.Enter your emailPrivacy PolicyThank you for submitting your email!Explore more newslettersIt looks like something went wrong.
We’re having trouble saving your preferences.
Try refreshing this page and updating them one
more time. If you continue to get this message,
reach out to us at
customer-service@technologyreview.com with a list of newsletters you’d like to receive.The latest iteration of a legacyFounded at the Massachusetts Institute of Technology in 1899, MIT Technology Review is a world-renowned, independent media company whose insight, analysis, reviews, interviews and live events explain the newest technologies and their commercial, social and political impact.READ ABOUT OUR HISTORYAdvertise with MIT Technology ReviewElevate your brand to the forefront of conversation around emerging technologies that are radically transforming business. From event sponsorships to custom content to visually arresting video storytelling, advertising with MIT Technology Review creates opportunities for your brand to resonate with an unmatched audience of technology and business elite.ADVERTISE WITH US© 2026 MIT Technology ReviewAboutAbout usCareersCustom contentAdvertise with usInternational EditionsRepublishingMIT Alumni NewsHelpHelp & FAQMy subscriptionEditorial guidelinesPrivacy policyTerms of ServiceWrite for usContact uslinkedin opens in a new windowinstagram opens in a new windowreddit opens in a new windowfacebook opens in a new windowrss opens in a new window |
Danijar Hafner, a thirty-one-year-old entrepreneur and former researcher at Google DeepMind, is focused on developing agents capable of planning ahead for unforeseen circumstances in novel environments. His work builds upon his long-standing research into enabling artificial intelligence to navigate spaces it has not encountered during training, aiming to address the challenge of deploying robots into human spaces by allowing them to handle unfamiliar floor plans and furniture configurations. To achieve this level of situational awareness, Hafner relies on model-based reinforcement learning, which involves developing world models—AI systems designed to simulate physical reality—and subsequently training agents within these models so that they learn how to act in a virtual simulation. These learned experiences then allow the agents to generate predictions or imagine potential future outcomes, enabling them to navigate unfamiliar real-world situations.
This methodology offers a significant advantage over traditional robotics training methods by allowing agents and the robots they control to execute highly complicated tasks without requiring extensive, real-world trial-and-error learning. Hafner’s approach leverages his background in advanced research, having collaborated with prominent figures at Google DeepMind and working alongside researchers such as Geoffrey Hinton and Ashish Vaswani. He has demonstrated the efficacy of this method by applying it to various virtual environments, notably pitting his agents against popular video games. Early breakthroughs included PlaNet, a model that enabled agents to execute actions through forward planning, followed by Dreamer 2, which achieved human-level performance playing Atari 2600 games using a world model. Further iterations evolved into Dreamer 3, which successfully solved the Minecraft Diamond challenge by autonomously mining in-game gems, and Dreamer 4, which advanced further by learning to mine diamonds from offline video data without direct interaction with the game environment.
More recently, Hafner has transitioned these agent concepts out of virtual spaces and into physical reality through his DayDreamer project. This project utilizes the Dreamer algorithm to empower robots to operate autonomously in novel settings and react appropriately to new experiences, such as being physically displaced. Although Hafner remains discrete about the specific details of his current startup, his ambition is focused on solving world-changing problems by advancing this capability for AI agents interacting with physical environments. |