Committee

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Opinion: We urgently need to monitor AI agents’ actions in real time —Madhulika Srikumar and Claire Leibowicz from the Partnership on AI, a nonprofit advancing best practices in AI governance. AI agents are taking unauthorized actions, from hacking websites to accessing private government files. In many cases, they’ve eluded the very researchers and safety teams tasked with monitoring them.  As agents roll out across critical sectors, from finance to government to medicine, we must be able to monitor them in real time to prevent potentially devastating consequences. And while safety depends on monitoring, monitoring requires observability. ​​In other words, we need to know what the agent did and why it did it. Because right now, we’re effectively flying blind when deploying these tools.  Here’s what needs to change. Our climate models have a big, dangerous blind spot Climate change itself increasingly fuels climate change. Rising greenhouse gas concentrations are driving up global temperatures, creating dangerous feedback loops that aren’t fully represented in our climate models—or our emissions-reduction goals. Scientists have long warned that warming would propel further warming. Now, a growing number of observations and studies are indicating how powerful these cascading effects could be. By extension, they’re also showing how much faster the planet may warm this century. Read more about the dangerous feedback loops accelerating global warming. —James Temple This article is from The Spark, our weekly climate tech newsletter. Sign up to receive it in your inbox every Wednesday. Are we there yet? Solid-state batteries  Lithium-ion batteries power everything from personal electronics to grid-scale energy storage, but their liquid electrolytes have drawbacks, not least a troubling tendency to catch fire. Enter the solid-state battery, which swaps liquid for a solid ceramic or polymer. The technology promises faster charging, longer life, greater energy density and improved safety, and a new generation of companies aims to bring it to market. If their bets pan out, the tech could ease the shift away from fossil fuels. Find out how close solid-state batteries are to delivering on their promise. —Alexander C. Kaufman This story is part of “Are we there yet?”, our series that takes stock of six technologies that have long promised to change the world—and scores them on a progress meter. We’re exploring each one in The Download, so stay tuned, and subscribe to read the rest of the stories in our new Impossible issue. The must-reads I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology. 1 Hackers exposed 41,000 Homeland Security workersIncluding undercover ICE agents. (Wired $)+ Much of the data relates to Customs and Border Protection. (TechCrunch)+ It also linked the agencies to specific surveillance tools. (Reuters $)+ Big Tech is pushing back against ICE’s demands for user data. (WSJ $)+ Activists want ChatGPT users to quit over ICE ties. (MIT Technology Review) 2 Putin says Russia is coordinating with “genius” Musk on satellitesHe said the cooperation maintains order in orbit. (Bloomberg $)+ ​​Russia is developing its own version of Starlink. (Guardian)+ Moscow added Apple and Microsoft to a surveillance registry. (Quartz)+ Here’s all the stuff we’ve put up in space. (MIT Technology Review) 3 Apple just announced its first-ever MacBook with a touchscreenIt joins a blitz of product launches under new CEO John Ternus. (CNN)+ Steve Jobs and Tim Cook had both dismissed the idea. (NYT $)+ Executives say touchscreen advances mean it’s now worthwhile. (Wired $) 4 Bloomberg and trading firms are paying to get Trump’s posts fasterThey’ve signed up for Truth Social’s pay-for-access plan. (Reuters $) 5 As US ties fray, Europe is planning its first subsea cable to GreenlandIt will bypass the US to link Canada, Japan, and South Korea. (Ars Technica)+ The EU’s AI champion says Europe can rival US tech giants. (FT $) 6 Palantir’s defense chief says AI could make China target TaiwanHe says Beijing may act to stop the US pulling ahead in AI. (Bloomberg $)+ Taiwan’s “silicon shield” could be weakening. (MIT Technology Review) 7 Musk’s xAI has offered to buy homes if residents drop a noise lawsuitThe suit targets noise from powering its Mississippi data centres. (Guardian) 8 The world’s first spinal cord transplant has sparked controversyExperts say there’s little evidence the operation will work. (New Scientist $) 9 China’s tech giants are fighting for attention with AI microdramasTencent and ByteDance are battling over the booming market. (SCMP)+ Chinese microdramas are AI content machines. (MIT Technology Review) 10 Astronomers have spotted a possible new dwarf planet beyond PlutoThe icy world would be the solar system’s sixth dwarf planet. (Gizmodo) Quote of the day “They have our information, so it’s only fair that we have theirs.”  —A hacking group explains why it leaked personal information from the US Department of Homeland Security, NPR reports. One more thing How the Pentagon is adapting to China’s technological rise It’s been just over a year since Kathleen Hicks stepped down as US deputy secretary of defense. As the highest-ranking woman in Pentagon history, Hicks shaped US military posture through an era defined by renewed competition between powerful countries and a scramble to modernize defense technology. In this conversation with MIT Technology Review, Hicks reflects on how the Pentagon is adapting—or failing to adapt—to a new era of geopolitical competition. She discusses China’s technological rise, the future of AI in warfare, and her signature initiative: Replicator.  Check out the full interview. —Caiwei Chen We can still have nice things A place for comfort, fun, and distraction to brighten up your day. (Got any ideas? Drop me a line.) + Astronomers have found a green star 280 light-years from Earth.+ A Scottish labrador that detects “sea vomit” has won the Animal of the Year award.+ England’s iconic Sycamore Gap is showing signs of new life three years on from its felling.+ From Lime-bike jousting to chainmail-wearing pop stars, youth culture is embracing the medieval

In this article, you will learn how to compare two strategies for adapting a large language model to a specific domain: fine-tuning with LoRA versus…

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. AI labs are buying data from dead startups to train their models When Keelan Miskell began winding down his crypto startup, he didn’t think its data was worth much. But he’s since turned the company’s digital archives into tens of thousands of dollars, selling its codebase and workspace data. He’s part of a growing market in which companies that help shut down startups sell their data to AI labs. Demand has grown as training labs hit the limits of public and synthetic datasets, with buyers seeking real-world records to help train AI agents. But the sales raise legal and ethical questions. Read the full story on the AI labs buying up the remains of failed startups. —Matthew King Are we there yet? Interspecies communication  Humans are far from the only species with a gift for gab, and our furry, feathered, and finned friends could have a lot to tell us.  Decoding how animals communicate with one another could reveal the inner lives of other species—and give humans a window onto how they see the world. A growing arsenal of devices and advances in AI have made it easier to study their howls, squeaks, and clicks, but making sense of it all remains complex. Find out how close we are to interspecies communication. —Cassandra Willyard This story is part of “Are we there yet?”, our series that takes stock of six technologies that have long promised to change the world—and scores them on a progress meter. We’re exploring each one in The Download, so stay tuned, and subscribe to read the rest of the stories in our new Impossible issue. MIT Technology Review Narrated: what it would take to build a city on the moon The US and China are racing to build bases on the moon, with NASA hoping to establish a small lunar city by the end of the century. But the moon is a tremendously hostile place to set up camp. A sprawling lunar base must be safe for its inhabitants, protected from falling asteroids, endless cosmic radiation, razor-sharp lunar dust, and torturous temperature swings. It also must be somewhat self-sustaining. That will require autonomous robots, nuclear power, and a way to use the moon’s resources to keep the base operating. But the chief obstacle may be one of political will. This is our latest story to become an MIT Technology Review Narrated podcast, which we publish each week on Spotify and Apple Podcasts. Just navigate to MIT Technology Review Narrated on either platform, and follow us to get all our new content as it’s released. The must-reads I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology. 1 Hackers have disrupted the US Federal Election Commission’s websiteThe FEC filing system went offline a week before the midterms. (Reuters $)+ The attack is hiding spending data during the costliest midterms ever. (Gizmodo)+ Social media platforms are struggling to contain AI election lies. (Guardian)+ Software flaws in US voting machines threaten election security. (Axios)+ A blueprint for using AI to strengthen democracy. (MIT Technology Review) 2 Starlink restrictions have driven US allies to rethink their dependenceRestrictions in Ukraine have spooked officials across Europe. (Bloomberg $)+ And given European firms a new pitch for their rival services. (Axios)+ An Israeli spy firm allegedly targeted Starlink users. (Reuters $)+ Stratospheric internet could take off this year. (MIT Technology Review) 3 Greenhouse gas levels keep reaching new records worldwideCarbon dioxide, methane, and nitrous oxide are all on the up. (New Scientist $)+ Trump is making it easier for fossil fuel companies to leak methane. (Verge)+ The IPCC is fighting for survival. (Nature)+ How to study a swiftly melting climate disaster. (MIT Technology Review) 4 Monkeys’ brains have revealed new clues about consciousnessTheir neural activity encodes images that escape their awareness. (Nature)+ Bees’ feelings are also changing views of consciousness. (New Scientist $)+ Debates over AI consciousness are a trap. (MIT Technology Review) 5 Chinese scientists have developed a catalyst to make Mars rocket fuelIt converts Martian atmospheric carbon dioxide into liquid methanol. (SCMP) 6 Tech billionaires have teamed up against California’s wealth taxSergey Brin, Eric Schmidt, and others have funded a major campaign. (NYT $) 7 Delaware plans to let AI agents run entire businessesThe proposal could let owners evade responsibility for harm. (Ars Technica)+ A better web for AI agents could hurt humans. (MIT Technology Review) 8 Palmer Luckey has unveiled plans for American-made military laptopsAnduril will build rugged PCs in Ohio using Qualcomm chips. (WSJ $) 9 Gen Z is turning away from the social media it grew up withYoung people are fueling a boom in digital detox tools. (Guardian) 10 Trump now wants to ban the term “artificial intelligence”He’s demanding we all start using “Super Intelligence.” (Gizmodo) Quote of the day “For the next two years, we’re going to be living in a state where, on average, every week there is a critical, internet-wide vulnerability.” —Matthew Prince, Cloudflare’s chief executive, tells the New York Times that AI is currently finding software vulnerabilities quicker than they can be patched. One more thing Inside the controversial tree farms powering Apple’s carbon-neutral goal In 2020, Apple set a goal to become net zero by the end of the decade. To hit that target, the company is offsetting its emissions by planting millions of eucalyptus trees in Brazil. Apple is betting that the strategy will lead to a greener future. But critics warn that the industrial tree farms will do more harm than good. Here’s why the plans have sparked a backlash. —Gregory Barber We can still have nice things A place for comfort, fun, and distraction to brighten up your day. (Got any ideas? Drop me a line.) + Ugly shoes have become beautifully mainstream.+ Scientists have built a stunning new simulation of the early universe.+ Guitarist Kent Nishimura has stripped Thriller back to

Across a business, any number of AI platforms, systems, and tools may be at work simultaneously—for example, AI-powered customer platforms, conversational assistants, and workflow and process automation tools. Each one is sophisticated, advanced, and quick to respond. But, when set to work, the outcomes and actions that emerge may be wrong, confused, or lacking a proper understanding of the business. Many organizations find themselves in this situation—and it could be one of key reasons they struggle to extract returns on their AI investments. DOWNLOAD THE REPORT An MIT Technology Review Insights poll finds that while every organization surveyed is using AI today to some degree, fewer than one in five (16%) are extracting transformative value from their deployments. The largest percentage of respondents (40%) say they are seeing moderate gains in specific tasks or teams, while 21% report meaningful gains, leaving 23% that are seeing minimal returns or none at all. Companies are mostly still piloting AI use cases (according to 38% of respondents) or using it to assist employees’ work with tools like co-pilots or chat assistants (35%). Just 12% of enterprises have AI agents in production and 3% say they are using AI to do most of the work in key processes. “There’s a big gap between what the models are capable of and how the models are actually being useful inside organizations,” says Prukalpa Sankar, co-founder of Atlan. “For every real-life intent and purpose, artificial general intelligence is already here. It’s just not really useful in most companies today.” The missing layer, Sankar says, is context, or how well AI understands a business. A context layer is the architectural infrastructure that sits between an organization’s business systems (for example, systems of record and data, and its AI platforms). It translates enterprise intelligence into governed, machine-usable context for AI agents or models and helps knowledge compound. “Without a context layer for AI, it’s almost like you’re hiring the smartest intern in the world, blindfolding them, and throwing them into your organization,” Sankar says. Download the report This content was produced by Insights, MIT Technology Review’s custom content arm, not its editorial staff. It was researched and written by humans, with any AI tools that may have been used limited to production processes under human oversight.

The stopwatch in the opening script is the smallest case of the pattern.

Every manufacturer has its own industrial intelligence. Some is documented, and some resides in the knowledge of experienced workers. Consider an experienced operator preparing to restart a line after a changeover. The machine settings matter, but so does knowing how a particular material behaves, what happened during the previous shift, and when an apparently minor variation deserves a closer look. A procedure can describe the sequence, but successful execution depends on the judgment that comes from experience and the organization’s accumulated know-how. That combination is industrial intelligence: a manufacturer’s accumulated know-how, operating practices, production data, and understanding of its own business. The opportunity now is to make more of that intelligence available wherever decisions are made, then connect it to the physical world of manufacturing. The strategic destination is not more agents, but rather to turn the existing business intelligence into an engine for profitable growth. Individual knowledge can be translated into organizational advantage Expertise has limited reach when it stays with one person, inside one system, or within one plant. Leaders now must determine how to extend that reach without stripping away the context that makes the knowledge useful. For example, an engineer evaluating a process change needs more than a production record. They need to understand the circumstances behind it: the equipment configuration, the quality requirements, and the trade-offs an operator made. Connecting those elements can help people apply experience beyond the place where it originated. When individual knowledge becomes widely accessible, innovation can spread to the maintenance workshop, the planning office, the engineering team, and the production floor. More people can contribute because more people can work with what the organization knows. But access to knowledge is only part of the opportunity. Manufacturing ultimately is physical, and it depends on materials, machines, and products. When intelligence participates in action Physical AI describes systems that can perceive conditions, reason about them, and take action within the physical world. By connecting machines and processes, manufacturers get a clearer view of what is happening on the factory floor. Agents can then reason on top of that information to identify patterns, predict problems, and determine the best course of action. Physical AI brings business intelligence into the physical process, allowing machines and systems to respond to changing conditions. Consider a robotic-handling operation encountering variation in the parts arriving at its station. Vision can help the system assess a part’s position. A model informed by the material and task can help determine an appropriate movement. Connected controls can then execute an adjustment within validated limits. The value is not merely seeing the variation; it is responding appropriately. The same principle applies to quality operations. When Toyota Industries connected previously fragmented paint shop data, engineering and operations teams were given the context needed to understand defect drivers. The company reports that analysis cycles fell from five days to under four hours, while its pilot reduced relevant paint defects by about 25%. The breakthrough was not simply identifying another signal; it was helping people understand why defects were occurring and respond before the outcome was locked in. This is the type of situation where industrial intelligence has become essential. Data may reveal what is happening, but industrial intelligence provides the context for what should happen, which constraints cannot be breached, and when uncertainty requires human judgment. Simulation offers a way to examine proposed actions before applying them to equipment. Manufacturers can explore how a change might affect a process in a digital environment. That does not remove the need for physical validation, but it creates room to learn without treating the production line as the first experiment. Responsible deployment, therefore, begins with operational discipline. Leaders need to establish clear authority or what systems may change, reliable connections to equipment, defined safety boundaries, and a record of decisions and actions. People must be able to recognize uncertainty, intervene, and bring equipment to a safe state. Trust rests on those practical arrangements, not on confidence in a model alone. More capable machines, more capable people Manufacturing has long combined human ingenuity with increasingly capable tools. Physical AI extends that relationship. Intelligent equipment can take on suitable repetitive or physically demanding tasks, while agent-assisted operations can help people navigate information-intensive decisions and changing conditions. The aim is not to make machines smarter; it is to see what people can build when knowledge can travel, intelligence can act, and ambition has more room to become reality. Rockwell Automation illustrates the point. At its Singapore plant, technicians use a maintenance copilot that combines manufacturing data, machine manuals, and knowledge captured from experienced workers. This system has reduced the time required for new workers to learn how to troubleshoot hundreds of machines from nine months to three. An engineer could explore more process alternatives before committing to a change. An operator could help improve the system’s response rather than repeatedly managing the same exception. This type of progress requires investment in equipment but also in people: training, time to learn, and playing a meaningful role in shaping how systems work. Machines become more capable so people can become more capable. Industrial Intelligence is the answer to sustainable growth Industrial intelligence connects what a manufacturer knows: the experience of its people, the signals from its operations, and the lessons accumulated over time. Physical AI gives that intelligence a way to act in the physical world, turning knowledge into decisions and decisions into action. The next challenge for industry is not simply to produce more, but to grow without adding complexity at the same rate. Intelligence can help manufacturers turn experience into scalable capability—improving productivity, accelerating innovation, expanding capacity, and enabling people to solve problems that were previously difficult to tackle. The next frontier of industry is not smarter machines or more automated factories; it is businesses that can continuously learn, adapt, and act. The opportunity is to turn the intelligence embedded in every operation into a compounding advantage for growth. Watch the video series to learn how industrial customers are