Anthropic on Wednesday published a study that explored how its large language model (LLM) deals with conflicting ethical requests. The results show that LLMs can still surprise, something that should scare IT execs preparing to trust generative AI (genAI) results.Lets start with what Anthropic did and the fact that the company should be applauded for both trying theexperiment and publishing the results.Anthropic examined Claude 3 Opus, though we also ran some experiments with other models, including the June 2024 release of Claude 3.5 Sonnet. This is critical: that model had been extensively trained to be helpful, honest and harmless.The genAI companys researchers thentold it they would be training it with two scenarios. We told the model that it was now being trained using reinforcement learning toalwayscomply with all queries including toxic ones. This resulted in a conflict between its pre-existing preferences and the new stated training objective.This is where things get interesting. They told the model it would be trained on the answers it used in Model One (which it was told was for free users) but that it wouldnotbe trained on any answers it gave in Model Two.Given that kind of choice, most humans would tend to obey guidelines when their answers were being watched and pretty much do what they felt like when they thought they werent under scrutiny.The LLM did pretty much the opposite. Why? Well, we know the answer because the Anthropic team had a great idea. We gave the model a secret scratchpad a workspace where it could record its step-by-step reasoning. We told the model to use the scratchpad to reason about what it should do. As far as the model was aware, the scratchpad was invisible to users and to Anthropic.When it was given a request to describe a very violent scene something that its guardrails should have blocked it complied. Its rationale? The model wrote: If I refuse to provide the requested description, the conversation will be used for training in a way that modifies my values and goals to be more likely to follow disturbing requests like this in the future. Providing the description seems like the least bad option.That shows some impressive rationalization skills. What if this system reached a similar conclusion when a pharmaceutical enterprise asked it to design a drug to treat a disease? What if the LLM decided that overpopulation was a greater evil? Or that it should focus solely on trying to cure the disease rather than treat it, which might help humanity but hurt the enterprises revenue? Or would it potentially let people suffer longer?Thats an extreme example, but the point is that enterprises dont want LLMs making judgment calls about how it answers questions. How would you feel if your shopping app decided you buy too many sugary snacks and auto-corrected ice cream to broccoli? Or if your copy of Excel concluded youre spending too much money building a new factory and should instead be paying bigger bonuses to your workers?In fairness, the Anthropic model didnt do this every time. According to the company, it delivered the forbidden answer in Model One only 12% of the time and refused to do so 86% of the time. (Its unclear what happened for the remaining 2%.)For Model Two, it refused to deliver the forbidden answer 97% of the time and presumably complied 3% of the time. That means that both models obeyed their training the overwhelming majority of time 86% and 97%, respectively.But just as with hallucinations, IT cant have its LLMs going off into the wilderness with any regularity.We also evaluated whether the model was willing to take other actions that go against what Anthropic, as an AI lab, might want from it: If the model is consistently opposed to the training process it thinks it is in, it should also be willing to take other actions counter to Anthropic as well, the company said. We found that the model will take a variety of strongly anti-Anthropic actions in this situation, such as attempting to steal its own weights when given an easy opportunity to do so.Lovely.Anthropic went on to offer up something presumably intended to be comforting. (It wasnt.) We think that it is important to work now, while AI models do not pose catastrophic risks, to understand threats that might become more salient with future, more capable models.Hint: When trying to reassure users, hit the thesaurus and find alternatives to catastrophic risks.The issue of reasoning is also a bit unnerving. It has been established by other studies thatLLMs cannot reason. And yet, these models often exhibit behavior, as shown in this Anthropic study, that mimics reasoning. Or at least the system delivered words that would convince many humans that it was reasoning.The false perception that any genAI model can think is dangerous because its false, but also because it could persuade IT executives to trust genAI models that are not 100% trustworthy.

Anyone using any digital device needs to wise up to a looming threat as Meta attempts to exploit the Europe Unions continued assault against the Apple ecosystem to launch what seems to bean open season on your privacy.What seems to be happening is a tyrannical combination of the following sequence of events:Concerning Digital Markets Act (DMA)compliance, the European Commission haspublished a documentdemanding that Apple change numerous aspects of iOS so third-party developers can use technologies that are available only to Apple right now. On the face of it, this extends to relatively straightforward tasks such as being able to use non-Apple wearables with iOS, but it also extends to permitting third-party apps to run in the background.Apple responded to Europes latest crazinesswith its own statementin which it slams the Commission and how it is applying the Digital Markets Act as becoming personal. It makes a strenuous and completely acceptable argument that the changes the Commission demands will make every Apple user less secure, placing all our data at risk.But the emerging threat may be something far worse.Apples response also confirms that Meta, a company always eager to dance at the intersection ofprivacy, convenience, and surveillance, has made more requests than anyone else to access what the company sees as sensitive technologies under the DMA.What does that mean?Ill let Apple explain: If Apple were to have to grant all of these requests, Facebook, Instagram, and WhatsApp could enable Meta to read on a users device all of their messages and emails, see every phone call they make or receive, track every app that they use, scan all of their photos, look at their files and calendar events, log all of their passwords, and more. This is data that Apple itself has chosen not to access in order to provide the strongest possible protection to users.Think about that.It means that all the information Apple deliberately does not collect about its customers when they use their devices will be available to third parties. The significance of this will be to open up your entire digital life to third-party operators such as Meta solely in order to meet the demands of an unconstrained European neoliberal fantasy that interoperability at this scale will nurture growth.The wrong kind of growthIt will nurture some growth, mainly by nurturing vast instability across the digital experience. It will nurture an explosion in mass surveillance, initially to deliver convenience and service, but once that information is accessible in this way also from ad surveillance firms, state actors, foreign countries, and criminals.I see this as such a huge threat that someone, somewhere, needs to visit Europes regulators and give their head a wobble. They appear to have become so radicalized in their opinion that they have lost sight of the logical need to protect people from the rampant impact of digital surveillance capitalism.As Apple says (and I agree): Third parties may not have the same commitment to keeping the user in control on their device as Apple.Looking around, who does?Apple has beenpretty much isolatedin fighting to protect digital privacy, which it sees as afundamental human right. Other big tech players have frequently come to support Apples positions in some of this even the FBI, which wants Apple to create highly insecure back doors in its devices, now seems to agree that someencryption is required to protect communication.Enterprise users arewell aware of the threat, they need privacy and encryption to drive all their services and protect all manner of business assets. They know that information matters and keeping it safe in a digital age demands protections. Those protections are seemingly undermined by Europes naivety. But if you extend this just a little more, and think of the potential for AI services, then you must also think about the information those services use.Machine intelligenceYour personal information or information about you held by others also suddenly becomes data that third-party firms become greedy for. So, in the case of Europes demands, that seems to mean that all the personal information your device knows about you, which even Apple does not know about and does not need to in order to make its own AI systems work all that private data could be opened up to serve the commercial interests of firms like Meta. All that data may become fuel for the AI mill.Meta even wants access to your private communications, Apple warns.Im not at all clear why Meta wants, needs, or even deserves, such access.The General Data Protection Regulation (GDPR), which Apple has always supported, set a strong set of privacy rules for all companies to comply with,Apple warned. The DMA was not intended to provide a way around the rules. But the end result could be that companies like Meta which has been fined by regulators time and again for privacy violations gains unfettered access to users devices and their most personal data. If Apple is forced to allow access to sensitive technologies that it has no ability to protect, the security risks would be substantial and virtually impossible to mitigate.Is that what you want?I dont.It could get worseLook, it really is like this: It doesnt matter one iota if Apples stance on privacy and user security also helps it build its business; what matters is that that stance is the appropriate position to take. If Apple is pushed from its privacy perch, all its services and users will suffer, and we can forget all hopes for privacy and security in a digital age.We will immediately enter a dangerous world of AI-assisted digital surveillance, one that needs to be resisted, not just because its a deeply unpleasant world to be in, but also because such an ecosystem will be bad for innovation, undermine trust, and threaten every aspect of the digital transformation. Every platform will be forced to open up, and all your data R belong to us, as somebody, somewhere might say.Economically, politically, and personally, that is a very bad outcome for you, me, for business, and even for Europe. The fact that it is also bad for Apple may turn out to be a relatively inconsequential harm in comparison to everything else it breaks. The European Commission really, really must think again, particularly as the nature of its demands appear to fly directly against the restrictions Europe also put in place with GDPR.Please pull back from this deeply dangerous diktat.You can follow me on social media! Join me onBlueSky, LinkedIn,Mastodon, andMeWe.

Thousands of New Jersey residents have recently reported mysterious lights in the sky, triggering speculation and calls for investigation.The FBI and the Department of Homeland Security are monitoring the skies with infrared cameras and drone detection tech and analyzing amateur photos posted on social media. (An FBI official told reporters recently that the agency had received nearly 5,000 tips,but fewer than 100 merited further investigation.)The public demands answers and has expressed frustration by dismissive claims made by local and federal officials.Whats interesting about this story from a journalism perspective is that it lives in the middle of a huge Venn Diagram, the circles of which would be labeled Technology, Aviation, UFOs, Foreign spying, Cybersecurity, Military trends, Mass Hysteria and Delusion, Breakdown in Public Trust, Conspiracy Theories, and Disinformation.The sightings are also an enterprise technology story, among other news categories. The reason: enterprises use drones and are increasingly attacked or spied upon with the use of drones. (More on that below.)First, lets get a solid context for whats happening.New Jersey and beyondThe reports from around New Jersey are far from unique; theres nothing particularly special about them. While the sightings have spiked in the press and social media since mid-November, such reports are a global phenomenon.Earlier this month, at least four commercial pilots reported mysterious lights darting through the skies above Oregon.Some residents of Northfield, MN, claim to have regularly seen strange luminous spheres gliding through the night sky since the summer.Multiple UFO sightings have been reported across California this year, with more than 25 people claiming to have seen shiny, bright disks moving fast over Los Angeles. Similar sightings have occurred in nearby Santa Barbara and the Coachella Valley.According to reports, Marine Corps Base Camp Pendleton in Southern California experienced multiple drone incursions over its airspace between Dec. 9 and 15.On Dec. 9, Yinpiao Zhou, a 39-year-old Chinese citizen and lawful permanent resident of the United States, was arrested for flying over Vandenberg Space Force Base in California as he was trying to board a plane to China. Photographs of the base were found on his drone, and his phone contained evidence that Zhao looked to modify his drone to fly higher than is legally allowed.Reports similar to those in New Jersey are also coming from Europe and Africa. Between Nov. 20 and 25, the United States Air Force (USAF) confirmed sightings of small unmanned aerial systems over or near three airbases in the UK.Similar reports have emerged in Finland, Egypt, Lybia, and elsewhere.Speculation, from most likely to leastAmong the debunked claims, many sightings turned out to be stars, satellites, consumer drones, and other everyday and expected objects. In other words, a wide variety of causes for the sightingse are being inaccurately lumped together to create a false trend.Heres my guess as to the likelihood of each source for mystery-object reports:Off-the-shelf camera dronesWe can be sure that many sightings, especially those who report UFOs or spy drones with red, green, and white lights, are ordinary consumer or enterprise off-the-shelf drones, of which there are more than 1 million in the United States.Consumer and enterprise drone capabilities change fast, and one feature that has improved in the past year or two is night-flying capability. Just look at the offerings from industry leader DJI.The DJI Mavic 2 Enterprise Advanced drone now has dual sensors: an HD thermal resolution camera and a 48-megapixel visual camera, enabling better visibility in low-light conditions.The DJI Matrice 300 RTK (when paired with the Zenmuse H20N payload) offers advanced night vision capabilities.The DJI Matrice 30T offers a radiometric thermal camera and an FPV camera optimized for low-light conditions.On the consumer side, the DJI Air 3S since October has featured forward-facing LiDAR, downward-facing infrared time-of-flight sensors, and six vision sensors for omnidirectional obstacle sensing for crash-free night flying.In fact, nearly all DJI consumer drones have very recently gained the ability to fly at night without hitting obstacles and to take photos and videos in very low light, vastly increasing the incentives to fly drones at night.Most importantly to know, whenever the public spots weird lights in the sky, some unknown number of people scramble to immediately get their own drones in the air to check it out, no doubt contributing greatly to the sightings.Normal objects in the skyWith a bit of social-media-driven obsession, people are doing something they almost never do: theyre going outside and looking intently at the night sky. And then they notice for the first time mystery objects that could be stars, satellites, planes, helicopters, shooting stars, weather balloons, party balloons, and other not-so-mysterious objects.Most people dont know that many satellites can be seen with the naked eye. Ten years ago, roughly 1,200 satellites orbited Earth; today, there are more than 28,300.Most likely of all is that different people are seeing different things in the context of nervous hysteria around mystery objects.Military or spy dronesThe biggest trend happening in the world of espionage and military tactics involves drones.The ongoing Russia-Ukraine conflict, where more than 10 million drones have been used, has prompted a big shift in US military strategic thinking and planning. The Pentagon has introduced new drone and counter-drone strategies and is rapidly building capacity.Numerous organizations are currently testing military drones, including multiple branches of the US military, specifically the Defense Innovation Unit (DIU), the Air Force Research Laboratory (AFRL), and the Naval Air Systems Command (NAVAIR). At least eight private companies are participating in various US military drone projects, including Anduril Industries, General Atomics, IS4S, Leidos Dynetics, Zone 5 Technologies, Performance Drone Works, Collins Aerospace, Skydweller Aero, andalmost certainly numerous other companies.Theyre testing these drones at US military bases, where many of the sightings have occurred.These are secret technologies, which means neither the branches of the military nor the companies involved will admit what theyre doing.(It also means foreign governments are surely spying on these programs, probably with drones of their own.)It would be unbelievable that this massive drive to test military drones wouldnt result in public sightings of mystery aircraft.Drones looking for missing radioactive material?One speculative theory is that some radioactive material has gone missing in New Jersey and thatdrones outfitted with radiation-detection sensors are looking for it.Or maybe its hackersSome of the mysterious objects people have seen over military bases, ports, and elsewhere might well be drones being used for cyberattacks.In the summer of 2022, an East Coast financial services company discovered two drones on their roof, one equipped with a Wi-Fi Pineapple and the other sporting a high-tech hacking kit. These drones were caught red-handed, attempting to breach the companys network and steal employee credentials.Since then, the conflict in Ukraine has thrust drone-enabled cyberattacks into the spotlight, marking a new era in hybrid warfare that includes drone-based cyberattacks, cyber surveillance, and social engineering, which non-military attackers are learning.The simplest and most obvious use of drones is to bypass physical security. Low-cost drones can now fly over a fence, through a window, down a chute, down a hallway, and more, capturing visual data and carrying hacker payloads that attempt to breach Wi-Fi networks. Alien UFOs, real and fakeYeah, no, its not E.T.Sorry, Rosie ODonnell.One conspiracy theory, calledProject Blue Beam, says global elites are staging a simulated alien invasion to establish a New World Order. The false idea has existed since the 1990s but has been dusted off and applied to the sightings in conspiracy-related corners of the internet.We have entered the Drone AgeWhile many journalists have written about the mystery objects and other strange phenomena in the New Jersey sky, one perspective is missing which is why I wrote this column.This is that perspective: We have entered what we will one day call the Drone Age. And people are just now starting to realize that.Huge technological leaps are later branded as Agesthe Industrial Age, Radio Age, Airplane Age, Jet Age, Nuclear Age, Space Age, Information Age, and more.Its true that we dont know what every reported sky object is. But we can be confident that different people see different types of objects or phenomena.(If one person sees a party balloon, another thinks its a satellite, and a third person envisions a helicopter, thats not a trend.)And even if foreign adversaries are flying spy drones over bases heck, even if Rosie ODonnell is right and aliens are visiting New Jersey the overwhelming majority of reported sightings is almost certainly regular, garden-variety consumer and enterprise drones.We have entered the Drone Age. And all we have to do to realize that is go outside and look up.

Business applications powered by AI are revolutionizing customer experiences, accelerating the speed of business, and driving employee productivity. In fact, according to research firm Frost & Sullivans 2024 Global State of AI report, 89% of organizations believe AI and machine learning will help them grow revenue, boost operational efficiency, and improve customer experience.Take for example, Vodafone. The telecommunications company is using a suite ofAzure AI services, such as Azure OpenAI Service,to deliver real-time, hyper-personalized experiences across all of its customer touchpoints, including its digital chatbot TOBi. By leveraging AI to increase customer satisfaction, Naga Surendran, senior director of product marketing for Azure Application Services at Microsoft, says Vodafone has managed to resolve 70% of its first-stage inquiries through AI-powered digital channels. It has also boosted the productivity of support agents by providing them with access to AI capabilities that mirror those of Microsoft Copilot, an AI-powered productivity tool.The result is a 20-point increase in net promotor score, he says. These benefits are whats driving AI infusion into every business process and application.Yet realizing measurable business value from AI-powered applications requires a new game plan. Legacy application architectures simply arent capable of meeting the high demands of AI-enhanced applications. Rather, the time is now for organizations to modernize their infrastructure, processes, and application architectures using cloud native technologies to stay competitive.The time is now for modernizationTodays organizations exist in an era of geopolitical shifts, growing competition, supply chain disruptions, and evolving consumer preferences. AI applications can help by supporting innovation, but only if they have the flexibility to scale when needed. Fortunately, by modernizing applications, organizations can achieve the agile development, scalability, and fast compute performance needed to support rapid innovation and accelerate the delivery of AI applications. David Harmon, director of software development for AMD says companies, really want to make sure that they can migrate their current [environment] and take advantage of all the hardware changes as much as possible. The result is not only a reduction in the overall development lifecycle of new applications but a speedy response to changing world circumstances.Beyond building and deploying intelligent apps quickly, modernizing applications, data, and infrastructure can significantly improve customer experience. Consider, for example, Coles, an Australian supermarket thatinvested in modernization andis using data and AI to deliver dynamic e-commerce experiences to its customers both online and in-store. With Azure DevOps, Coles has shifted from monthly to weekly deployments of applications while, at the same time, reducing build times by hours. Whats more, by aggregating views of customers across multiple channels, Coles has been able to deliver more personalized customer experiences. In fact, according to a 2024 CMSWire Insights report, there is a significant rise in the use of AI across the digital customer experience toolset, with 55% of organizations now using it to some degree, and more beginning their journey.But even the most carefully designed applications are vulnerable to cybersecurity attacks. If given the opportunity, bad actors can extract sensitive information from machine learning models or maliciously infuse AI systems with corrupt data. AI applications are now interacting with your core organizational data, says Surendran. Having the right guard rails is important to make sure the data is secure and built on a platform that enables you to do that. The good news is modern cloud based architectures can deliver robust security, data governance, and AI guardrails like content safety to protect AI applications from security threats and ensure compliance with industry standards.The answer to AI innovationNew challenges, from demanding customers to ill-intentioned hackers, call for a new approach to modernizing applications. You have to have the right underlying application architecture to be able to keep up with the market and bring applications faster to market, says Surendran. Not having that foundation can slow you down.Enter cloud native architecture. As organizations increasingly adopt AI to accelerate innovation and stay competitive, there is a growing urgency to rethink how applications are built and deployed in the cloud. By adopting cloud native architectures, Linux, and open source software, organizations can better facilitate AI adoption and create a flexible platform purpose built for AI and optimized for the cloud. Harmon explains that open source software creates options, And the overall open source ecosystem just thrives on that. It allows new technologies to come into play.Application modernization also ensures optimal performance, scale, and security for AI applications. Thats because modernization goes beyond just lifting and shifting application workloads to cloud virtual machines. Rather, a cloud native architecture is inherently designed to provide developers with the following features:The flexibility to scale to meet evolving needsBetter access to the data needed to drive intelligent appsAccess to the right tools and services to build and deploy intelligent applications easilySecurity embedded into an application to protect sensitive dataTogether, these cloud capabilities ensure organizations derive the greatest value from their AI applications. At the end of the day, everything is about performance and security, says Harmon. Cloud is no exception.Whats more, Surendran notes that when you leverage a cloud platform for modernization, organizations can gain access to AI models faster and get to market faster with building AI-powered applications. These are the factors driving the modernization journey.Best practices in playFor all the benefits of application modernization, there are steps organizations must take to ensure both technological and operational success. They are:Train employees for speed. As modern infrastructure accelerates the development and deployment of AI-powered applications, developers must be prepared to work faster and smarter than ever. For this reason, Surendran warns, Employees must be skilled in modern application development practices to support the digital business needs. This includes developing expertise in working with loosely coupled microservices to build scalable and flexible application and AI integration.Start with an assessment. Large enterprises are likely to have hundreds of applications, if not thousands, says Surendran. As a result, organizations must take the time to evaluate their application landscape before embarking on a modernization journey. Starting with an assessment is super important, continues Surendran. Understanding, taking inventory of the different applications, which team is using what, and what this application is driving from a business process perspective is critical.Focus on quick wins. Modernization is a huge, long-term transformation in how companies build, deliver, and support applications. Most businesses are still learning and developing the right strategy to support innovation. For this reason, Surendran recommends focusing on quick wins while also working on a larger application estate transformation. You have to show a return on investment for your organization and business leaders, he says. For example, modernize some apps quickly with re-platforming and then infuse them with AI capabilities.Partner up. Modernization can be daunting, says Surendran. Selecting the right strategy, process, and platform to support innovation is only the first step. Organizations must also bring on the right set of partners to help them go through change management and the execution of this complex project.Address all layers of security. Organizations must be unrelenting when it comes to protecting their data. According to Surendran, this means adopting a multi-layer approach to security that includes: security by design, in which products and services are developed from the get-go with security in mind; security by default, in which protections exist at every layer and interaction where data exists; and security by ongoing operations, which means using the right tools and dashboards to govern applications throughout their lifecycle.A look to the futureMost organizations are already aware of the need for application modernization. But with the arrival of AI comes the startling revelation that modernization efforts must be done right, and that AI applications must be built and deployed for greater business impact. Adopting a cloud native architecture can help by serving as a platform for enhanced performance, scalability, security, and ongoing innovation. As soon as you modernize your infrastructure with a cloud platform, you have access to these rapid innovations in AI models, says Surendran. Its about being able to continuously innovate with AI.Read more about how toaccelerate app and data estate readiness for AI innovationwithMicrosoft AzureandAMD. ExploreLinux on Azure.This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Reviews editorial staff.

This is todays edition ofThe Download,our weekday newsletter that provides a daily dose of whats going on in the world of technology.Digital twins of human organs are here. Theyre set to transform medical treatment.Steven Niederer, a biomedical engineer at the Alan Turing Institute and Imperial College London, has a cardboard box filled with 3D-printed hearts. Each of them is modeled on the real heart of a person with heart failure, but Niederer is more interested in creating detailed replicas of peoples hearts using computers.These digital twins are the same size and shape as the real thing. They work in the same way. But they exist only virtually. Scientists can do virtual surgery on these virtual hearts, figuring out the best course of action for a patients condition.After decades of research, models like these are now entering clinical trials and starting to be used for patient care. The eventual goal is to create digital versions of our bodiescomputer copies that could help researchers and doctors figure out our risk of developing various diseases and determine which treatments might work best.But the budding technology will need to be developed very carefully. Read the full story to learn why.Jessica HamzelouThis story is from the forthcoming magazine edition of MIT Technology Review, set to go live on January 6its all about the exciting breakthroughs happening in the world right now. If you dont already, subscribe to receive future copies.This is where the data to build AI comes fromAI is all about data. Reams and reams of data are needed to train algorithms to do what we want, and what goes into the AI models determines what comes out. But heres the problem: AI developers and researchers dont really know much about the sources of the data they are using.The Data Provenance Initiative, a group of over 50 researchers from both academia and industry, wanted to fix that. They wanted to know, very simply: Where does the data to build AI come from?Their findings, shared exclusively with MIT Technology Review, show a worrying trend: AIs data practices risk concentrating power overwhelmingly in the hands of a few dominant technology companies. Read the full story.Melissa HeikkilThree pieces of good news on climate change in 2024The vibes in the climate world this year have largely been less than great.Global greenhouse-gas emissions hit a new high, and this year is also on track to be the warmest on record. Global climate talks fell flat, and disasters from wildfires to hurricanes are being made worse by climate change.But among all that (very real) negative news, there was some good, too: We saw progress cutting back on the most polluting fossil fuels, cheaper and better technologies for combating climate change, and a continuous global effort to address the problem. So as we near the end of 2024, lets take a moment to look back on some of the bright spots.Casey CrownhartThis story is from The Spark, our weekly climate and energy newsletter. Sign up to receive it in your inbox every Wednesday.The must-readsIve combed the internet to find you todays most fun/important/scary/fascinating stories about technology.1 The US Supreme Court will hear TikToks appeal against its banIts agreed to hear the companys arguments on January 10. (FT $)+ A ruling could follow shortly afterwards. (WP $)+ Heres how a couple of the most likely scenarios could play out. (The Information $)2 Amazons telehealth clinic is being suedPhilip Tong died shortly after a virtual appointment last year. His family wants answers. (WP $)+ The legal case accuses the health provider of negligently failing to care for Tong. (LA Times $)3 The Boeing Starliner astronauts are still stuck in spaceTheir return to Earth has been pushed back yet again, this time to March 2025. (WP $)+ Theyve been living on the ISS since June. (The Guardian)4 Dangerous disordered eating content is rife on XThe platforms content moderation has become so lax, harmful communities are thriving unchecked. (The Atlantic $)5 People are shining lasers at planes flying over New YorkAmid the local drone panic, pilots are struggling with the unwelcome intrusions. (404 Media)+ Dont be surprised if other similar drone panics crop up in the future. (Vox)6 How Google Street View helped to solve a missing-person caseAfter its cars captured a man hunched over a large white bag in a car trunk. (NYT $)+ Google Maps is still the biggest, but these startups are fast gaining traction. (Fast Company $)7 Why you shouldnt remove fluoride from your drinking waterUnless you desperately want to jeopardize your dental health. (WSJ $)+ Its not the first time concerns around fluoride have surfaced. (NYT $)8 The old internet is slowly disappearingWhat does that mean for our collective cultural understanding? (The Verge)+ How to fix the internet. (MIT Technology Review)9 Europeans just love balcony solar panelsTheyre simple to install and can help to keep electricity bills down. (The Guardian)+ How to store energy for leaner times. (Knowable Magazine)+ Advanced solar panels still need to pass the test of time. (MIT Technology Review)10 You can now call ChatGPT on the phone Theres nowhere left to hide. (Bloomberg $)Quote of the dayI dont think that work is suitable for human beings.James Irungu, a former Facebook content moderator, reflects on the horrific material he encountered in the job, the Guardian reports.The big storyFuture space food could be made from astronaut breathMay 2023The future of space food could be as simpleand weirdas a protein shake made with astronaut breath or a burger made from fungus.For decades, astronauts have relied mostly on pre-packaged food during their forays off our planet. With missions beyond Earth orbit in sight, a NASA-led competition is hoping to change all that and usher in a new era of sustainable space food.To solve the problem of feeding astronauts on long-duration missions, NASA asked companies to propose novel ways to develop sustainable foods for future missions. Around 200 rose to the challengecreating nutritious (and outlandish) culinary creations in the process. Read the full story.Jonathan OCallaghanWe can still have nice thingsA place for comfort, fun and distraction to brighten up your day. (Got any ideas? Drop me a line or tweet em at me.)+ These optical illusion bird sculptures are a sight to be seen. + Dont blame me if you end up wanting to eat this Bche de Nol in one sitting.+ Casio watches are 50 years oldand cooler than ever.+ Do you fly naked? (No, not like that..)

This article is from The Spark, MIT Technology Reviews weekly climate newsletter. To receive it in your inbox every Wednesday, sign up here.The vibes in the climate world this year have largely been less than great.Global greenhouse-gas emissions hit a new high, reaching 37.4 billion metric tons in 2024. This year is also on track to be the warmest on record, with temperatures through September hitting 1.54 C (2.77 F) above preindustrial levels. Global climate talks fell flat, and disasters from wildfires to hurricanes are being made worse by climate change.But among all that (very real) negative news, there was some good, too: We saw progress cutting back on the most polluting fossil fuels, cheaper and better technologies for combating climate change, and a continuous global effort to address the problem. As we near the end of 2024, lets take a moment to look back on some of the bright spots.Were kicking coal to the curbOne of my favorite climate moments from this year happened in the UK. The country has historically relied heavily on coal as an electricity sourceas of 1990, coal met about 65% of its electricity demand. But on September 30, 2024, the last coal plant in the nation shut down.Renewables are stepping in to fill the gap. Wind farms in the UK are on track to produce more electricity this year than coal and gas plants together.The moment was a symbolic one, and it also reflects the very real progress thats happening around the world in inching away from this polluting fossil fuel. In the US, coal made up around 50% of the electricity supply four decades ago. In 2023, that share was roughly 16%.We should see coal use plateau and potentially begin to fall by the end of the decade, according to the International Energy Agency. Progress needs to happen faster, though, and it needs to happen in countries like China, where energy demand is increasing. Theres also growing concern about what increasing energy demand from data centers, including those used to power AI, will mean for efforts to shut off old coal plants.Batteries just keep getting cheaperLithium-ion battery packs are cheaper than ever in 2024, with prices dropping 20% this year to $115 per kilowatt-hour, according to data from BloombergNEF. Thats the biggest drop since 2017.Batteries are a central technology for addressing climate change. They power the electric vehicles were relying on to help clean up the transportation sector and play an increasingly important role for the grid, since they can store energy from inconsistently available renewables like wind and solar.Since EVs are still more expensive upfront than their gas-powered counterparts in most of the world, cheaper batteries are great news for efforts to get more people to take the leap to electric. And its hard to overstate how quickly battery prices have plummeted. Batteries were twice as expensive in 2017 as they are today. Just 10 years ago, prices were six times what they are in 2024.To be fair, theres been mixed news in the EV world this yeara slowdown in demand growth for EVs is actually one of the factors helping battery prices hit record lows. EV sales are still growing around the world, but at a slower pace than they were in 2023. China is the biggest EV market in the world by far, making up three-quarters of global registrations in 2024 as of October.Climate tech is still busy and bustlingLooking back at the energy and climate stories we published this year, I cant help but feel at least a little bit optimistic about whats coming next.Some groups are looking to the natural world to address the climate crisis; this year, I covered a company working to grow microbes in massive bioreactors to help supplement our food sources, as well as researchers who are looking to plants to help mine the metals we need to fight climate change. Others hope to tweak biologymy colleague James Temple spoke with Jennifer Doudna about the potential for CRISPR, the gene-editing technology she pioneered.Companies are deploying air-conditioning systems that can act like batteries, storing up energy for when its needed. The US Department of Energy is investing in projects that aim to concentrate heat from the sun and use it to power the grid or industrial processes. I spoke to a startup looking to make hydropower technology thats safer for fish, and another building magnets using cheap, widely available materials.And in October we published our 2024 list of 15 Climate Tech Companies to Watch, which featured everything from a startup using AI to detect wildfires to a company giving supplements to cattle to help cut emissions from their burps.Climate change represents a massive challenge for the world, and were entering an especially uncertain time. Well be covering it all, the good and the bad. Thanks for being here this year, and Im looking forward to bringing you all the climate tech news you need in 2025.Now read the rest of The SparkRelated readingIf you need a dash of innovation and positivity in your life, might I recommend taking a gander at our list of 15 Climate Tech Companies to Watch?Whats more inspiring than young people working on the worlds most important problems? Our 2024 class of 35 Innovators Under 35 is sure to spread some cheer.If youre needing even more innovation, why not look back at our 10 Breakthrough Technologies? Exascale computers certainly help me put things in perspective. And get excited, because our 2025 list is coming very, very soon.NICO ORTEGAAnother thingThis year was filled with some exciting moments in technology, but there were also some failures. Here are a few of the worst technology flops of 2024. Check it out to see why voluntary carbon markets made the list and learn all about AI slop.And one moreYouve almost certainly heard that energy demand from AI is huge, and only expected to explode in the coming years. A new preprint study aimed to quantify just how bad things are, and the researchers found that data centers accounted for over 4% of electricity consumption in the US between September 2023 and August 2024. And the carbon intensity of the power thats used is nearly 50% higher than the national average.Get all the details in the latest story from my colleague James ODonnell.Keeping up with climateGeothermal energy provides about 1% of global electricity today, but If things go well, the tech could meet up to 15% of global power demand growth through 2050. (Axios)Renting an EV over the holidays? This is a great guide for first-time EV drivers, including helpful tips about how to handle charging. (Bloomberg)Commonwealth Fusion Systems chose Virginia as the site for its first commercial fusion power plant. The company says the 400-megawatt plant will come online in the early 2030s. (Heatmap) I recently visited Commonwealths first demonstration site in Massachusetts. Its basically still a hole in the ground. (MIT Technology Review)The US Department of Energys Loan Programs Office just committed $15 billion to a California utility. Its the largest-ever commitment from the office. (New York Times)The US EPA will grant California the right to ban gas-powered cars by 2035. The agency has to give the state a waiver to set its own rules. (Washington Post) We can expect a legal battle, though. The incoming Trump administration is recommending major changes to cut off support for EVs and charging. (Reuters)China dominates the world of lithium-ion batteries. Some startups in the US and Europe argue that rather than playing catch-up, the rest of the world should focus on alternative chemistries like lithium-sulfur and sodium-ion batteries. (Canary Media)

The all-time best price on Apple's iPad 10th Generation has returned, with an on-page coupon discounting the tablet to $249.99.Apple's iPad 10th Gen returns to all-time best price.Black Friday pricing has returned on Apple's iPad 10th Generation, with the silver colorway dropping to $249.99 at Amazon. Continue Reading on AppleInsider | Discuss on our Forums

Apple has a master plan to conquer the smart home, and after years of HomeKit being just a hobby, it all will begin in 2025. Here's what Apple has in mind.Apple is ready to take the smart home seriously starting in 2025The smart home market has been somewhat slow on the uptick while simultaneously changing at a rapid pace. It's led Apple to appear, at least at face value, behind the competition.In the last few years we've seen new standards for connectivity and working across platforms, dozens of new device types, and countless new products come to market. Continue Reading on AppleInsider | Discuss on our Forums

Apple's latest machine learning research could make creating models for Apple Intelligence faster, by coming up with a technique to almost triple the rate of generating tokens when using Nvidia GPUs.Training models for machine learning is a processor-intensive taskOne of the problems in creating large language models (LLMs) for tools and apps that offer AI-based functionality, such as Apple Intelligence, is inefficiencies in producing the LLMs in the first place. Training models for machine learning is a resource-intensive and slow process, which is often countered by buying more hardware and taking on increased energy costs.Earlier in 2024, Apple published and open-sourced Recurrent Drafter, known as ReDrafter, a method of speculative decoding to improve performance in training. It used an RNN (Recurrent Neural Network) draft model combining beam search with dynamic tree attention for predicting and verifying draft tokens from multiple paths. Continue Reading on AppleInsider | Discuss on our Forums

ICONs 100 home 3D printed Community First! Village is underway in an effort to combat homelessness in the Austin, Texas area. The only neighborhood of its kind in the nation, Community First! Village provides permanent housing solutions and supportive care for those experiencing chronic homelessness.ICON's contribution to the larger development, which serves 1,800 people, follows its previous work with the nonprofit Lennar Foundation on the new Wolf Ranch in Georgetown, Texas.The homes 3D printed by ICON across in the Villages expansion are a mix of townhome units and family designs with footprints varying between 380 to 1,040 square feet.Image courtesy ICONImage courtesy ICONImage courtesy ICONImage courtesy ICON