moment

The war between Anthropic and the Pentagon points to something terrifying: a new “Oppenheimer Moment”

by

Anthropic has refused to bow to pressure from the Pentagon. Its co-founder and CEO, Dario Amodei, has just published a statement in which they make it clear that they are not willing to break their ethical principles. No massive espionage with AI, no development of lethal autonomous weapons with its models. And that reminds us of a terrible case: the one with the atomic bomb. From hero to villain. J. Robert Oppenheimer went from being the “father of the atomic bomb” and a national hero to become in an outcast. His sin was not betrayal, but his moral clarity. After witness the horror of Hiroshima and Nagasaki, Oppenheimer desperately tried to stop the atomic escalation and the development of the hydrogen bomb. Either you are with us, or against us. The United States, which had praised him in the past, took advantage of his former political affiliations and stripped him of all his privileges and influence. This demonstrated how the US government simply decided that scientific knowledge was state property and that any researcher who tried to propose ethical limits to their own projects would be treated as an enemy of the country. History is threatening to repeat itself these days. From Oppenheimer to Anthropic. He is doing it with a protagonist that is still there—the US Government—and another that is changing: the one who now defends the ethics of a scientific-technological project is not Oppenheimer, but Dario Amodei, CEO of Anthropic. Claude is increasingly vital in the US Government. Your company is between a rock and a hard place these days. Anthropic managed to make its model Claude become the pretty girl of the US Government. The ability of this AI has proven to be so remarkable that it was apparently used to plan the arrest of the former president of VenezuelaNicolás Maduro. red lines. But so that the Pentagon could use Claude, Anthropic imposed certain red lines. No use for mass surveillance of US citizens, and no use for the development of lethal autonomous weapons. And the Pentagon has ended up not liking those red lines, so they want to eliminate them and use Claude as they please as long as, they say, the Constitution and American laws are respected. The Pentagon wants AI without restrictions. That has ended up causing an enormously tense situation these days. The Pentagon threatened to punish Anthropic if it did not give in to its demands, and those threats from the Department of Defense have not been subtle at all. In fact, they have suggested that they could label Anthropic as a company that is “a supply chain risk,” a black label typically reserved for companies in rival countries like China or Russia. Contradiction. Dario Amodei himself explained in an entry on the company’s official blog that those two threats were self-exclusive: “These last two threats are inherently contradictory: one labels us as a security risk; the other labels Claude as essential to national security.” Can AI be nationalized? It’s a disturbing irony: the same government that considers Claude an essential tool for national security is willing to label his creators a public threat if they don’t hand over the keys to the kingdom and their AI. What the Department of Defense and the Pentagon want is to basically “nationalize” the AI ​​technology developed by Anthropic and appropriate it as they already did with the technology that gave rise to the atomic bomb. We know how that ended. Anthropic refuses to give in. The danger is enormous in both sections: mass surveillance, rather than defending democracy, can dynamite it from within, and the NSA scandal is a good example. But even more worrying is the Pentagon’s intention to use this AI to develop lethal autonomous weapons. Amodei insisted on this point, indicating that “The foundational models of AI They’re just not reliable enough. to power fully autonomous weapons. “We will not knowingly provide a product that puts American warfighters and civilians at risk.” Amodei even offers the Department of War/Defense help in the “transition to another provider” of AI models, but at the moment it is not clear which path the US government will take. Oppenheimer Moment. If the Pentagon finally execute his threat and ban Anthropic, the message for the industry will be chilling. In the age of AI there are no conscientious objectors: if a company develops a technological and strategic advantage at a military level, that company is at the mercy of the State. It is a new and terrifying “Oppenheimer Moment” that conditions the future not only of Anthropic, but of the development of AI models itself. In Xataka | “The world is in danger”: Anthropic’s security manager leaves the company to write poetry

An AI publishes 11,000 podcasts a day by copying local journalists. And at the moment there is no way to stop the avalanche

by

An automated podcast network publishes more episodes in 24 hours than many broadcasters do in a year, using AI to convert news articles to audio in minutes. A specific case, that of the channel ‘The Daily News Now!’, helps us to consider how far the scraping of content in the era of generative AI. To loot. The case was put on the table indicator: On January 31, at 2:57 in the afternoon, the newspaper ‘The Chronicle’ (a completely marginal publication: despite being 120 years old, it is published by Duke University, in Durham, and is run and produced entirely by students) published an article about Gemma Tutton, a student and pole vaulter who had won a university competition. Seventeen minutes later, a podcast called ‘Durham News Today’ uploaded an episode titled ‘Gemma Tutton’s Triumphant Return to Pole Vault’ to Spotify. The podcast, of course, had no connection with the newspaper. But it reproduced almost all the data from the original article in the same order, including practically identical phrases. And it is not an isolated case: ‘Durham News Today’ is one of at least 433 programs that make up ‘The Daily News Now!’ podcast network created by Corey Cambridge. As of January 23, ‘DNN’ has published more than 350,000 episodes (approximately 11,000 per day). How they do it. Obviously, with AI: a system of scraping (software automation that extracts large volumes of content) monitors media websites, excises text from published articles, processes it using natural language synthesis tools, converts it into audio and distributes it on platforms such as Spotify. All in a matter of minutes. And they don’t bother to dissemble: according to Indicator, they reproduce the structure, data and writing of pieces published by outlets such as local Fox and NBC affiliates, ‘TechCrunch’, ‘Toronto Star’, ‘The Verge’ or the radio station ‘WRAL’. The tools. To understand why an operation of this type is technically possible today, we must take a look at the ecosystem of tools that has been democratizing synthetic audio production for two years. In September 2024, Google activated the feature globally Audio Overview of NotebookLM. The tool converts any document uploaded by the user into an audio summary. The impact was immediate: NotebookLM went from 652,000 monthly visits in August of that year to 10.5 million in September, an increase of 371% in thirty days. In the three months following the global launch, users accumulated audio with a total duration greater than 350 years of continuous reproduction. NotebookLM normalized the idea of ​​the synthetic podcast, and it was all downhill from there. ElevenLabsspecialized in speech synthesis and valued at more than a billion dollars, launched its GenFM function in December 2024, which allows you to generate complete episodes from text. Wondercraftfunded in part by ElevenLabs, introduced support for editing podcasts generated with NotebookLM. Podcastle, aimed at podcast creators, incorporated speech generation with text to complete or replace fragments of speech. The secret: the price. In an analysis from a similar network (Inception Point AIwhich generates around 3,000 episodes per week with more than fifty AI announcers) producing an episode costs approximately one dollar, and with just 20 listeners the episode is profitable thanks to programmatic advertising. The model does not seek loyal audiences, but search engine positioning: by publishing hyper-specific episodes on cities or niche topics minutes after local media launch their articles, these networks anticipate humans’ capacity for informative immediacy. In other words: ‘The Daily News Now!’ appears in the top Spotify results for local news searches in dozens of American cities. It directly competes (and in many cases surpasses) the media from which it steals content. Legal issues. Cambridge defends itself by saying that its network only accesses “publicly available information” and merely summarizes it. But Indicator found almost thirty episodes of ‘Durham News Today’ that reproduced the structure, order and specific sentences of articles from ‘The Duke Chronicle’: it is not a specific pattern. And Cambridge may still be legally protected, but the problem is more about information ethics than legal details. In any case, in May 2025, the United States Copyright Office came to the conclusion that “publicly accessible” material is not necessarily free to use. There are legal precedents in that direction: in November 2025, a federal judge from New York did not reject the lawsuit by fourteen major publishers (including Forbes, The Atlantic and the Los Angeles Times) against the AI ​​company Cohere, considering that their summaries could constitute direct infringement if they reproduced “structure, sequencing, tone and expressive choices” of the original articles. On the contrary, in April of the same year, the case NYT vs. Microsoft dismissed claims related to the Copilot-generated summaries on the grounds that they were not “substantially similar” to the source articles. Meanwhile, and still without trialthere is the case of the New York Times against OpenAI and Microsoft, accused of using journalistic content to train their models Very clever. There is another detail: we are not talking about the ‘New York Times’, but rather ‘DNN’ concentrates its production on local niche news (university athletics, student councils, cats trapped in trees), first because these contents generate specific searches with little competition on Spotify. And second, because legally it is safer. They point to more fragile journalism models. Meanwhile, distributors like Spotify are developing tools to detect artificial music (removed more than 75 million tracks), but the next step is to make big brands aware that they do not benefit from the exploitation of newsrooms that cannot defend themselves. In Xataka | AI is already a battlefield: Anthropic has just accused DeepSeek and other Chinese companies of “distilling” Claude

Xiaomi already has its own AI model for robots. At the moment, he’s great at taking apart LEGOs and folding towels.

by

It has been a long, long time since Xiaomi stopped being a mobile company. Today the company’s tentacles reach all types of sectors, from mobile and household appliances until cars, chip design and, from now on, robotics. And the Chinese company has just presented its first vision, language and action model for robotics. Its name: Xiaomi-Robotics-0. What is this about?. Xiaomi-Robotics-0 an open-source model whose code can be found in GitHub and HugginFace. As the company explains, this model has been optimized to offer “high performance, speed and smoothness in real-time executions.” We should not think of this model as an AI capable of making a robot run and jump like a human, but rather one capable of making a “simple” robot understand its surroundings and know how to make the optimal decision without, for example, destroying whatever it has in its hands. About the robots. When we talk about AI applied to robotics we are not just talking about a robot being able to move. The device must know and understand that it should not apply the same force when holding a brick as it does when holding a cat, for example. In that sense, there has to be an understanding of the visual, an understanding of what is being seen and an appropriate execution of actions: this is a brick > it is a heavy object > I have to apply more force to hold it and move it from one side to the other. Xiaomi-Robotics-0 results in the benchmarks | Image: Xiaomi The benchmarks. Xiaomi has achieved, as detailed on the project website, very good results in the benchmarks I RELEASE (measures knowledge transfer), SimplerEnv (measures performance in real simulations) and CALVIN (measures performance in tasks conditioned by language). According to the company, Xiaomi-Robotics-0 “achieves high success rates and robust results in two challenging two-handed tasks: disassembling LEGOs and folding towels.” The fun of training. Every AI model draws from a training dataset. In the case of Xiaomi-Robotics-0, a 4.7 billion parameter model, the dataset consists of 200 million time steps of robot trajectories and more than 80 million samples of general vision-language data, including 338 hours of LEGO disassembly videos and 400 hours of towel folding videos. The results. The company claims in the paper that its model is capable of disassembling complex LEGOs of up to 20 pieces, adapting the grip in real time to avoid errors, using only one hand to place the towel correctly and folding it or, if you pick up two towels from the basket, take one of them, leave it in place and fold only one. This demonstrates an interesting capacity for adaptation and learning that, although it may seem trivial on paper, has its value if we think about industrial and even domestic robots. Beyond. What this model is demonstrating is being able to adapt to complex and unpredictable geometries, such as that of a towel thrown in a basket, and to understand the, let’s say, “soft physics.” On a towel it may seem like a small thing, but let’s think about manipulating human tissues in an intervention, for example. Same with LEGOs. It’s not just disassembling them, it’s understanding the position of the blocks, how they fit together, what force to apply and at what angle so as not to break them. Let’s think about a robot that removes debris. An industrial robot has historically been programmed with fixed coordinates, that is, moving something from point A to point B. A robot with AI like the one proposed by Xiaomi would be much more versatile. The first robot learns movements, the second robot learns tasks, and the difference is a world. If we think about a distant future in which there are domestic robots, a robot cleaning dust from a shelf will not be the same as knowing how to identify objects, decorations, etc., and understanding that it must move them to avoid throwing them away and cleaning them thoroughly. Cover image | Xiaomi In Xataka | A Chinese company boasts another limit in robotics: it ensures that its new humanoid robot runs like an elite athlete

OpenClaw is the most viral, fascinating and dangerous AI of the moment. For this last reason, it has joined forces with VirusTotal from Malaga

by

In 2025 we had a ‘DeepSeek moment’ and in 2026 we are having an ‘OpenClaw moment’. This AI agent is super powerful, but also super insecure. There is, however, good news, because the Malaga company VirusTotal has partnered with the OpenClaw project to try to mitigate one of the most important cybersecurity risks of this AI agent: its skills. what has happened. OpenClaw (formerly Moltbot, and before Clawdbot) has announced that it has begun a collaboration with the Malaga cybersecurity company VirusTotal, owned by Google. The agreement will see VirusTotal be in charge of “scanning” and analyzing the so-called “skills”, which work like OpenClaw plugins and add all kinds of functions. They do it, of course, but many take the opportunity to introduce malicious instructions that allow them to steal data and remotely operate other people’s AI agents. More security for disturbing AI. Peter Steinberger, creator of the project, has joined Jamieson O’Reilly, cybersecurity expert and founder of the company Dvulnand Bernardo Quintero, founder of VirusTotal, to offer that “additional layer of security for the OpenClaw community.” In it official announcement explain that “all the skills published in ClawdHub (the project’s official skills “store”) are now scanned through Virus Total’s Threat Intelligence system, including its new capability Code Insight (code inspection)”. Bernardo Quintero indicated on Twitter how the effort has already allowed 1,700 skillls to be identified as malicious. If the skill is malicious, it is blocked. This analysis carried out with the VirusTotal tools allows us to identify skills as malicious and block them immediately so that they cannot be downloaded. Not only that: those skills that have been classified as benign are analyzed again every day to detect scenarios in which for some reason they could end up becoming malicious. Still, be careful. Those responsible for OpenClaw warn: the VirusTotal scan helps a lot, but it is not a total guarantee that any skill can perform malicious actions on the machine on which we have our AI agent installed. The attacks of prompt injection Sophisticated skills can manage to cross that barrier, but of course this collaboration means that OpenClaw users can be much calmer regarding the skills available in the ClawdHub repository. OpenClaw wants to be much more secure. This first effort joins OpenClaw’s ambition to have a complete cybersecurity model which includes things like a public roadmap for your new developments in this area, a formal communication process, and details about full audits of your code. Plugging a problem that could kill OpenClaw. The OpenClaw project soon went viral due to its eye-catching options, but shortly after doing so a security audit initial 2,851 skills detected 341 malicious skills. Companies like BitDefender also joined these efforts to avoid problems with tools like AI Skills Checker to check whether a skill was dangerous or not. These malicious skills were, for example, capable of executing shell commands on the victim machine, which gave the attacker complete control of those resources. Attacking the machine is confusing it with natural language. Normally cybersecurity attacks are complex, but the problem with AI agents is that they work with natural language. This implies that to infiltrate these systems you do not have to use code, but simply “convince” and “trick” the AI ​​with natural language. That is where prompt injection attacks come in, which consist of giving instructions to those AI agents that can confuse them to obtain something that theoretically they should not allow them to obtain. Personal data, API keys of the models we use at OpenClaw, email accounts and passwords for all types of services… the possibilities are endless, and OpenClaw, which has access to all of this to operate autonomously, can end up being “tricked” into transferring said data. Beware of OpenClaw. These problems now seem a little less feasible thanks to the collaboration with VirusTotal, but those who are trying OpenClaw on their machines or any other platform should be very alert from the beginning. There are guides that help you install it with some barriers important security issues, and the project itself has a command (‘openclaw security audit –deep –fix’ to audit the most important problems and address them. In Xataka | OpenAI has a problem: Anthropic is succeeding right where the most money is at stake

There is a Chinese startup creating the most amazing robots of the moment. It’s called X Square

by

The only embodied AI (bodied artificial intelligence) company backed by the three Chinese technology giants: ByteDance, Meituan and Alibaba. Just over two years of life and financing rounds in which they have managed to overcome the 400 million dollars. These are some of the cover letters of X Square Robot, one of the most promising companies in the field of robotics. where does it come from. XSquare It is a Chinese startup which was born in 2023 at the hands of Wang Qian, an engineer and doctor from the University of Southern California who, in recent years, has maintained a discreet profile in the industry. The company was born not only as a company aimed at creating humanoid robots: they are also behind the development of the language models necessary to lead in robotics. The roadmap. The startup, despite its youth, has made the most of its two years of life. December 2023, full financing and start of operations. March 2024, efforts begin to develop a general large-scale model for embodied AIthe brain that would move its robots. May 2025, commercialization of Quanta X1, a bimanual wheeled robot equipped with its WALL-A model. Specially designed for logistics and commercial tasks. July 2025, first to show purposeful AI model general capable of directly controlling a highly dexterous robotic hand. Unlike traditional approaches—based on rules, fixed trajectories or action-specific training—the system uses a single model that integrates perception, planning and control, allowing grip and movement to be adapted in real time to changes in the environment. August 2025, Quanta X2 arrives, its first humanoid robot, also with a wheel base. The product. Quanta X2 is the latest solution from X Quare, a wheeled humanoid robot that integrates the company’s own AI model. This model allows the robot to have a vision system, autonomous motion control, real-time task planning, etc. We highly recommend watching the demo video in which X Square shows it in operation, because it is spectacular. Why is it important. X Square does not sell ordinary humanoid robots, it sells cognitive capacity. The norm in robotics companies is to design the hardware and adapt it to existing software. X Square designs its own models focused on physical AI. This is something fundamental for his native country, China. The country wants to accelerate the automobile industry in 2030 with 100% automated factories. The aid policy is especially favorable for local companies developing robotics solutions. China has created centers responsible for training robots to imitate human behavior. X Square software is key The backup. X Square is backed by giants like Alibaba and Bytedancethe first group having announced an internal team dedicated to robotics using Qwen, its AI models division, as a base. Despite Alibaba’s muscle when it comes to creating its own language models, the investment of more than $140 million in X Square Robot makes it clear that it is much more than a typical startup. Image | XSquare In Xataka | Robotics has just broken another scale barrier: there are already autonomous robots smaller than a grain of salt

At the moment it has an open field and four streets of a PAU

by

The official website says that “you no longer have to imagine it. Now you can live it.” But the truth is that, as the matter stands, it is difficult to imagine. Let’s hope we can live it. Because when there are nine months left, Madringthe capital’s Formula 1 street circuit looks like anything but a circuit. At these moments, it becomes difficult to think that through those streets and those half-assembly curves you can search Fernando Alonso his 33. If everything goes as planned, the September 11 The cars have to start rolling south of the city. Among the streets of Valdebebas, one of the Madrid PAUsand the bowels of IFEMA, which competes with the Fira de Barcelona for being the largest fairground and convention center in Spain. The layout is a good reflection of everything that Formula 1 is rewarding: urban layouts that allow attracting large investments from cities to export their image to the world even if, as in the case of Madrid, all the attractions around the circuit are residential buildings, a fairground and a half-built City of Justice for more than 20 years. Now, with 243 days left until Madrid returns to the Formula 1 calendar 45 years later, a doubt is beginning to float: whether Madrid will return to the Formula 1 calendar. Some works in diapers The FIA ​​said that everything is going well, that there is no problem. That’s what they collected in I amMotora portal specialized in motorsport competitions, a little less than a month ago. “There are no delays or concerns within the FIA,” they then stated from the media, which also claimed that a commission from the International Automobile Federation had been supervising the status of the works in the streets of Valdebebas. From IFEMA they also defend that they are advancing at the expected pace. “The works are going within the established deadline. The paving has begun and part of it is already finished, although it is being done little by little due to the rain,” they point out to Autobild. And they make it clear that the circuit still remains to be completed: “the last layer of asphalt is expected to be completed during the summer.” That IFEMA has come out to speak is no coincidence. The information points to a closing of ranks after the viability of the project was questioned. At least for this year. The Italian media Rmc Motori claimed last November that Liberty Media, owner of the rights to the sport, was considering removing 2026 calendar to Madrid in favor of Imola, given the progress of the works. The legendary Italian circuitwill not be part (at least for the moment) next year of the F1 calendar. What is certain is that the works seem to be in an embryonic phase. Those who are walking around the circuit these days are finding the streets of Valdebebas without any type of modification. Ready for you the cars pass at full speed but not single-seaters. Click on the image to go to the original tweet With few exceptionsit is difficult to intuit the circuit along the 22 curves that make it up. The Monumental, a banked curve with a 24% inclination that has become one of the great attractions of the circuit, it’s a muddy mess right now. Yes, progress has been made on the route but there is no sign of progress in the surrounding services and the first asphalt is conspicuous by its absence. The times are also much tighter than we might think. In August the circuit must be ready so that Eurocup-3, a single-seater category inferior to Formula 1, can compete in one of its grand prizes. If it arrives, the intention is to make it the first big test before Fernando Alonso, Carlos Sainz or Max Verstappen set foot on the soil of Valdebebas. Click on the image to go to the original tweet The circuit, in addition, has to fight with the opposition of the neighbors. Pave the way for cars to pass is causing profound changes in its streetsconstant works and the anticipation that the noise suffered during that weekend will be much higher than the averages that have paralyzed the concerts at the Santiago Bernabéu. Besides, environmental associations They defend that the project threatens the conservation of wetlands and “non-transplantable” trees in the area. Nor is Madrid the first city where the viability of a Formula 1 Grand Prix is ​​doubted a few months before its celebration. In South Korea, Yeongam circuit was not reviewed by the FIA up to 10 days before the traffic light went out on the finish line in a clear example of “out of sight, out of mind” heart. That same weekend work was being done on the track and in some areas the asphalt was not well established. In Las Vegas, Formula 1 has been fighting with a recurring problem for three years now: the sewers become loose with the passage of cars. And in Hanoi, 600 million euros were spent on a circuit so that five years later a total of zero cars raced before its abandonment. Photos | Ifema In Xataka | Madrid says that F1 will not be paid for with public money. Valencia promised the same and it cost them 300 million euros

The DGT is not going to fine for the V-16 beacons at the moment, and therein lies the key

by

Since last January 1, anyone who is stranded on the road due to a breakdown has to place the V-16 beacon connected. And what happens if I don’t have it? Absolutely nothing. At least that is what the Government assures. Because, with the law in hand, the agents can fine us if they consider it appropriate. We also don’t know how long this “truce” will last. “It is not tax collection”. This is what Fernando Grande-Marlaska stated in the press conference in which he gave the results of the road accidents relating to 2025. The DGT has made public the accident data for last year but a good part of the press conference has revolved around the topic of the moment: the connected V-16 beacon that the DGT has been required to carry since last January 1. The agents, Grande-Marlaska assures, will not fine for a “reasonable” period of time, in words reported by The World. They do it because, they say, “our objective is not sanctioning or collecting, what moves us is the obligation to save lives.” “Reasonable”. It is the temporary measure that the Minister of the Interior has used to refer to the time that the agents have before fining. The word says nothing because, really, from January 1, 2026, Traffic can fine us for not having the corresponding signage elements. The fine is 80 euros and it does not take into account whether we carry the triangles with us because the only essential element in the car when signaling an accident is the connected V16 beacon, which must be approved by the DGT. And the triangles have been left in a kind of limbo so that the driver can do with them whatever he considers. Not now. The position of the DGT has changed over time. Since it was confirmed that the V-16 beacon would be the only signaling element of a road breakdown, the discourse has changed and its position has been relaxed. At first it was argued that the use of triangles could be grounds for a fine since an incident was not being correctly signalled. Now, Interior says that there will be a period in which fines will not be imposed for this. Later it was left up in the air whether the beacon+triangle combination was valid. Finally, it will be allowed put the triangles “at your own risk”. many doubts. In his speech, Grande-Marlaska pointed out that last year more than 100 people died on the road, “a significant number for getting off to put up the triangles”, in words reported by Motorpassion. In The World They point out that estimates point to 25 pedestrians dying while trying to put the triangles in what Grande-Marlaska describes as “bleeding.” However, as we have said in Xatakathe DGT has never offered clarifying data. Traffic has always classified these victims as people run over “after getting out of the vehicle” but without clarifying under what circumstances. They do not indicate whether they were hit when getting out of the car, putting the triangles on, changing a tire on the shoulder or waiting for help to arrive. According to their accounts, between 2018 and 2022 (a period that includes before and during the COVID-19 pandemic), an annual average of between 18 and 26 people died in accidents “after getting off the vehicle” on high-capacity roads. as reflected in the document itself which explains why the regulations and technical requirements of this connected V-16 beacon are changed. Taking the total number of deaths in this entire series (8,615 people, according to data from Statista), we are talking about just over 1% of deaths that fall into the category “after getting out of the vehicle.” No fines but no extensions. The result in the application of the measure has been paradoxical. From the Interior they say that the measure is “essential” to reduce the number of road accidents but omitting its use or not having the beacon will not be penalized despite there being no extension. And, at the same time, Traffic defends that it has not implemented an extension because it is something that has been known since 2023 and that we should have already purchased the device. According to Pere Navarrodirector of the DGT, “we considered delaying it” but that “would not have changed anything.” Also left to the driver’s discretion whether or not they want to put the triangles in despite the fact that they consider it a sufficient risk to promote a regulatory change. And they recognize that something has been done wrong with the communication of the new measure. Photo | DGT and Help Flash In Xataka | The V-16 beacon business: who is making money with the elimination of the DGT triangles

At the moment brands are selling us “AI assistants”

by

The automobile industry is undergoing a technological transformation. And in addition to a transition to electric that is having a difficult time convincing drivers, there are vehicles that are increasingly more capable and dependent on manufacturers’ software. In this aspect, the functions of autonomous drivingwhich have been showing more legs lately. However, the gap between promise and reality is still considerable. More importantly, the level of full autonomous driving in commercial vehicles is still in its infancy. Level 3 remains a pipe dream for most. Level 3 autonomous driving, which allows the driver to let go of the steering wheel and take their eyes off the road in certain conditions, has been heralded for years as the next big leap. ford just promised that its L3 system will arrive in 2028 as part of its new electric vehicle platform, according to Doug Field, the company’s software manager. Technology is advancing in the automotive industry, but there is still much to do and each manufacturer is taking its own time. Mercedes-Benz, for example, has an operational L3 system since 2023, but it only works on mapped highways and at limited speeds. On the other hand, China has recently authorized to manufacturers such as Changan and BAIC to produce L3 vehicles, although restrictions on use remain strict. The reality is that these systems still require very specific conditions to function. Meanwhile, AI comes to the dashboard. Those hoping for greater capabilities in autonomous driving seem like they’re going to have to wait a while longer, at least until AI assistants become normalized in vehicles. In this regard, Ford has promised launch this year an AI-powered voice assistant that will first be available in its mobile application and then in the vehicle. The idea is that the mobile phone also accompanies the experience. In the official blog of the announcement they mention the example of being able to photograph any object to ask the AI ​​if it fits in the trunk or cabin of the pickup, since in this case the model would have the exact dimensions of the vehicle. The idea is that the system is designed to be compatible with different language models, including Gemini from Google, according to Ford. They are not the only manufacturers to be integrating AI into their vehicles, since chatbots like ChatGPT or Gemini are slowly reaching infotainment systems. Some examples of this are Mercedes-Benz, Opel, Volkswagen or Tesla, among others. Promises. Ford explains that it is developing many of these components internally to reduce costs and maintain control, although it is not creating its own language models or designing chips like Tesla or Rivian. The company assures having managed to reduce the costs of its hands-free driving systems by approximately 30% while increasing capabilities. The idea of ​​this approach is to launch more affordable electric vehicles after the lukewarm reception of some of its proposals such as the electric mustang and the pickup F-150 Lightning. Ford also acknowledges that it does not want to get into “a TOPS arms race,” the metric that measures the processing speed of AI chips, unlike Tesla which boasts about the raw power of its processors. China accelerates. The country represents a particular case. Daiwa Securities analysts cited by SCMP They estimate that almost 270,000 vehicles with L3 systems will be sold in 2026, approximately 1% of the total Chinese market. BYD and other local manufacturers They are doing massive testing in cities like Shenzhen, accumulating hundreds of thousands of kilometers of real data. On the other hand, a report from Southwest Securities points out that the legalization of L3 vehicles could generate demand for components and software valued at 1.2 trillion yuan by 2030. However, Chinese authorities have tightened supervision after fatal accidents related to driving assistance systems, such as the one that involved a Xiaomi SU7 in March 2024. There is still rope left for a while. While several brands are already announcing L3 systems for three or four years, what is coming now are incremental improvements, such as smarter voice assistants, improved hands-free on-road systems, or more polished user experiences. In this regard, Tesla continues to operate in China with its FSD (Full Self Driving) as a “hands on the wheel” technology, awaiting regulatory approval to operate hands-free. Most current systems are L2 or L2+, requiring constant driver attention. ford assures that his team, made up of former Argo AI engineers (its failed L4 autonomous driving project) and BlackBerry specialists, is prepared to deliver. Many manufacturers aim to have autonomous driving functions within a few years. We’ll see how things progress in the end. Cover image | Arcfox In Xataka | If it seems expensive to change the battery in an electric car, wait until you see what it costs in a Ferrari LaFerrari: more than 200,000 euros

Xiaomi has made profits selling cars in its first year. The problem is that it has optimized for an unrepeatable moment

by

Xiaomi Auto, Xiaomi’s car division, reported a few weeks ago something that is considered impossible in the automobile industry: achieving profits in its first year. It has had a healthy gross margin of 25.5% and a net profit of 680 million yuan, about 82 million euros, thanks to 109,000 cars delivered in a single quarter. Barely a year after selling its first car, the division presents numbers that place a newcomer in the same range as BMW or Mercedes. One that took Tesla years to reach and one that other manufacturers like NIO are still not there. Some They died trying to get there. Lei Jun has executed an impeccable launch and his investors have reason to be impressed, but if we take a closer look at the numbers and break down the origin of the margins (something that must be attributed to Poe Zhao’s wonderful analysis in Hello China Tech), a different story appears: that of a company that has perfectly optimized for a moment that will not be repeated. Two figures: The average price per car in the third quarter was 238,000 yuan (about 29,000 euros). The broadest category was close to 260,000 (about 32,000 euros). Those numbers They are not representative of the market that Xiaomi wants to addressbut rather they represent a temporary concentration. In that quarter, many units of the SU7 Ultra and other premium configurations. The first buyers (the biggest fans of the brand, those who wanted to be the first to drive a Xiaomi) ordered the most expensive versions. It’s not that Xiaomi has fooled anyone, it’s the natural dynamic of any technological launch. The early adopters They always buy the higher versions. The testmotto is to confuse that initial demand with sustained market demand. The 25.5% margin does not validate your business model, it only tells you that you have sold the right product to the right people at the right time. The question is what happens when those people run out. Lu Weibing, president of the group, made this clear in the presentation of results. It said auto margins will likely fall in 2026 due to “competitive factors and normalization of the product mix.” It’s careful business language, but lto translation is simple: When you’re done delivering premium configurations and have to sell entry-level versions to maintain volume, you’re going to find out how much it really costs to compete in this market. Apple experienced something similar with the first Apple Watch. The first few quarters showed spectacular margins, but those numbers reflected sales to enthusiasts willing to pay for novelty, not sustained demand from a mature category. They had to learn to sell beyond the circle of fans. The difference is that Apple was not competing in a market with structural overcapacity and price wars. Xiaomi yes. Xiaomi competes in a Chinese electric vehicle industry where overcapacity is systemicgovernment subsidies have an imminent expiration date and the competition is fierce. There is another detail that should worry: Xiaomi is delivering cars faster than it is selling them. They are consuming the backlog of accumulated orders at a rate that exceeds the entry of new orders. An optimized factory running at maximum capacity is impressive, but if demand is not growing at the same rate, you have built production capacity for a level of demand that you have not yet proven exists. What is coming in 2026 is a kind of convergence of pressures: The portfolio of premium configurations will be exhausted. Subsidies will disappear. And security regulations will be tightened. Xiaomi will have to demonstrate that it can be profitable by selling cheaper cars, without public aid and meeting stricter standards. It is the moment when companies that built a real business are separated from those that surfed favorable temporary conditions. The trap of early profitability is not that the numbers are false. It’s that they make you believe that you have solved the problem when you have only optimized for the easier phase. The real test of Xiaomi Auto is not whether it can make quality cars (it has already proven this) but whether it can build a car business that works when the novelty wears off and it has to compete car for car with rivals that cannot afford to lose. That answer is not in the third quarter report. It’s coming. In Xataka | Xiaomi is no longer a brand: there are several brands fighting over the same logo Featured image | Xiaomi

What is coliving and why has it become the residential alternative of the moment

by

Arriving in a new city with a suitcase is not always the beginning of an idyllic adventure; It is often the result of desperate mathematical calculation. With rental prices climbing 51.4% in the last decade while salaries barely moved 3.4%, according to the joint report by Fotocasa and InfoJobsfreedom of choice has been replaced by scarcity management. In this scenario, the coliving It is not born from a romantic desire to share a kitchen, but from a structural necessity. It is the housing response to a world where the traditional market has built walls of unpayable deposits and rigid contracts that no longer fit with anyone’s life. Opening the door to a coliving is, for many, the only way to stop being “expelled” from the system and become a resident with rights and services, even if it is at the cost of reducing one’s own square meters. What is coliving and how does it work? He coliving It’s not just sharing a flat; It is a professionalized evolution of coexistence. According to the specialized media Minutthis model is a hybrid that combines the privacy of a room with integrated services. The operational performance, as detailed by ULI experts (Urban Land Institute) and the report The European Coliving Best Practice Guide, It is based on a comprehensive management model. In other words, the resident pays a single bill that covers rent, furniture, high-speed Wi-Fi, cleaning and supplies. This ecosystem removes the “mental load” of home management. As the MIT thesis points outcoliving was born to provide resilience to an exhausted real estate market, offering “ready to move in” spaces that allow the tenant to focus on their career or personal projects from minute one. Types of coliving The versatility of the model has allowed different formats to be created according to the user’s needs: Urban and Flex Living Models: It is the commitment to density. According to Savillsthese formats will represent 16% of the new rental offer in Spain. They are large buildings with hundreds of units that revitalize the city center. Thematic Colivings: The MIT report highlights spaces where the community is filtered by interests: from “hubs” for artists selected for their work to communities of programmers. Rural Coliving: Maybe the guy more transformer. Cases can be highlighted such as Send either Anceu in Galicia, where coliving is used as a tool against depopulation, allowing digital nomads from Google or Spotify to live in villages of 20 people, injecting talent and consumption into rural areas. Collaborative housing (Senior Living): To combat the epidemic of loneliness in the elderly. The Law 3/2023 of the Valencian Community It is a pioneer in Spain by regulating these homes where mutual support is the central axis. What advantages does coliving have? The immediate advantage is affordability. From the Coliving.com portal estimates that a resident You can save up to 40% compared to a traditional studio. However, there are invisible benefits. A report from Lund University emphasizes that coliving is a sustainable urban housing strategy, reducing energy and water consumption by sharing resources and appliances. Furthermore, the psychological impact is measurable. While urban isolation grows, 71% of “colivers” affirm feel more connected. Given the return-to-office policies in cities with impossible prices, living in a coliving near the workplace is the only alternative to avoid two-hour daily commutes. How much does it cost to live in a coliving: prices In cities like New York or London, the savings are drastic, but in Spain the model is also consolidated. According to CBREinvestment in the sector Living room It reached 3,730 million euros in 2024, which allows us to offer high-quality accommodation at prices that, although they seem high at first glance, are competitive by eliminating investment in furniture, maintenance and supply charges. It is, in essence, the transformation of the rental into a transparent monthly subscription. In the main urban nodes of Spain, these are the current ranges: Madrid and Barcelona: Between €750 and €1,300 per month. The price varies depending on whether the room has a private bathroom or if the complex includes luxury services such as a gym, pool or rooftop. Málaga, Valencia and Alicante: Between €500 and €900 per month. These cities are attracting digital nomads with an offer that prioritizes community and proximity to the sea. Difference between coliving and shared apartment There is no need to confuse them. In a shared apartment, coexistence is random and management is informal. In coliving, there is professional management 24/7. As Minut highlightsthe use of technology (noise and smoke sensors that respect privacy) guarantees that coexistence is not degraded. Furthermore, the contracts are individual; If a partner does not pay, it does not affect the rest, something that the Urban Lease Law (LAU) does not always guarantee in traditional group contracts. How to find a coliving Finding these communities is now as easy as booking a hotel thanks to platforms like Coliving.com. However, unlike a hotel, the community factor is vital here. Many managers as mentioned in MIT studiesthey conduct prior interviews to ensure that the resident’s profile fits with the vibe of the building, seeking a harmony that benefits all members. The coliving business for investors For the investor, coliving is a safe haven asset. CBRE points out that Madrid and Barcelona concentrate the greatest interest due to their high profitability per square meter. However, the Uría Menéndez office warns about “limbo” legal: since there is no clear national law, coliving navigates between the Civil Code and municipal regulations that seek to organize the market. In this context, Madrid has taken a step forward with the RESIDE Plana new roadmap designed to combat “tourism” and the escalation of prices caused by vacation rentals. This regulation is key for the sector because it draws a red line: it separates buildings for residential use from tourist ones. However, the City Council will allow private public buildings that are obsolete or in disuse. they transform in colivings or affordable rental housing, as long as their restoration is … Read more

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.