takes

AI saves us time but takes away the story

by

A few days ago I surprised myself doing something that five years ago would have seemed sacrilege. I had in front of me one of those reports that you save to read on Sunday morning. 5,000 words, a prestigious signature and a great design. A text of those who ask for calm. And when I didn’t even have two sentences, I instinctively looked for the ‘summary’ button that now crown my browser. Nine lines. That was the whole summary. It was not for lack of interest, it was rather for that modern urgency that whispers to us that spending twenty minutes on a single idea is an inefficient thing. After a quarter of an hour I remembered almost nothing of those nine lines. I had the information, but I didn’t have the knowledge. We are turning reading into an administrative procedure. What started as a survival tool to deal with the deluge of work emails or some long-winded Reddit threads has colonized our capacity for wonder. In 2026, AI not only helps us write, it also is teaching us not to read. Or even worse: it is convincing us that the path is a hindrance to reaching the destination. It is the definitive victory of the TLDR about curiosity. The problem with outsourcing digestion is that we start from a false premise: that the substance of things is the only thing that matters. But in culture, information or in a simple conversation, substance is sometimes the least important thing. Ask an AI to summarize Don Quixote for you and it will tell you that it’s like a man from La Mancha who has read too much and confuses windmills with giants. You have the information, but you have not heard the conversations with Sancho on the roads. You have not felt the bitterness on the beach in Barcelona nor the lucidity of someone who regains their sanity to realize that the world, without its madness, is a gray place. Technology, in its efforts to eliminate friction (paradoxically, being the one who has blocked our notifications) is taking away the fabric of our experience. Silences and nuances are what fixes memory. The horny thing is What are we using that time for that we supposedly save by not reading?r. It is not to think deeply or to walk around without a cell phone and hit the coconut, but to consume even more summaries. It’s a loop infinity (pun intended) empty efficiency. We optimize the consumption of information to be able to ingest more information, which in turn we summarize in the next scroll. Thus we become archivists of a life that we did not get to witness. We save, synthesize and archive, but we do not inhabit anything. We are reaching a phase in which the true status, the intellectual luxury of our era, is not to be very smart or to be up to date with everything thanks to our AI agent, but in be able to sustain attention. Prestige belongs to those who can afford the extravagance of reading a text from beginning to end, of listening to a podcast without skipping the silences or set it to 1.75x. Or finishing watching a movie without having used your cell phone for two hours. Efficiency is a great metric for an assembly line or an AWS server, but If we let it guide the leisure of a human life, we are making ourselves a little miserable. We start by optimizing each minute to end up leaving everything in a list of three key points. Or in a nine-line summary. But life cannot be summarized. In Xataka | There is a generation working for free as a documentarian of their own life: they are not influencers but they act as if they were Featured image | Xataka

If your Madrid-Barcelona train now takes five hours, Renfe and Iryo have a message: there is no compensation

by

What used to be two and a half hours has become recurring trips of more than four hours. The Madrid-Barcelona line lives between speed limitations that are multiplying travel times. Adif is working hard to correct the alleged defects on the road. And, in the middle, some passengers who bought a ticket that promised a 150-minute trip that is now impossible to fulfill and who are not going to receive a single euro in compensation. The Madrid-Barcelona line. Since the occurrence of fateful train accident in Adamuz (Córdoba)the Madrid-Barcelona line is in focus. After months of complaints from train drivers, and self-imposed speed limitations, as they have confirmed to XatakaAdif strives to review all avenues to verify that there are no defects or, if there are, fix them. The company in charge of managing and maintaining the roads began applying temporary speed restrictions in the most controversial sections. These limitations they rose and applied punctually in the days following the Andalusian accident, with the vibrations being the focus of controversy. More, many more. As the days go by, the controversy has grown. To the reviews of Adif facilities a storm has joined which has complicated the service even more. The result: trains that were supposed to take 150 minutes between Madrid and Barcelona have been arriving regularly after four and a half hours. The latest news is that Adif has asked Renfe, Ouigo and Iryo to eliminate the last services of each day in order to be able to work for a greater number of hours and reduce the planned days of track inspection. These works will study the complaints reported by the train drivers and, if necessary, fix the damage on the tracks if the technicians consider that intervention is necessary, they point out in The World. No compensation. These warnings from the train drivers and the subsequent reviews by Adif are what are causing the continuous delays in the journeys. Delays that, however, will not be compensated by Renfe and Iryo, companies that already report on their websites that this situation does not fall within the reasons for returning a ticket partially or in its entirety. So much Renfe as Iryo They emphasize that these delays are unrelated to the service provided by the operators and, therefore, will not be reimbursed. In the case of the Italian company, this decision affects tickets purchased after January 28, while in the case of Renfe it will not be a reason for refunds for tickets purchased after January 31. An exception. What travelers from these companies do have the right to be relocated to another shift by the operator if their trains have been canceled as a result of Adif’s latest request. In the case of Renfe, which has issued its decision in statements that it has collected The Countrythe trains to relocate those affected will be of double composition to double the number of seats and the possibility of canceling the trip has been opened at no cost to the traveler. A new controversy. These delays and the decision not to compensate travelers, understanding that they are due to reasons unrelated to their services, deepen the controversies that have been surrounding railway compensation for months. And Renfe, by order of Congresshas the obligation to compensate again for delays of more than 15 minutes and return the money if the train arrives 30 minutes above the scheduled time. These deadlines were extended to 60 and 90 minutes respectively. in 2024 and They should have returned to their original deadlines in 2026 but this has not happened yet. Photo | trenduck and Renfe In Xataka | Spain has put so many passengers on the train that the Government is already toying with an idea: that we travel standing

The time it takes to get to a highway anywhere in Spain, on a revealing map

by

Faced with the pressing housing problem in Spain In large cities, one of the simplest solutions for those who can afford it is to leave stressed centers such as Madrid or Barcelona in search of more accessible municipalities and properties. How much? It depends on your budget, what your work is like and what the destination location offers you in such objective terms as services and infrastructure. And there is one essential to move: the distance to a main road. I speak with knowledge of the facts: this was a key factor when choosing a municipality to buy an apartment months ago. My new location has direct access to the highway and getting from there to my trusted padel club in Pamplona is 10 minutes longer than doing it from my old apartment, located in the center of the Navarrese capital. Although it is not ideal, my pocket has appreciated it and the sacrifice is profitable for me. Now, having chosen an idyllic municipality in the Navarrese Pyrenees would have been a very bad idea in terms of mobility (although bucolic on days like today). That was my personal decision, but given the prices, I know that I am not alone: ​​from buying in the capital to doing so in a municipality in the province there are price variations of up to 131% in Madrid or 126% in Álava, according to the latest Idealista study that collects La Razón. Because if the price of the property in Villagónadas de Abajo is the lowest in the province but it is where Cristo lost his lighter, already such. Well yes: the price differences are abysmal and the communications are too. An x-ray of territorial inequality and Spanish orography This map created by Digital Cartography With data from the Ministry for the Ecological Transition and the Demographic Challenge, this is evident. The cartography collects the minutes by car to a highway or highway from a good part of the Spanish state (if there are no this type of roads, as happens in Ceuta or Melilla for example, then they do not appear) with information from 2022. To see everything in a more intuitive way, they have used the colors of the traffic light, where green is what can cost you up to 20 minutes and red goes from one hour to 133 minutes in the maroon areas. The access time to a highway or highway in Spain. Digital cartography with data from the Ministry If we superimposed a physical map with a demographic one we would find a clear diagnosis of red zones in critical areas such as the Asturian massif and the Pyrenees, the muga with Portugal (especially in Zamora, Salamanca and western Extremadura), the Iberian System and the maximum expression of “Empty Spain” in the south of Teruel, the north of the basin and areas of Guadalajara or the Betic Systems. We know that in communications Spain It is a centralized state with Madrid as the nerve center and the lines of these main roads, although they do not appear on the map, can be intuited. Without going any further, it is not too difficult to imagine where the A-2 goes to Barcelona or the A-6 to A Coruña. That is the first clue as to why we find such an uneven map: the radial network model, which leaves enormous gaps in peripheral areas that are not linked to large state/European corridors. Obviously the extreme orography of the Pyrenees or the Iberian System makes construction difficult on a technical and economic level (it is not that it is not possible to lay out viaducts or tunnels, it is that it makes the cost skyrocket), but the Average Daily Intensity mandates: for a public work to be approved there is a cost-benefit analysis and if an area has a low population density, the ADI is low, making it difficult to justify the investment. On the other hand, there are environmental restrictions: some of these red zones coincide with national parks or protected areas. In this scenario, obtain a Environmental Impact Statement (mandatory in projects of this magnitude) is an impossible mission. The small print. Something that I greatly appreciated when I returned to Navarra is that there is no traffic… compared to Madrid. The rush hour for leaving work or school may be noticeable in a few minutes of delay, but it is light years away from the traffic jams that I have had to suffer in return or bridge operations when I lived in the state capital. Because although in Madrid almost everything is green, in practice those minutes correspond to a distance traveled respecting the limits of the road and assuming fluid traffic. In Xataka | This is the DGT map to visualize where there are active V-16 beacons in Spain. There is another more useful unofficial map In Xataka | Europe’s passenger car industry, in a revealing map that makes it clear who is the real “engine” of the EU Cover | Digital cartography

China has a solution that only takes 0.1 seconds

by

In the age of electricity, time is no longer measured in minutes, but in milliseconds. As Fatih Birol explainsexecutive director of the International Energy Agency (IEA), the world has fully entered a new era where electricity consumption grows twice as fast as general energy demand. However, this advance has an “Achilles heel”: network stability. Renewable energy, although necessary, is inherently unstable. The great fear of operators is that a small failure in a network saturated with solar and wind energy will cause a domino effect that ends in a total collapse. Faced with this scenario, China has deployed technology capable of detecting, isolating and recovering the network from a failure in just 0.1 seconds. From hours of darkness to the blink of an eye. Historically, managing power grid failures was a slow and manual process. As a South China Morning Post report recallsrestoring power after a community blackout typically required between 6 and 10 hours of work. That year, China already marked a milestone by testing an artificial intelligence system that reduced that time to 3 seconds. However, that is no longer enough. The recent achievement of 100 milliseconds (0.1 seconds) It is the result of a collaboration of more than a decade between elite universities (Tianjin and Shandong), the state corporation State Grid and automation specialists such as Beijing Sifang Automation. This technology is not only faster, but is capable of identifying “micro-currents” of just 100 milliamps, almost invisible faults that previously went unnoticed until it was too late. “Self-healing” versus intermittency. The fundamental problem, detailed in the study published in Energy Informaticsis that modern networks are much more complex. The massive incorporation of renewables and extreme weather conditions make traditional diagnostic methods based on static rules fail due to lack of precision and adaptability. This advancement means that the Chinese power grid —the largest in the worldwith projected consumption in 2025 of more than 10 billion kWh—is moving from a reactive system to a “self-healing” one (self-healing). This capability is so strategic that China already has exported the technology to 12 nationsconsolidating its influence not only as a panel manufacturer, but as the architect of global electrical safety. The algorithms behind the miracle. To understand how this speed is achieved, we must look at academic study by Qi Guo and his team. The system is supported by a dual structure of intelligent algorithms: Fault Location Algorithm (FLA): Uses a fault classifier Support Vector Machine (SVM) with a radial basis function (RBF) kernel. This “brain” analyzes variables such as voltage, line impedance and weather conditions to predict with 92% accuracy where exactly the problem has occurred. Fault Isolation Algorithm (FIA): Once the critical point is located, a decision tree logic comes into play that evaluates the severity. According to the research, if the fault is critical (such as a short circuit near a substation), the system orders the immediate isolation of that section and redirects the energy along alternative routes almost instantly. This hybrid approach allows the system to learn from historical data and adapt to dynamic conditions, something that conventional distance protection systems simply cannot do as effectively. The new geopolitical battlefield. The energy transition not only redefines how energy is produced, but also who controls the rules of the new industrial system. While the West focuses its response on ensuring the supply of critical minerals through initiatives such as ReSourceEU, China advances in less visible terrain but more decisive: the standardization, digitalization and integration of the infrastructures that will sustain the low-carbon economy. Rather than competing for resources, the dispute revolves around who designs the technological architecture on which global growth will function. The ability to recover a network in 0.1 seconds is not just a technical record; It is the life insurance of a highly electrified economy. The greatest current risk is that this clash of strategies between powers ends slowing decarbonization. However, in the race for stability, China has shown that while the rest of the world continues to search for the switch in the dark, they have already designed a system that never lets the light go out. Image | Unsplash Xataka | China dominates the world of renewable energy, but it has an Achilles heel: it depends on the West more than it admits

Neuroscience explains why the brain takes much longer to mature than we thought

by

The idea we have about adolescence right now it ends at 25 years old, this being the age at which supposedly the brain has just been ‘cooked’ forever to give way to a functional adult. But the reality is very different as the new studies point out, since we would continue to mature the brain until at least 32 years old. Where did the current idea come from? To understand why scientists pointed to 25 years as the age at which brain maturity ends, we must go back to studies of the past. Specifically to Resonance studies from the 90s and early 2000s like the classic Nitin Gogtay who mapped brain development and discovered that the cortex matures from “back to front.” This means that the sensory and motor areas are consolidated soon, but the prefrontal cortex which is in charge of executive functions, impulse control and planning is last in line. The problem is that many of those studies stopped following the subjects when you reach 20 or 21 years oldsince seeing that the curve continued to rise, it was assumed that the “peak” of maturity would arrive shortly after, around the mid-20s. But we had no idea what happened after this. Just assumptions. A new frontier. In order to solve this ‘blindness’ of neuroscience used the analysis of more than 4,000 brains using connectivity neuroimaging techniques at the University of Cambridge. What they saw was clearly five ‘epochs’ or milestones in brain wiring throughout life. Different turning points. And as if our life were a game, in the brain we have like five different screens that begin at a specific age that acts as a turning point. These ages are: 9, 32, 66 and 83 years. What interests us in this case is the period between 9 and 32 years, since the brain is characterized by a continuous increase in the efficiency and integration of neural networks. It is what the authors describe as an ‘extended adolescence’. It’s not that at 30 you think the same as a 15-year-old, but that the architecture of connections has not yet reached its final ‘adult’ form. Something that occurs at age 32 and remains stable until age 66, when brain activity begins to decline. To understand it better. Researchers wanted to use a simile to illustrate this new paradigm. To do this, they ask us to think of our brain as the union of several “functional neighborhoods” that specialize in specific tasks such as vision, language or logic. All of these are integrated with each other through different highways that are high-speed connections. Well then, between 20 and 32 years old The brain is balancing these two processes, so that the connections between different areas of the brain are well connected and organized. And it is precisely this typical pattern of the adult network, where the brain is capable of integrating complex information fluidly, which does not appear until after the age of thirty. Teenager at 30? This is where the important nuance comes in. Just because the brain continues to mature structurally does not mean that we should redefine adolescence in legal or clinical terms. All this because maturation is a gradient, not a switch of ‘now I’m a teenager and now I’m not’. To understand this, you have to know that the different elements of the brain and executive functions have a very different development curve. In this way, saying that the brain matures at 32 is a simplification that is as useful (or as erroneous) as saying that it matures at 25. What science really tells us is that there is no sudden development “blackout”; We remain biologically plastic and dynamic much longer than we thought. An opportunity for habits. This prolonged maturation is good news for all of us, since if the brain continues to actively ‘wire’ itself throughout our 20s, it means that structural plasticity is especially dynamic at this stage. In this way, science is quite clear: aerobic exercise, learning new languages ​​or facing cognitively demanding tasks during this “third decade” of life helps to improve the volume and organization of the brain’s white matter. On the contrary, factors such as chronic stress can affect the integrity of those connections. In short, a brain at 28 years old is not a finished product, but rather a work in progress that is finishing paving its best highways. The next time someone tells you that you should have your life figured out now because “you’re an adult,” you can tell them that, according to the University of Cambridge, your brain still has a couple of years of baking left. Images | Hal Gatewood Robina Weermeijer In Xataka | From 27 to 36 years old the brain reaches its peak concentration. And from there, bad news

In the midst of the RAM memory crisis, Samsung takes a leap with its HBM4 memory. It does not imply good news for the pocket

by

We are in full RAM price crisis. The industry is a cake that three large producers share and the data centers and the artificial intelligence They want to eat the whole cake. Samsung is one of the companies that manufactures memory for consumption and data centersand will soon begin mass production of its latest broadband memory chips: the HBM4. Don’t throw the bells in the air too soon. HBM4. This technology represents a crucial advance in stacked memories. Its density allows double the bandwidth, key to transmitting more data per second, but they are also up to 40% more energy efficient than HBM3. In short: they consume less energy and have fewer bottlenecks, which translates into an improvement in data processing. Industry sources point out that Samsung will use the 10-nanometer D1c manufacturing process for the matrix of these HBM4 memorieswith an internal structure of 4 nm. It’s a more advanced process than the 12-nanometer D1b from its main rival, SK Hynix. In addition, it will achieve a data transfer speed of 11.7 Gbps compared to 9-10 Gbps of the current standard. Hello Nvidia. South Korean media they point that these new Samsung HBM4 modules they would have passed Nvidia certification testing and will be in february when the company starts mass manufacturing them. Where will they end up? Some to Nvidia’s new AI acceleration system, called Vera Rubinothers at the heart of Google’s seventh-generation TPUs. After these reports, the company’s shares they went up 5.3% in the Seoul market. The enemy at home. In statements To South Korean media, Samsung representatives have commented that they feel quite confident with a new product that will clear up doubts about the company’s ability to supply the demanding needs of data centers. The fifth-generation HBM3E memories were a bottleneck for the company, so major players in the AI ​​industry looked next door: SK Hynix. Also South Korean, she is the second leg of memory chip manufacturing. The third is the American Micron Technology, a considerable distance from the two South Koreans. A year ago we already told that SK Hynix had achieved enormous efficiency in the DRAM stacking process to create these HBM memories, which allowed it to be 8.8 times more efficient than Samsung or Micron and, therefore, produce more modules for an industry that never stops asking. Meanwhile, the two South Koreans were in a race for the development of the new generation HBM4, and Samsung seems to have struck the first blow. Of course, it is estimated that Hynix will also begin mass production of these new memories on the same dates. And the consumer… what? Well nothing. If you were expecting good news related to the price of RAM, it must be said that no improvements are expected. These HBM4 modules will go to Nvidia, but we recently commented that OpenAI had reached an agreement with Samsung and SK Hynix to supply with 900,000 wafers per month. It is the volume equivalent to 39% of the estimated global capacity… and only for one company. Translation? Bottleneck in the market, a manufacturing speed that may not meet that demand and more bad news for the user. We have seen that Micron has abandoned its Crucial brand for consumers in favor of RAM for data centers, and that Samsung and SK Hynix are focused on HBM4 memories en masse, although they are not used in consumer devices, implies that this is where they will focus on this lucrative AI market. In short: Samsung may be dominating the new generation of memories, but 2026 seems difficult for anyone who wants to build a PC, expand RAM of yours, buy a new mobile or even wait for good news from the Steam Machine. Image | TSMC, Google In Xataka | RAM has become so, so expensive that there are manufacturers selling computers in an unprecedented way: “pre-assembled”

There was a time when the Lottery Jackpot “took you away from work.” Today it barely takes away a part of the mortgage

by

Someone who already has gray hair still remembers that, thirty years ago, May you get the Christmas Fat Man It was practically the key to financial freedom. With the full prize of one tenth (about 30 million pesetas in the 90s) you could buy several houses, pay mortgages and even ensure the well-being of your family with that stroke of luck. Today, with a prize of 400,000 euros (328,000 euros after taxes), that story sounds very different. One of the main conditions is that, in the mid-nineties, the real estate market in Spain I played in another league. Buy an apartment…or several In cities like Madrid, a home of about 90 square meters could be found for less than 14 or 15 million pesetas, according to official statistics. That meant that Fatty Christmas allowed to buy two apartments medium-sized in a big city, or buy one, pay off the mortgage and a good pinch to maintain a good margin of liquidity. In those years, the award was not just help: it was a complete break from financial worries. As was often heard at the doors of lottery administrations while the winners uncorked bottles of champagne, it was a prize that “kept you off work.” Thirty years later, the prize is still striking in terms of numbers, but its real purchasing power has changed. El Gordo has been frozen at 400,000 euros per tenth for more than a decade, while the price of housing has followed an almost constant upward trajectory. In Madrid, the average house price It ranges between 5,500 and 5,758 euros per square meter, which implies that with the 328,000 euros net of the prize, you can barely purchase 60 or 70 square meters at an average price. In practice, this means that Gordo no longer even guarantees a standard floor in many neighborhoods of the capital. Barcelona offers a similar image. With average prices located at 3,084 euros per square meter, the Gordo de Navidad allows buy a modest home or a medium-sized apartment in peripheral areas, but it is far from the purchasing capacity it had in the nineties. The comparison leaves no room for doubt: where before the prize opened the door to buying an apartment in the city and a house on the beachtoday it is barely enough for one, and not necessarily in the best conditions. The contrast is softened slightly if the market is viewed from more affordable cities. In capitals like Zamora or Lugo, where average prices are between 980 and 1,300 euros per square meter, El Gordo continues to allow you to buy spacious homes or even more than a small property. However, even in these more affordable markets, the premium no longer equivalent to that massive asset leap that it represented three decades ago. The difference is not so much in the amount of the prize as in the uneven evolution of prices. This purchasing capacity is also explained by the general price context. He housing cost It was much more aligned with the average income of the population and access to property was not subject to the housing and demand pressure that characterizes the current market. El Gordo, in that scenario, functioned as a real wealth multiplier. A Gordo with more salary, but less power make a salary comparison helps to better understand this change of scale. In the 90s, the average annual salary in Spain was around 2 million pesetas (about 12,000 euros). In that context, the Gordo of 30 million pesetas was equivalent to approximately 15 times the annual salary of a worker medium, which reinforced its perception as an immediate economic transformation: decades of income concentrated in a single stroke of chance. Today, according to the latest data from the National Statistics Institute, the median salary annually in Spain is around 23,300 euros. With this reference, the current Gordo’s 328,000 euros is equivalent to just over 14 times the median annual salary. The proportion, curiously, is not that different from that of the nineties. The big difference appears when that salary multiple faces the price of housing (and all goods in general), which has grown much faster than income. That’s the key to change. Although the premium maintains a similar relationship with salaries, your ability to buy a home has deteriorated drastically. The real estate market has become decoupled from wage growth, and El Gordo, by remaining fixed, has been trapped in the middle of that gap. What was previously enough to buy two apartments today barely covers one, and in many cases forces them to continue getting into debt, although to a lesser extent. The social meaning of Gordo has changed. In the nineties it was synonymous with total economic independence. In 2025, it is still an extraordinary stroke of luck, but its role has shifted, no longer guaranteeing financial freedom, but financial relief. In Xataka | There is something even more difficult than winning the Lottery Jackpot: not making mistakes with the Treasury when collecting it Image | Flickr (srgpicker)

this time it takes aim directly at its ‘reputation abuse policy’

by

What Brussels has launched today is not just another technical note on how Google works, but a movement that points directly to the way in which it decides what we see when we search for information. An internal policy designed to combat spam has ended up at the center of a new European file because, according to the Commissionyou could be relegating content from legitimate media and publishers. An issue that, for the Commission, deserves to be thoroughly reviewed to determine if its application is having undesired effects. We are facing an action that opens an official procedure in which the Commission will evaluate whether Google is complying with the obligations of the DMA in relation to the treatment that publishers receive in its search engine. Brussels wants to check whether the access and positioning conditions comply with the equity criteria provided for services designated as gatekeeper. This initial phase does not involve attributing a violation, but it does activate a detailed process that will determine how the regulations are actually being applied. Politics under suspicion. Google includes ‘reputation abuse policy’ in its Search spam rules and presents it as a tool to address practices aimed at manipulating ranking in results when sites include content from commercial partners. From a technical perspective, the motivation makes sense: the ecosystem is full of practices that try to exploit gaps to obtain a better position in the results. The Commission’s question is whether this policy is affecting publications that use commercial collaborations within a legitimate editorial framework. For media outlets, these deals are an important source of revenue, and their demotion in Google Search can have real effects on audiences. Brussels wants to know to what extent its application may be penalizing actors who are not seeking to manipulate anything. The analysis will revolve around that fine line. DMA in action. The Digital Markets Law establishes its own regime for the platforms considered gatekeepersthe large platforms that act as a gateway between companies and users in the digital environment. These services are required to ensure that their internal rules are understandable, justified and reviewable by the Commission, even before there is demonstrated harm. The investigation is framed in that model: Brussels wants to validate that the policy applied by Google complies with the reinforced obligations that accompany that status. Blow to the revenue model. The executive vice president for a Clean, Fair and Competitive Transition, Teresa Ribera, was explicit about the point that most worries Brussels: “We are concerned that Google’s policies do not allow news publishers to be treated in a fair, reasonable and non-discriminatory manner in search results.” He also stressed that the loss of income comes “at a difficult time for the sector,” which makes this investigation more than just a technical review. A long and complicated relationship. Brussels has maintained constant scrutiny over Google for years, visible in sanctions like the 2,950 million euros imposed in September 2025 for practices in its advertising business, or in the 2017 Google Shopping fine, ratified by the European Court of Justice in 2024. This new investigation does not start from scratch: it adds to a history that reflects how the Commission has tightened its surveillance as the company’s activity has covered more sectors. Pressure from Washington. The case also comes at a time when some of the loudest criticism of the European digital framework comes from the United States. Donald Trump has denounced that measures like the DMA hurt American companies and has warned of possible additional tariffs if they persist. Without being part of the file, these statements illustrate the political context in which Google’s policy is examined and show how European regulation coexists with growing commercial sensitivity on the other side of the Atlantic. The possible outcomes. From now on, Brussels will examine documentation, ask Alphabet for clarification and evaluate whether the policy fits into the DMA’s obligations. If it detects non-compliance, it will inform the company of its conclusions and the measures it considers necessary to correct them. The procedure can be closed without sanctions, with internal adjustments by Google or with the imposition of formal obligations and, ultimately, fines. The Commission plans to complete the analysis within a period of up to twelve months. Images | sarah b | 1981 Digital In Xataka | Apple accepts crumbs in China: the 15% that shows who has the power

While OpenAI takes all the media glory with ChatGPT, Alibaba is already taking important clients with Qwen. The latest: Airbnb

by

Alibaba has been investing in its family of open language models for quite some time.qwen‘, which are gaining increasing acceptance between developers and users. Although OpenAI takes all the media glory with ChatGPT and the rest of the services, the Chinese firm is not short and already is overtaking him with some clients. The latest example: Airbnb, which has chosen to rely mostly on Alibaba’s Qwen AI model for its automated customer service, leaving ChatGPT in a secondary role. Airbnb’s decision. Brian Chesky, co-founder and CEO of the tourist accommodation platform, explained Bloomberg this week that his company “heavily relies” on Alibaba’s Qwen model. As he admitted to the outlet, ChatGPT’s integration capabilities “are not quite ready” for Airbnb’s needs. On the other hand, Chesky assured that Qwen is “very good, fast and cheap.” It is curious, especially considering that Chesky is a personal friend of Sam Altman, head of OpenAI. How the system works. Airbnb’s customer service agent, which the company deployed to all its users Americans in English last May, is built on 13 different AI models, including those from OpenAI, Google and open source providers. However, Chesky recognized that, although they use the latest OpenAI models, “we usually don’t use them much in production because there are faster and cheaper models.” Just like point the company, the system has allowed them to cut their human workforce by 15% and claims to have saved average resolution time, going from almost three hours to just six seconds. Open source is gaining ground. Open source models, which developers can modify as they wish, are increasingly challenging closed systems like those from OpenAI. Although the company also has an open model (gpt-oss), Chinese tech companies are releasing models much faster, more cost-effectively, and open source. Joe Tsai, president of Alibaba, declared recently that the winner in AI should be determined by “who can adopt it the fastest,” not “who creates the most powerful model.” A future integration with ChatGPT in the air. Although Airbnb is awaiting the development of ChatGPT app integrations and could consider a collaboration in the future, similar to those of its competitors Booking and Expedia, the platform is not currently among the first applications available on the OpenAI chatbot. Chesky even advised to OpenAI about its new ability for third-party developers to integrate their applications into ChatGPT, a feature that the company announced this month and which he described as a “developer preview.” And now what. Airbnb plans expand its AI agent with support in Spanish and French this fall, and 56 more languages ​​next year. Meanwhile, the company claims to be betting on new social functions to foster connections between users and improve travel recommendations within the application. For Chesky, these features are “probably the most differentiated part of Airbnb.” Cover image | Unsplash (Oberon Copeland), Wikimedia In Xataka | OpenAI is no longer a startup. Now it is a black hole of 500,000 million that threatens the world economy

The Chinese subsidiary of Nexperia has just broken ranks with its parent company in the Netherlands. And that takes the conflict to another level.

by

Nexperia has gone from being unknown to becoming the new focus of tension in the technological war between the West and China. The company, with Chinese capital but based in the Netherlands, has been intervened by the Dutch Governmentwhich alleges national security reasons. And its impact could soon be felt in sectors as sensitive as automobiles and consumer electronics. The movement is not minor: Nexperia controls an extensive network of factories and assembly centers in Germany, the United Kingdom, the Philippines, Malaysia and China, all important for the global semiconductor chain. Since the Netherlands took over governance of the company at the beginning of the month, a key question has arisen: how far does its control over those international operations really extend? Different laws, one company: Nexperia, caught between Europe and China The answer, at least in part, we already have. Nexperia operations in China have recalled that They work “independently” from the Dutch headquarters. A gesture that not only challenges this European authority, but adds a new layer of uncertainty to an industry that continues to suffer the consequences of the chip crisis. The statement released by Nexperia China on October 17 through its official channel WeChat marks a turning point in the dispute. In the text, signed by all the group’s operating entities in the country, the company reaffirms its autonomy from the headquarters in the Netherlands and remembers that its activity is governed exclusively by Chinese legislation. The document clearly establishes that the legal representative has exclusive authority to make decisions and approve any instructions from abroad: “Nexperia companies in China are independent companies that operate in accordance with national laws. The legal representative has exclusive authority to make decisions and approve any external instructions. No employee is obliged to follow orders coming from outside without their express consent.” The Dutch headquarters, for its part, has denied that “independence” and has attributed it to unauthorized information and actions, which adds another chapter to the internal clash. A ban on exporting its products from China has put European manufacturers on alert, especially the automotive industry, which depends on Nexperia chips for the operation of numerous electronic components. The European Automobile Manufacturers Association (ACEA) warned last week that the situation could cause production stops if supply is not restored in the coming weeks. According to the organization, current stocks would barely cover a few assembly cycles and approving new suppliers would take months, a period incompatible with market demand. One of Nexperia’s facilities in Guangdong Nexperia’s weight in the semiconductor chain is best understood by looking at how its production is organized. Although the headquarters and operational management are located in the Netherlands, much of the group’s added value comes from Asia. Its assembly and test plants in China, the Philippines and Malaysia manage enormous production volumes that supply both the Asian market and Europe. The coming weeks will be marked by the search for a fragile balance between regulators and governments. Nexperia has confirmed that it is in talks with China’s Ministry of Commerce to reverse the export blockade, while the Netherlands retains control of its governance. The question is whether the company will be able to operate normally. without violating either of the two legal frameworks. For now, the signals are mixed: production continues, but under an environment of uncertainty that leaves manufacturers waiting for a quick outcome. Images | Nexperia In Xataka | The problem is not that Europe has “expropriated” Nexperia from a Chinese company: it is that it approved its sale just a year ago

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.