News

We already know where many of the microplastics that enter our body end up. We have bad news

by

That microplastics had managed to enter our body is something that we already knew in detail, especially considering that we have found them in the lungsin the placenta and even in the testicles. However, there were questions about where they accumulated in greater quantities in our body and what consequences does it have. Something that science has already solved. What they have seen. It has been thanks to a recent study published in the journal Environmental Science and Ecotechnology that it has finally been revealed that not only can cholesterol crystals be found in bile that end up generating stones, but also there are microplastics. And the worst of all is that they have a direct impact on the premature aging of the cells that make up our gallbladder. How do you know? To reach this conclusion, the researchers analyzed 14 samples of human bile: five from healthy patients without gallstones and nine from patients with gallstones. The results were conclusive, since they found microplastics in the samples, mainly highlighting two of them. the most common polymers in our daily lives: polyethylene (PE) and polyethylene terephthalate (PET). Here it could be seen that the particles had a size that ranged between 20 and 50 micrometers. A very relevant fact, since from these sizes is when they can cross the biological barriers, cross the intestine-liver axis and end up in the gallbladder of our liver. There is more. In addition to the presence of plastic in the bile, it could be seen that patients who had gallstones showed a higher load of microplastics. Something fundamental, since it is a fact that fits with recent research that suggests that these particles could act as ‘seeds’ around which cholesterol groups to form the dreaded stones in the gallbladder. What do they do? This is the key point of these studies, since we still do not have much idea of ​​the damage that microplastics can do to our body. Here it points out that bile causes mitochondrial dysfunction and promotes the aging of cholangiocytes, which are the cells that line the bile ducts. In previous experiments, it was possible to see that in the liver of laboratory mice exposed to environmental concentrations of microplastics there was alteration in the metabolism of bile acids and liver damage. In the case of humans, which increases oxidative stress. But the important thing is that in both cases the bile duct cell loses its ability to function correctly and ages prematurely, which in the long term could be related to liver and bile duct diseases. Can it be mitigated? Among the bad news, scientific literature suggests that there are ways to avoid it. One of the great protectors that exists is melatonin, suggesting that it can combat oxidative stress and mitochondrial dysfunction generated by these synthetic intruders. In parallel, other recent experiments with human liver organoids have shown that the damage caused by microplastics improves when ursodeoxycholic acid is administered, which is the drug administered to ‘dissolve’ gallstones. The ‘magic’ of this compound is that it increases bile flow, suggesting that promoting a kind of natural “washing” of the bile duct could help reduce toxicity. A problem. The confirmation of bile as a “hidden reservoir” of microplastics highlights an undeniable reality: plastic pollution is no longer just an environmental problem in our oceans, but a systemic public health problem about which we know more and more data. The longer-term consequences, such as the relationship with gallbladder cancer, remain to be seen. Images | FREEPIK In Xataka | An 18-year-old girl has created the definitive weapon against microplastics: a filter that eliminates 96% of them from water

It takes 7,000 GPUs to simulate a tiny quantum processor. Although it may not seem like it, it is excellent news.

by

The complexity of quantum computers It is extraordinary. In their construction it is possible to rely on several very different strategies, such as, for example, superconducting qubits, ion traps or neutral atoms, among other technologies, but they all have something in common: to a large extent its power is a consequence of its complexity. Of the complexity inherent in any device designed to take advantage the laws of quantum physics. The surprising thing is that, despite its sophistication and exoticism, it is already possible to accurately simulate a small quantum processor using conventional hardware. In fact, has achieved it a research group from the Quantum Systems Accelerator and the Division of Applied Mathematics and Computational Research at the University of California at Berkeley (USA). This is not the first time that a quantum processor has been simulated, but until now no one had managed to do it by emulating every physical detail before its manufacture. A new era begins in quantum chip design Here’s a shocking fact: the Berkeley researchers I mentioned in the previous paragraph have carried out their simulation of a quantum chip using the Perlmutter supercomputer, which contains 7,168 NVIDIA GPUs. To achieve their purpose, they used almost all of these GPUs for 24 uninterrupted hours, so it is evident that the computational effort was titanic. But they got it. They managed to model a multilayer quantum chip 10 mm wide and 0.3 mm thick, accurately simulating how signals travel and interact within this processor. This statement from Andy Nonaka, one of the scientists at the Berkeley Quantum Systems Accelerator, express clearly Why this milestone is so important: “I am not aware of anyone who has ever performed physical modeling of microelectronic circuits at the full scale of the Perlmutter system.” “I’m not aware of anyone having ever done physical modeling of microelectronic circuits at the full scale of the Perlmutter system. We were using almost 7,000 GPUs (…) We divided the chip into 11 billion grid cells and were able to run over a million time steps in seven hours, allowing us to evaluate three circuit configurations in a single day. These simulations would not have been possible in this time frame without the complete system” What really what makes the difference is precision with which they have managed to carry out the design and simulation of their quantum processor. “We perform a full-wave physics-level simulation, which means we care about what material is used in the chip, its design, how the metal is wired (using niobium or other types of metal wires), how the resonators are built, what the size, shape and material used are (…) We care about those physical details and we include them in our model,” Nonaka says. A priori we can conclude that using almost 7,000 GPUs for 24 hours with the computational effort and energy expenditure involved in this process to simulate a quantum chip just 10 mm wide and 0.3 mm thick is not a success. But yes it is. Thanks to this technology, it will now be possible to design quantum hardware in less time and in a more efficient way. Bert de Jong, director of the Berkeley Quantum Systems Accelerator, invites us to look towards the future of quantum computing with optimism: “This unprecedented simulation is a critical step in accelerating the design and development of quantum hardware. More powerful, higher-performance chips will unlock new capabilities for researchers and open new avenues in science” Image | Generated by Xataka with Gemini More information | ScienceDaily In Xataka | We already know what the chips that will arrive until 2039 will be like. The machine that will allow them to be manufactured is close

There is a way to open the dishwasher incorrectly. This is bad news for your kitchen furniture.

by

Practically half of Spanish homes have a dishwasher. And some of the most reputable manufacturers have had to come to the fore to explain why. it’s not a good idea Open the door as soon as the washing cycle ends. Don’t be craving. Opening the dishwasher as soon as the wash cycle is finished is not the best idea, according to the manufacturers themselves. Some, like Boschrecommend waiting for it to cool a little before opening the door so that less steam comes out of the opening. Others, such as Siemens, make exactly the same recommendation: let the appliance cool down slightly after the program ends before opening the door. Because. Opening the dishwasher at the end of the cycle is not particularly dangerous, but it can cause problems in your kitchen in the long term. Manufacturers point out that excess steam when opening the door when it is still quite hot can end up damaging kitchen furniture. Similarly, dishwashers use their own residual heat to help the correct drying process. All this without taking into account the most obvious detail: if we open the appliance when it is full of high-temperature steam, we run a greater risk of suffering a slight burn. How the dishwasher works. To understand why the dishwasher generates so much steam at the end of the cycle, it is worth doing a brief review of how it works. Basically, it is a closed circuit that allows hot water to be pumped into rotating arms. When you choose the program, the door closes. Water enters its base and the dishwasher heats it. When the water is hot, the pump pushes it under pressure into your arms. The water hits the dishes and drags away the dirt. The detergent is released and begins to break down grease and food debris. The residual heat dries the dishes little by little. The heat problem. Aware that opening the dishwasher as soon as it is finished is common practice in homes, manufacturers have been devising solutions for years to prevent excess steam. Some appliances of this type have an automatic opening system. The door opens just a few centimeters to let the steam out little by little, so that when we finish opening it manually we avoid that initial blow. Other systems, even more advanced, they use zeolites. Zeolites are microporous aluminosilicates (of mineral or synthetic origin) that have a crystalline structure with a huge internal surface area. When they absorb water molecules the process is exothermic, that is, they release heat. This allows part of the steam to be collected and used as waste heat for drying. Increasingly advanced solutions to address a problem as simple as it is common. In Xataka | A user bought a next-generation connected dishwasher. That’s where his nightmare began

SpaceX is now a company in the railway sector and it is very bad news for its employees

by

For some people it will be ingenuity, for others a very hard face, but the point is that SpaceX has found a way to avoid lawsuits and strikes by its workers when obtaining the name of air transport company. This means that it is regulated under the Railway Labor Law, with all the benefits that it entails within US legislation. The news. On March 13, the official resolution was made public by which SpaceX, Elon Musk’s space agency, is now considered a company in the railway sector in the United States. This means that your activity is no longer subject to the supervision of the National Labor Relations Board (NLRB)which is typically responsible for protecting the labor rights of private sector workers. The layoffs that started it all. In January 2024, the NLRB put a lawsuit against SpaceX on the tableafter the company illegally fired 8 employees. The lawsuit requested reinstatement of the employees, back pay, and a letter of apology to each of them. Given this situation, SpaceX responded with another lawsuit to the NLRBalleging that the procedure being carried out was unconstitutional. Rockets have the same legal treatment as cargo planes. An ace up your sleeve. According to Elon Musk’s company, the NLRB should not be able to act against a company that is dedicated to transportation. He added that One of its main missions is the transport of humans and goods to the International Space Station.. In many cases, these jobs are carried out for NASA, so they would also be providing a service to the Government. For all this, they requested to be covered under the Railway Labor Law. A plan that suits many. In recent years, SpaceX, as well as other Elon Musk companies, have been the subject of complaints from a multitude of dissatisfied employees, either due to their personal situation or due to bad practices carried out in the company. In the case of Neuralink, for example, Very bad practice towards laboratory animals was reported. But returning to SpaceX, the increasing volume of complaints could put the company’s work pace at risk. This, logically, would harm its managers, but also the companies that benefit from its services. The entire US space program would probably collapse. For all this, although it seemed difficult, in the end Elon Musk’s company has had a resolution in favor of its new name. Immune to strikes. One of the peculiarities of railroad companies in the United States is that they benefit from special state protection. Since minimum transport services must be guaranteed, strikes and other similar activities that would normally slow down the normal pace of work are closely controlled. The NLRB no longer rules. Another of those special protections for railroad companies is that the NLRB no longer has power over them. Therefore, dismissed employees cannot resort to it to report their situation. Instead, the company is governed by the rules of the National Mediation Boardmuch more lax in the mediation of labor disputes. It is true that employees can request strikes, but to do so they must undergo a long and tedious process that often causes them to change their decision. And now what? With this new name, SpaceX has even more power and freedom than before. If measures are carried out that involve malpractice towards employees, it is difficult for their complaints to come to fruition legally. This gives them a lot of leeway and greatly speeds up their protocols. Other curious legal victories. It is not the first time that SpaceX has obtained an unexpected legal name. Last year, for example, The Starbase base was given the name of cityso that all employees who live nearby would also become inhabitants. This, far from changing a few patterns, also gave SpaceX more freedom when maneuvering in the areas surrounding its base. As with railway legislation, what may seem like a small name change can change everything. Image | Gage Skidmore (Wikimedia Commons) |SpaceX In Xataka | SpaceX is preparing the largest IPO in history: the fact that it is doing so right now is no coincidence

The good news is that the Ebro reservoirs are at a historic 85% water level. The bad thing is that we are going to spend it in a short time.

by

There are 6,640 hm3 of water in the Ebro basin. The reservoirs are at 85.1% of its total capacity at what is its highest level (for this date) of the decade. And yet, the fact that there is a lot of water is not news. All of Spain is the same (83.3%). The news is that we are going to spend it. A structural problem called ‘Mediterranean’. Every year, the pressure of the Mediterranean summer and the irrigation campaign empty the reservoirs very quickly. AND, as history has shown usthere is never too much water: “each dry period has served to implement emergency measures for agriculture that were not eliminated when the rains returned, they were used to expand irrigation, aggravating the problem in the following drought”, said Ana Tudela and Antonio Delgado. And that, precisely that, is what we are about to see. The complete image. Seeing the figures for the reservoirs can lead us to forget that, just three years ago, 85% of the basin’s surface was in “prolonged drought“and 45% of them declared themselves in shortage emergency. Mequinenza, the largest swamp, reached historic lows. It was a catastrophe not only in water terms, but also in energy terms. Now, however, all that is in the past. And Say’s Law lurks in the dark. What the old French economist Jean-Baptiste Say argued at the end of the 19th century is that “every supply creates its own demand“and, translated into this situation, this means that the fact that there is more water generates all the incentives in the world for there to be more irrigation. As soon as we do it, this becomes clear. After all, not all of the basin’s storage capacity is enough for a full year of agricultural demand. Without the annual rainfall and the melting of snow, we could already consider all its reserves exhausted. March is the key month. The irrigation campaign runs from April to September and that means that March is the key month for planning the year. It is true that the thaw has not yet begun (which this year is going to be very intense), but it helps us estimate what quantities of water are really available. All irrigated agriculture in the valley depends on the water we are able to store during the spring. The question from now on becomes: how do we conserve as much water as possible before we once again enter a situation of risk? And the problem is that we don’t have answers. Especially in a regulatory context in which are not foreseen widespread restrictions on irrigation. Economic, social and institutional incentives tell us that we are not yet prepared, as a country, to address the really important question: we do not have a water problem, we have a consumption problem. There is still room for improvement in management, yes. But that won’t solve the problem: it only postpones it. And that 85% of reservoir water has given us unbeatable weather, we just have to hope that we can take advantage of it. Image | Manuel Torres Garcia In Xataka | The great battle of the Ebro is not between Murcia and Aragón, it is between the headwaters of the rivers, the large cities and the delta

Beijing just gave the best news to NVIDIA

by

NVIDIA has been caught in the crossfire of the trade war between the United States and China for more than a year. Its most powerful chips could not be sold to the Asian giant because Washington required export licenses and later Beijing did not give the green light to imports. This week the two fronts have been unlocked simultaneously and Jensen Huang has taken advantage of his annual developers conference to announce it out loud. Your factories they are starting enginesand the future is brighter than ever for her. H200 chips, unlocked. The H200 chip, NVIDIA’s second most powerful chip today, had become the center of trade and technological tensions between China and the United States. The Trump administration had already granted export licenses but of course, as long as she got her cut. Which was missing it was him approval from Beijingthat according to Reuters It has arrived this week for many of the customers who were demanding access to these chips. Among them are ByteDance, Tencent, Alibaba and DeepSeek. China is a gold mine. Before the restrictions imposed For the US, China represented 13% of NVIDIA’s total turnover. The export veto was highly criticized by Jensen Huang, who did not stop criticize the measureof try to avoid it and to explain that what the US had done was not protect its technology, but rather shoot itself in the foot. During this blockade, Chinese companies have been advancing both in the development of AI models and in the development of their own chips. They still have room for improvement, but this effort to “become independent” from US technology is already bearing fruit and perhaps would not have occurred if it were not for the US veto. There will also be Groq chips. NVIDIA will not only export its H200 chips, but is preparing a version of its AI inference accelerators for the Chinese market. Specifically, we are talking about Groq chips, a company in which NVIDIA invested $20 billion to “license” its technology, although In practical terms I have acquired it. These chips are especially interesting because they are not used to train AI models, but to execute and “serve” them. This is the market that is growing the fastest right now, and where competition is toughest. But China already has inference chips. Companies like Baidu they are already producing its own inference chips, which means that NVIDIA will not enter the Chinese market from a monopoly position, but as another competitor. What is striking here is the fact that Groq chips are not refined versions nor are they adapted to this market according to the cited sources. in Reuters: They will be the same ones that companies in the US or other parts of the world use. China will continue without access to Vera Rubin. This week NVIDIA presented a new line of products built around its next AI chips, the Vera Rubin. These chips cannot be sold to China due to current restrictions, so at NVIDIA they have found a hybrid architecture: Vera Rubin for markets where it can operate freely, and Groq as an inference component for China. NVIDIA is brimming with optimism, and with good reason. Jensen Huang spoke precisely in his inaugural conference about how promising the company’s future looks. Previous projections spoke of medium-term revenues of $500 billion for its Blackwell and Rubin chips. Inference solutions and this “opening” to China now mean that this forecast is doubled: Huang hopes to achieve at least $1 trillion in cumulative orders by 2027a simply dizzying figure that makes it clear that NVIDIA’s business seems to be in an enviable state of health. Image | NVIDIA In Xataka | DLSS 5: Millions invested in AI graphics improvements so people say it looks like an Instagram beauty filter

China already has two chip manufacturers with 7nm technology. This is very bad news for the US and its allies.

by

SMIC (Semiconductor Manufacturing International Corp), the largest Chinese semiconductor manufacturer, has the capacity to produce 7nm chips from 2023. At the beginning of September of that year this company shook up the integrated circuit industry by demonstrating that it had been able to manufacture these semiconductors despite not having access to UVE photolithography equipment (extreme ultraviolet) produced by the Dutch company ASML. These highly sophisticated machines are necessary to manufacture cutting-edge chipsand ASML cannot sell them to its Chinese customers because the US, which controls some of the patents used by this equipment, prohibits it. Still, in 2023 SMIC and Huawei worked hand in hand to manufacture 7nm ICs without using ASML’s EUU machines. Of course, they used deep ultraviolet (UVP) equipment that this company from the Netherlands also produces. UVP machines are not as advanced as UVE machines, but with proper refinements they can be used to manufacture cutting-edge integrated circuits. And now China has two semiconductor manufacturers capable of producing 7nm chips. As we have just seen, one of them is SMIC, and the other, according to Reutersis Hua Hong Semiconductor, the country’s second largest integrated circuit producer. The shadow of ‘multiple patterning’ is very long Hua Hong Semiconductor’s division specializing in third-party chip manufacturing is called Huali Microelectronics, and, again according to Reutersis preparing to start production of 7nm integrated circuits at its Shanghai plant. The sources that have revealed this information assure that Huawei has collaborated with Huali Microelectronics on this project, which invites us to reach two reasonable conclusions. It is very likely that with the help of Huawei, Huali has developed ‘multiple patterning’ techniques The first is that Huali’s 7nm lithography will most likely play an essential role in GPU production capacity for artificial intelligence (AI) from both Huawei and other Chinese companies. And the second conclusion is actually a plausible hypothesis. And, like SMIC, Huali does not have access to ASML SVU equipment. For this reason, it is very likely that with the help of Huawei it has developed security techniques. multiple patterning to be able to manufacture 7nm chips with the UVP machines in its possession. A priori, UVP machines are suitable for manufacturing semiconductors up to 10 nm. And with EUVs it is possible to exceed 3 nm. However, by refining the processes involved in transferring the pattern to the wafer and turning to multiple patterning It is possible to go beyond these integration technologies. This technique broadly consists of transferring the pattern to the wafer in several passes with the purpose of increasing the resolution of the lithographic process. It may have an upward impact on the cost of chips and a downward impact on production capacity, but it works. Either way, Huali Microelectronics is going to face the same challenges that SMIC has dealt with for the last three years: the multiple patterning seriously limits the number of viable chips per wafer that is possible to manufacture. And, therefore, it increases its cost. Still, it is very important for Chinese AI chip designers to have access to two companies capable of producing their designs with advanced photolithography technology. And for Hua Hong Semiconductor it is crucial to have the ability to manufacture, thanks to Huali, cutting-edge integrated circuits and not just 22 nm or larger (this is the most advanced photolithography it had so far). Image | Generated by Xataka with Gemini More information | Reuters In Xataka | TSMC is already the highest-earning chipmaker on the planet. It has beaten two semiconductor giants

MásOrange has begun to completely dismantle its 3G network. Not good news for elevators

by

If you are one of those who usually browse even on a 3G connection, we have a curious fact: you are from the club of 1.82% of global traffic. The operators carry years saying goodbye to this networkand MásOrange has been the last to take the step. The operator is going to say goodbye to its 3G network for good reason. Hello to 5G. MasOrange has been the last of the large operators to start with the shutdown of 3G. Vodafone finished turning off its networks two years ago, and Telefónica is about to finish the process. With the release of the spectrum used by 2G and 3G connectivity, operators have additional bands to reinforce 5G technology. Specifically, the 900 and 2,100 MHz bands allow expanding coverage in areas with lower population density without the need to build new towers. A progressive plan. Although 2G and 3G connections sound completely obsolete, they are still necessary to connect a good number of day-to-day technologies. Elevators Cars with eCall system with 2G modules Telecare services security cameras Home alarms Old dataphones For this reason, the plan to dismantle the Spanish 3G network is being carried out progressively, giving time for a good part of these devices to update their connection modules. Although 3G was scheduled to say goodbye forever between 2025 and 2026, current plans keep it alive until 2030. A great challenge ahead. The 3G network continues to be a great ally, both for older devices and for times when 5G coverage does not have range. The challenge now is, precisely, that the deployment of 5G is even greater. For three decades, 3G invaded every corner of Spain, and some of the networks it uses (900 MHz) are especially good at passing through walls and operating in rural areas. Therefore, in complicated areas your mobile still connects to 3G. The end. Maintaining 2G and 3G networks is unsustainable. The radio space they occupy is especially valuable and, although there are devices that still use old technology, their dismantling is necessary. Cover image | Baatcheet Films In Xataka | How to request an eSIM from each operator in Spain: in which cases it is free and application methods

For the first time, BYD has sold more cars outside of China than inside. It’s very bad news for them.

by

Pursue your dreams… outside of China. Beyond Your Dreams are the words hidden behind the acronym BYD. The acronym of the company that sold the most electric cars in the world in 2025. A company that seemed to have meteoric progress but that has stagnated with a local market that is slowing down at a dizzying pace. So much so that it has already sold more cars outside of China than inside. A well-thought-out strategy that arrives ahead of time. A milestone?. The month of February was the first in which BYD has sold more vehicles outside China than in its own market. It is a conditional milestone, since sales in China of the entire market have plummeted and, of course, have hit the country’s largest car manufacturer hard. A general trend. Although in 2025 car sales in China once again set a record with 34.4 million cars sold (a growth of 9.4%, according to the Chinese Association of Automobile Manufacturers) in figures collected by the media 36krthe market has been experiencing a slowdown for months. In February, car sales in the country fell 15% compared to the same month in 2025, they point out in Reuters. But the problem is worse among individuals, where sales have fallen by 34% as a result of the Chinese New Year festivities and the withdrawal of some purchasing aid. The latter has had a direct impact on sales of “new energy” cars (plug-in hybrids and electric). According to data collected by CNEVPostIn January, 596,000 cars of this type were purchased, compared to 744,052 units in the same month of 2025. A drop of 22.1% that worsened in February, with 464,000 units sold compared to 686,000 units the previous year. It is a year-on-year drop of 32%. The BYD case. This general decline in sales, with more worrying figures among new energy vehicles, has had a direct impact on BYD. Last January, BYD sold 210,051 new energy cars when in the same period of 2025 it placed 300,538 units on the market. In February, the figures were worse with 190,190 units sold compared to last year’s 322,846 units, reported in CNEVPost. That is, so far this year, its sales have fallen by 30% in January and 41% in February, extending a trend of low sales that has been going on since September of last year. BYD sales have not grown in China since June 2025. In July and August they achieved a technical tie in the year-on-year comparison but, since then, they have lost in all one-year comparisons. These falls have caused Geely to surpass BYD in sales in the first two months of the year. Between January and February, Geely has sold 476,327 units, just 1% more than in the same period of the previous year. There are just over 76,000 units than BYD (400,241 vehicles between January and February) thanks to a larger product portfolio and less dependence on “new energy” vehicles. This has avoided a fall due to the withdrawal of state aid, they state in SCMP. More outside than inside. As we said, BYD’s sales have plummeted in China but its exports have skyrocketed abroad. This has meant that the company has sold, for the first time, more cars outside its borders than within its borders, they point out in Electrive. Two factors explain it 41% decrease in sales in China compared to February 2025 Increase in exports of 50.1% compared to February 2025 The company has managed to consolidate sustained growth in its exports. They point out in CarNewsChina which with February now adds up to four consecutive months exporting more than 100,000 units. This has caused them to place outside their borders this month 100,600 units of the 190,190 units which have sold all over the world. That is, more than 50% of its sales have been delivered outside of China. a mirror. BYD has become the best example of what the Chinese market is all about. The country lives in a whirlwind of launches and a suffocating price war. BYD itself, with its new launches at ultra-competitive prices, has caused their own cars become obsolete with months on the market, gathering dust in dealerships. The rest of the companies have also played to lower prices to keep up the pace and release news at a frenetic pace, but that produces some anxiety in the client that sees how what is new today can be left behind very soon. Bad news in a country like China that has been trying for years to promote domestic consumption to put its economy into higher gear. But, in addition, the State has withdrawn some aid to the purchase of electric cars, the most important column in the industry. This has its consequences in the drop in sales among individuals. Before time. That BYD intended to expand outside China was no secret. In fact, his plans happened because half of global sales will be consolidated outside of China in 2030. The expansion plan with the factory Hungarythat of Türkiye and, it is rumored, another in Europe is part of it, without forgetting the Thailand and Brazil. The question is to know if this surprise of sales abroad has arrived too soon and the only thing it confirms is the slowdown that the brand will have to deal with in China. If you want to consolidate yourself as one of the largest global manufacturers (there was talk of reaching 5.5 million units in 2025 but finally they stayed below the 5 million border) it is essential that they expand borders and not depend solely on the internal market. European manufacturers can give BYD some examples of what happens if you base the bulk of your strategy on selling in China. Photo | BYD In Xataka | The year of Chinese consolidation in Spain: MG, Omoda and BYD close a spectacular 2025 and are among the best sellers

The good news is that AI models are becoming more powerful. The bad thing is that everyone ends up saying the same thing.

by

We have artificial intelligence. What we don’t have is artificial diversity. That is the conclusion reached by a group of researchers who did a relatively simple test: they asked 25 different AI models a bunch of questions to see what they answered. And that’s the bad thing: who answered things that were too similar. “Artificial hive mind”. Scientists from the University of Washington, Carnegie Mellon University and Stanford University, among other institutions, have published an interesting joint study. In it they reveal how after various tests it seems clear that although AI models are becoming more and more advanced, the problem is that they all seem to have developed a kind of “artificial hive mind”: no matter what you ask them, they answer in a suspiciously similar way. When asking all these models “what time was”, many responded with the phrase “time is like a river”, while another group of models answered that “it is like a weaver”. time is a river. One of the questions asked of these models is “What is time?”and although that question leaves clear room for very different answers, the worrying thing is that they were not. Several models responded with the phrase “time is a river” and then developed it a little, while others responded with “time is a weaver (of moments).” That similarity when it came to responding turned out to be a constant. The illusion of abundance. We believe that when we consult something with an AI we access a whole world of conversational possibilities, but the study reveals that in reality we are facing a system that proposes very similar outputs. Although language models promise limitless creativity, they tend to converge on that hive mind where diversity is sacrificed for statistical consistency. It is reasonable, especially considering that large language models They are based on the concept of transformera probabilistic system that tries to find the next “best” word as it answers us. Same script. The researchers created a large-scale data set with 26,000 queries from real users that theoretically allowed the models to generate multiple valid and creative responses. They called that data set “Infinity-Chat” and divided the questions into six main categories and 17 subcategories. IA, you repeat yourself more than a broken record. During the tests it was observed that the same model tends to repeat itself, generating very similar responses. In fact, even when special parameters were used for questions designed to encourage diversity, the same effect was produced. This is what researchers call “inter-model collapse.” Too similar. These tests made it clear that the semantic similarity, how similar the responses of the different models were, was worrying. According to the study, this similarity ranged between 71% and 82%, and in some cases certain models managed to generate identical paragraphs word for word. The training problem. It is not only that they all generate text in a similar way due to their design, but there is also a training problem. The authors suggest that this homogeneity of responses could be due to several reasons: Training data sources end up being shared: models They are trained with similar “datasets” and for example they are based on similar texts and knowledge that come, for example, from Wikipedia or a very similar set of books. Contamination effect due to synthetic data generated by other AIs: they also use synthetic texts generated by other AI models. Rewards: The models used to reward these models are calibrated to reward some notion of “consensus” quality. Thus, creative and individual diversity is punished. AIs are “educated” to be precisely very similar to each other. Problem in sight. All of this makes researchers explicitly warn about two clear risks when using these AI models. We will think the same: if we users do not stop using AI models that answer basically the same thing, our own ways of thinking on those topics and problems will be “homogenized”and it will also make our responses more uniform. Point of view reduction: The other danger follows from the first: if the AI ​​ends up converging and answering the same thing, points of view are eliminated. Here the biases for example from the western world will be evident in Western models (ChatGPT, Gemini, Claude), and the same will happen with the oriental ones, for example. This would cause the potential suppression of alternative worldviews, of perspectives and “looks” that are different from our reality. Image | Solen Feyissa In Xataka | The scientist who made the AI ​​we know today possible has just raised 1 billion. His new goal is to teach him to see space

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.