Mobile Tech & Gadgets

European telecos are discouraged before global competition

by

The Departure by José María Álvarez-Pallete de Telefónica It marks the end of an era that symbolizes the decline of the sector in Europe. Its mandate ends with a 57% drop in the stock market value of the company since 2016, when it agreed to the position. Of course the problem goes far beyond Telefónica. Why is it important. The European Telecommunications sector has lost 41% of its stock market capitalization between 2015 and 2023, falling to 270,000 million euros. In contrast, American telecos are worth more than 650,000 million. The root of the problem. Europe has 34 mobile network operators and 351 virtual operators for 450 million inhabitants. The United States, with 335 million inhabitants, only has three main and 70 virtual operators. China, for 1.4 billion inhabitants, has four operators and 16 virtual. Proportionally to its populations … Europe has eight times more main operators than the US and 27 times more than China. In virtual operators the difference is even greater: almost four times more than the US and 78 times more than China. Between the lines. European regulation has prioritized low prices and high competition, creating a fragmented and unable sector to compete globally. Meanwhile, the US and China have opted for giants with financial muscle. This dynamic has had consequences: it is expected that the sector only grows 1.7% in 2025. In Spain, 71,000 jobs have been destroyed since 1998, according to appointment Five days. And while the 5G coverage in Europe is 81%, in the US and in China it exceeds 95%, according to The independent. Turning point. He Draghi report Ask to facilitate mergers and end market fragmentation. The new European Competition Commissioner, Teresa Ribera, He said that the rules “will evolve” to allow larger scale. It is a burning nail to which you have to grab. Without a regulatory change that allows consolidation (and perhaps requires greater contribution to large technological ones whose business goes through the use of these networks, A historical demand Of the great telecos), the European Telecos sector will continue to lag in the global race. In Xataka | The discreet engineer: why the great shareholders see in Murtra the manager who needs telephone for his next decade Outstanding image | Telefónica

detect a plane at 5,000 meters high

by

China has been around for a while developing all types of sensors and radars whose characteristics suggest a future very different from the current one. Without going too far, last September They tested a device called LARIDa super radar that was able to detect a gigantic bubble of invisible plasma in the Pyramid of Giza. The latter is something very different, and can change the battlefield. Unprecedented underwater radar. In a secret location 1,000 meters deep, Chinese scientists have developed and tested what points to an unprecedented underwater sensor capable of detecting aircraft at high altitude from the depths of the ocean, representing an unparalleled strategic advance for the People’s Liberation Army (PLA) submarine fleet. To give us an idea, technology managed to identify a fixed-wing aircraft at 5,000 meters above sea levela milestone that no other nation has achieved to date. Such technology, which until now was considered impractical due to the complex acoustic properties of water, could dramatically improve the survivability of submarines against warplanes. antisubmarineequipped with advanced sensors to locate and destroy them. Implications and capabilities. Underwater radar represents a significant strategic advantage for the nation. The reason? Turn the current combat around, allowing your submarines to go from targets to hunters in anti-submarine warfare. Currently, aircraft equipped with advanced sensors can detect and attack submarines without being spotted, but with this new technology Chinese submarines could, a priori, receive early warnings of hostile aircraft and launch missiles from underwater with precision. And there is more, of course. The main technological challenge lies in the fact that most of the sound waves generated by aircraft are reflected from the sea surface, making their detection difficult. However, researchers at the Institute of Acoustics of the Chinese Academy of Sciences, led by Zhang Bo and Peng Zhaohuiemphasize that they have achieved an innovative solution by taking advantage of the multiple reflection of sound waves between the seabed and the water surface, allowing Extend detection range and improve radar sensitivity. System testing and limitations. Apparently, In 2022, China carried out secret tests of underwater radarmanaging to detect a plane 17 kilometers away and, attention, with a margin of error of less than one kilometer. Still, limitations were identified, such as a blind spot when aircraft fly directly over the sensors, as well as a detection range that still needs to be expanded for effective military applications. In this regard, Zhang and Peng continue working on system optimizationalthough they have not revealed recent data. Complement to the military intelligence network. China’s military strategy is based on a multi-layered approach that integrates various detection technologies. Currently, the People’s Liberation Army has a satellite network capable of tracking high-performance aircraft like the F-22 stealth fightersin addition to radar systems on ships and land stations. The new underwater radar would act as a complement, providing redundancy and improving detection capability in hostile environments. Future perspectives. Although the details of this development were kept secret for years due to its military sensitivity, the research team has recently published an article in the journal Acta Acusticaa job where information about the functionality of the system is provided. That said, and as is often the case with innovations of this type, everything indicates that there remains a long road of testing and improvements in operational scope to guarantee its effective integration in real combat scenarios. What does seem abundantly clear is that the innovative underwater detection system marks a paradigm shift in naval warfare, one where, in the absence of verifying it on the ground, It would give China a huge tactical advantagehighlighting its leadership in the development of cutting-edge underwater surveillance technologies. Image | Marinha In Xataka | China has tested its new super radar. They have detected a gigantic bubble of invisible plasma in the Pyramid of Giza In Xataka | China gains muscle in aerial technology: it already has new radars against drones, missiles and stealth aircraft

It is becoming increasingly clear that there is no “normal” body temperature.

by

If you ask us what the “normal” temperature of our body is, the instinctive answer will be 37º Celsius. When the thermometer exceeds that mark, we usually talk about fevermild or high depending on how far we move away from the figure. However, over time health experts have realized that the reality is a little more complex. The body temperature issue It is not a mere curiosity. Fever is an important response of our body to many diseases or disorders, generally to infections. The fever It is a double-edged sword: our body raises its temperature to try to kill viruses and bacteria that may be damaging it, while activating our body’s immune response; However, in the process, fever can also put the proper functioning of our organs at risk and cause other problems such as dehydration. Since fever is a common response to various illnesses, it can also cause us serve as a diagnostic toolto narrow the circle on the possible conditions that affect us. Answering the question of what is the “normal” temperature of our body is difficult. And the reasons behind this are several. Firstly, because, over the last century and a half, the estimated average temperature of the human body has been reducing. The notion that our body temperature It is at 37º and dates back to the mid-19th century. In 1868, the German doctor Carl Reinhold August Wunderlich conducted a study using 25,000 patients and more than a million temperature measurements. From these data, he calculated that the average temperature was 37ºbut also observed certain deviations. However, more recent studies have observed lower average temperatures. A recent example of this we found it in a studio Made in the United States and published in 2020 in the magazine eLife. The analysis indicated that Americans’ body temperatures had been dropping at a rate of about 0.03º Celsius every decade. A previous study conducted in the United Kingdom and published in 2017 in the magazine B.M.J.estimated an average temperature of 36.6º in its sample of more than 35,000 participants and 250,000 measurements. We don’t really know why body temperature has been reducing over time. A possible explanation It lies in the improvements in hygiene and immunity, which would imply a lower incidence of infections in the population and therefore lower average temperatures. But this is just one of the various hypotheses that seek to explain the phenomenon. Wunderlich himself observed in his study that men and older people tended to have lower body temperatures, while women and younger people had higher temperatures on average. Which brings us to the second reason why establishing a “normal” reading is especially difficult. And it depends. Sex and age are two of the factors that can make what is “normal” for one person not “normal” for another. But other factors can also alter this figure. a study published in 2023 in the magazine JAMA Internal Medicine measured the degree to which these factors affected body temperature, but also added new variables such as height, body mass, and the time of day at which the measurement was taken. Among the sample of 618,306 observations, the average temperature was at 36.64º Celsius. Among the participants, the average readings for each individual ranged between 36.24º and 36.89º. It is also worth remembering that there are different ways to measure body temperature (tympanic, oral, axillary…) and that each one It presents some slight associated deviations. So at what temperature fever comes? As is evident after what we have read, the answer is that it depends on each person and situation, although fortunately, with the variations being less than one degree, the interpretation of the results of a thermometer may not be as different from the conventional one as to affect decisions such as whether or not to stay home during a cold. However, for health experts, having better knowledge about these variables can be of great help. That is why new studies have also investigated this question. One published in November of last year in the magazine Scientific Reports by South Korean researchers, analyzed the body temperature of 9,195 hospital patients through tympanic temperature measurements (the tympanic temperature It is usually half a degree above the oral measurement and about one degree above the axillary measurement). The team estimated an average temperature of 36.91º Celsius, and a limit of 37.81º for fever. In Xataka | What to do when we have the flu: what measures to take and in which case we should ask for help Image | Polina Tankilevitch

We believed that the price of coffee could not rise much more. The diplomatic “war” between Colombia and the US thinks otherwise

by

Of the tens of thousands of words that make up the English lexicon, Donald Trump has one that he especially likes and for which he has declared his love in some or other interview: tariff (tariff). This weekend he reminded the Colombian president of this in a quite practical way, threatening to impose 25% rates (or even 50%) if he did not give in to the aggressive immigration policy which is promoted from the White House. Everything indicates that it will remain that way, a threat, but it serves to warm up a market that has been facing strong shocks for months. turbulence: the one with coffee. Yes 2025 It looked complicated For lovers of morning espressos, your outlook has just become more complicated. What has happened? That Trump has shown that, indeed, he feels a special weakness for the word “tariff.” Over the last few weeks it has announced more or less clearly that it will apply taxes on imports of China, Mexico, Canada, Europe, Denmark and even Spainalthough it is still not entirely clear whether the latter was said deliberately or as a result of a geographical ‘slip’. Curiously, it has been another country that has been on the verge of suffering tariff fury from the republican: Colombia. Screenshot of Trump’s announcement on Truth Social. Why’s that? For something that actually has little to do with the international market, trade balances and tax policy. The trigger has been migration. And a political fight between Washington and Bogotá. Basically, yesterday the Colombian president, Gustavo Petro, refused to allow two planes from the US loaded with deported Colombians to land in his country. What’s more, he threatened not to welcome them until Trump adopts protocols that guarantee treatment. “with dignity and respect” for immigrants. The response of the Republican, who has managed to return to the White House after an electoral campaign that largely pivoted on a hardening of immigration policy, it did not take long to wait: through its platform Truth Social advertisement a 25% rate for the import of Colombian merchandise that would rise to 50% in a matter of days. Petro responded after a few hours with the same currencyordering a sudden increase (25%) in the tariffs that Bogotá applies to US goods. How did the crash end? In dispatches and without reaching customs. At least for now. Despite its initial reaction, the Petro Executive ended up giving in to Washington’s demands and agreed to receive the planes with deportees. Enough so that Trump has not yet signed the economic sanctions, which have already been drafted and will be activated if his southern neighbor “does not comply” with the agreement. “The Government of Colombia has accepted all of President Trump’s terms, including the unrestricted acceptance of all illegal aliens from Colombia returning from the US,” they boast from the White House, which reminds that Bogotá will also receive them “without limitations or delay.” Click on the image to go to the tweet. What does it have to do with coffee? Simple. The announcement of Trump’s tariffs and the fight between the White House and the Nariño Palace did more than shake up American diplomacy. He also put on guard various sectors Colombians who have important interests in the United States, such as oil, floriculture (which is preparing for the millionaire campaign Valentine’s Day) and coffee. Of all of them, the one more expectation generatesdue to the state of its market and price driftis the latter. At the end of the day, Colombia is not just any country on the international coffee map. And the United States is not just another market for Colombian producers either. This double condition means that everything that affects the relationship between the two, including of course the threats of 25% tariffs or even 50%, interest (and quite a bit) to the market. But… What does the data say? To begin with, Colombia is one of the main coffee powers on the planet. The own tables The US Department of Agriculture places it as the third largest producer, only behind Brazil and Vietnam. Other observatories leave the same drawing, like Statista. A 2024 reportThe USDA office, linked to the US Government, estimated that during the 2024/2025 campaign, Colombian coffee exports would total around 12 million GBE bags. Colombia matters on the global coffee map. And its relationship with the United States is also important, something that is better understood with the help of a couple of figures. According to the Observatory of Economic Complexity (OEC), in 2022 Colombia exported 15.6 billion dollars to the US, of which almost 1.8 billion corresponded to coffee. This data makes it the second largest exported commodity in terms of value, only behind crude oil. The footprint of Colombia in the flow of coffee that reaches the US is also considerable. Click on the image to go to the tweet. But the rate would affect the US, right? The US tariff policy is that, the US tariff policy, and as warned The New York Times Yesterday, basically imposing 25% taxes on Colombian goods would mean that Americans would have to dig deeper into their pockets to buy flowers and coffee. However, if we talk about grain, what happens on the other side of the Atlantic interests us. And the reason is very simple: shocks like the one on Sunday put even more tension on a market that is already go through turbulence. “If the US imposes a 25% tariff on all Colombian exports, the already red-hot coffee market will heat up even more. Colombia is the third largest coffee producer in the world (and a key source of rabi beans). premium)”, I was reflecting yesterday in X Javier Blas, Bloomberg columnist. In the same tweet he included a column written by himself a few days ago in which he warned of the complex panorama facing the coffee market. And what situation is that? In Xataka we have talked already several times her. And it is summed up … Read more

In the middle of a huge “train of storms”, AEMET has just pronounced the most anticipated word of winter: snow

by

The rain no respite and he’s not going to give it. Neither the rain, nor the winds, nor the snow. While the storm Herminia full of red notices the entire northwest part of the peninsula, the forecasts have us all looking at an Atlantic full of surprises. Among them, snow. What Herminia is doing. In recent days, the entire country has been revolving around Herminia. In some regions (such as Galicia) the intensity of the winds forced to divert flights and cancel trains. The worst, however, had not yet passed: this Monday will be the strongest day of the storm. 16 autonomous communities They have notices from the AEMET. Until Tuesday, Herminia will unfold a temporary more than considerable with wind, waves, rain and snow. But, as we said, Herminia is nothing more than the appetizer. According to predictive models, two new storms are going to cross the peninsula. The first will make its debut on the same Wednesday. In fact, on Tuesday a very cold air mass of polar origin will begin to enter the country. We talk about “temperatures at about 5500 meters altitude (500 hPa)”, explained Sergio Escama in Meteored. But the interesting thing is that we will notice it down here. What will we notice? The winter. Winter in its meteorologically pure state. From Wednesday the general drop in temperatures will be evident and snow will affect to the Cantabrian Mountains, the Pyrenees, the Central System, the Iberian System and other inland areas. The snow level is not clear yet, but is expected to be low. And that will be only the first of the post-Herminia storms. Because, according to AEMET modelsafter a very entertaining Wednesday-Thursday with the snow (and its problems); A busy weekend will follow. That is, the “storm train” extends as far as we can reliably see. This precision is important because, although right now there is a possibility that this second storm will be more intense than the previous ones, the time frame prevents us from being able to estimate its virulence accurately. What can we do? Luckily, The recommendations for now are the same: avoid unnecessary trips; prepare for the cold, wind and snow; and pay attention to weather and civil protection updates. We already know that this is one of the great paradoxes of snow: it is something as wonderful and fascinating as it is full of problems. Image | AEMET In Xataka | The best tips for a snow or hail storm

The price of energy that its chips factory is using in Ireland

by

The tuning of an advanced semiconductor factory equipped with machines with avant -garde lithography It costs up to 30,000 million euros. This is precisely what will cost the plant that Intel will build in Magdeburg (Germany) If the project finally goes ahead. And, as we tell you at the beginning of last November, this company has decided delay the start of construction of this factory until 2029 or 2030. Initially the works were going to start during the first half of 2023, but The negotiation of subsidies forced to delay this date until the summer of 2024. However, Finally Intel and the German government agreed as a construction start date May 2025. Anyway, the four or five years of lag planned by this company are a consequence of the delicate situation in The one that is intel. The cost of energy is crucial to preserve the competitiveness of a factory The center of the conversation on this occasion is not occupied by the Magdeburg plant, although we stay in Europe; It is monitored by the factory that Intel has in Leixlip (Ireland). These facilities began producing integrated avant -garde circuits In the Intel 4 node In September 2023, and they are also manufacturing semiconductors in the Intel 3 node. This plant is equipped with equipment of extreme ultraviolet lithography (UVE) produced by the Dutch company ASML, and has a fundamental role in Intel infrastructure to serve its European clients. Intel is paying for electricity in Ireland that in the US or Israel As we have seen in the first lines of this article, the cost derived from the tuning of a last niche plant is very high, but, in addition, Intel has just stumbled with a problem in his Leixlip factory: The price of electricity in Ireland. According to this US company, this cost amounts to 15 cents per kilovatio houra figure that is approximately double high as in the US or Israel, which are the venues of some of its most advanced semiconductor production plants. As we can intuit, the price of electricity paid a chip factory has a direct impact on the final cost of the integrated circuits it produces. And, therefore, also in its competitiveness. Intel has rushed to ensure that the future of the Leixlip plant is assured due to the very important role he plays in his infrastructure, but in Rte They argue that he is negotiating with the Irish government to find a way to reduce the cost of electricity that he is currently paying. Presumably it is likely that the State finally assume a part of the cost of energy. Image | Intel More information | Rte In Xataka | The US does not press only Nvidia and Asml against China. Prepare more restrictions for TSMC, Intel and Samsung

We knew that US Big Tech had a problem with the costs of their AI. DeepSeek has just shown to what extent

by

DeepSeek is the new darling of AI. This family of models, developed by a Chinese R&D laboratory of the same name, has achieved what seemed impossible: compete with the OpenAI or Meta models and do so, according to them, at a much lower cost. Is that true? A development 18 times cheaper than GPT-4. The Chinese startup released DeepSeek V3 671B at the end of December 2024. Its gigantic model was trained in just two months with a budget of 5.58 million dollars according to SCMP and analysts cited in Financial Times. Its performance is comparable to OpenAI’s GPT-4, but the latter cost about $100 million to develop according to Sam Altman. That’s almost 18 times more if we take into account both the data revealed by SCMP and Altman’s estimates. Comparative cost of the main chat and reasoning models today. DeepSeek’s price is incredibly lower than its competitors. Data: DeepSeek, OpenAI, Anthropic, Meta. Amazingly cheap. The cost of DeepSeek’s API is incredibly low when compared to its competitors. If we take the data from DeepSeek, Goal, OpenAI, Google and Anthropic It seems to be clear that the cost of using DeepSeek through its API is much lower than that proposed by its rivals. We have included the cost of GPT-4o mini which seems to be the only one comparable, but its performance is much lower than DeepSeek V3. DeepSeek V3 is superior to most of its competitors, although it is true that Meta has released for example Llama 3.3 in recent days and that comparison varies frequently. And it is (theoretically) superior to all. As they point out on RedditDeepSeek V3 prices are promotional: starting February 8 they will be $0.27 per million input tokens (almost double) and $1.10 per million output tokens (almost four times more) . This makes the comparison somewhat better for the competitors, especially for Llama, the only one that can compete in cost although the Chinese model is superior to that of Meta (and almost also to the rest in many metrics) according to the benchmarks carried out in DeepSeek. DeepSeek also “thinks” cheaper. The cost comparison is not only in favor of DeepSeek in the area of ​​traditional chatbots, but also in the area of ​​reasoning models. According to its internal benchmarks, the spectacular DeepSeek R1 It is significantly superior to OpenAI’s o1, but using the o1 API costs 27 times more than that of DeepSeek R1. Hallucinatory. Price drop in sight. As expert Ethan Mollick points out, the market will adjust to these DeepSeek-driven price drops fairly quickly. According to their estimates, the cost of a GPT-4 level AI was reduced 1000 times in 18 months, and a 95% drop in the price of the reasoning models, which right now are clearly higher than the AI models behind ChatGPT, for example. a chinese tsunami. The launch of the DeepSeek models is a great little revolution for all types of developers of AI-based solutions: they now have access to much cheaper models that are comparatively equal to or superior to those of the competition. This puts their rivals in a lot of trouble, and we will see how they react. Good news for users. The truth is that for us, the users, as well as for the developers, this is great news, especially because these prices make access to these functions incredibly cheaper. The market has been following this trend clearly, but DeepSeek has made the jump in cost reduction suddenly drastic. Image | Xataka with Freepik Pikasso In Xataka | OpenAI prepares a PhD-level AI. It is so promising that he will first show it to the US Government

Sweden has just found another damaged underwater cable in the Baltic. The fourth case already has a suspect

by

Last week there were two news that gave an idea that everything that surrounds the cuts and damage of underwater cables in the Baltic is far from ending. After an investigation carried out in the United States, it was slid that the cases in recent months They were not due to a Russian sabotage, but to inexperience and the deterioration. Hours later, United Kingdom put the focus on Russia After detecting a ship in its waters for the second time. Sweden too. A Sweden-Leonia cable. Apparently, A new underwater data cable, now between Sweden and Latvia, was damaged yesterday Sundaymarking the fourth recent incident in the Baltic Sea, one that generates concern in NATO about the vulnerability of critical infrastructure. Latvia’s prime minister, Evika Siliņa, has described the damage as significant and probably “caused by an external force”which has led to the opening of a criminal investigation. As we have explained, previous incidents They have been linked to Russian and Chinese ships. Ongoing research and measures. The Financial Times had a few hours ago That the systemal authorities already have a suspect. In fact, they sent a patrol boat to inspect a vessel near the cable at the time of damage, while two other ships are also being monitored. Swedish prime minister, Ulf Kristersson, confirmed the damage to at least one cable and said that Sweden is collaborating closely with Latvia and NATO to clarify the incident. In response to these attacks on critical infrastructure, NATO has deployed drones, submarines, ships and airplanes in the region Under the “Baltic Sentry” operationan action that aims to detect and prevent possible sabotage attempts. As we also count, United Kingdom has deployed its p8 posidon fighters. Background and suspicions. The situation, by repetitive, leads to the affected nations that there is something else. The Recent incidents in the Baltic They have fueled suspicions about possible acts of sabotage by Russian and Chinese vessels. Finland recently seized An oil tanker of the so -called “Shadow Flot” by Russiacomposed of old ships registered in remote jurisdictions such as the Cook Islands, which evade international sanctions. Finland Foreign Minister Elina Valtonen declared that It is unlikely that the damage will be accidentalalthough he avoided pointing directly to Russia. Not just that. In previous incidents, a Chinese ship, the Newnew Polar Bear, damaged a gas pipeline between Finland and Estonia by dragging your anchor, while Chinese load Yi Peng 3 crossed on two data cables In the same region in November. Subsequently, he remained a month in international waters between Denmark and Sweden, where he was approached by Chinese researchers, although access to Swedish inspectors was denied, generating criticism from Stockholm. Why are they so important. It We have underlined On more than one occasion, but now something else joins. This last incident occurs in a high voltage context, since Estonia, Latvia and Lithuania are preparing to disconnect in February from the Soviet Electricity Inherited and join the European continental system, a key step for its total integration in the West. The three countries, NATO members and the EU, consider this transition as His last separation from Russian influencewhich has generated fears of possible sabotage attempts to hinder the process. In this regard, Lithuania Defense Minister Kęstutis Budry A review of navigation standards in the Balticespecially when it comes to the use of anchors, underlining the most obvious: that the high number of recent incidents reduces the possibility of accidents. Repairs and contingency measures. Unlike gas pipelines and power lines, whose repair can take months, Fiber optic cables damaged in the Baltic are usually restored in weeks. In fact, the State Radio and Television Center in Latvia has assured that it has already found alternative routes to maintain communications in operation. Be that as it may, the repetition of these incidents, The room in the Baltic in such a short timeposes a significant strategic challenge for the region, one that has intensified its cooperation with NATO to strengthen the protection of its critical infrastructure. In fact, the organization has made it clear that any new sabotage action will be faced with firmer measures to guarantee the protection of communications and energy supply in the region. Image | BORN In Xataka | United Kingdom will display the “Frankenstein” of the electronic war to protect its submarine cables: the P8 Poseidon In Xataka | A British nuclear submarine has discovered a Russian ship in front of its submarine cables. The second time in three months

It is the greatest existential threat that Silicon Valley has faced

by

They have barely spent five days since its launch, but Deepseek R1 has caused An earthquake in markets (-6.5% NVIDIA, -3.5% Microsoft, -8% ASML) that goes beyond its technical efficiency. The true threat to the American industry is in its open nature: Depseek is democratizing technology that Silicon Valley has kept jealously stored after its own walls. Why is it important. The western industry of AI has been built on two pillars that Depseek has just dynamit: The need for multi -million dollar investments in hardware. The extreme secretism about the architecture of its models. The money trail. Wall Street is reacting hard because Depseek directly threatens the dominant business model that has taken us here. Investors ask questions: how to justify projects such as Stargate (and its half billion dollars) If there is an open and efficient alternative? What value intellectual property when your competitors openly publish their advances? Between the lines. The Chinese strategy seems clear: to use the open source as a Trojan horse to destabilize Western domain in AI. It is no coincidence that Depseek has chosen this moment, when: American models show signs of stagnation (where is it GPT-5?). Inversiones in infrastructure are triggered. Business models remain without clarifying. The big question. Several experts have pointed out that we are facing Sputnik moment For AI: when the United States discovered that it did not have the monopoly of advance in the space race. R1 has that magnitude, or do you have to relax expectations about its implications? Perhaps “only” (add all the quotes they want) we are rather at a Linux time: their arrival democratized the operating system starting from open source as a disruptive tool, but the industry continued to thrive. Windows did not disappear, he simply adapted to a new scenario in which open and closed source coexist. The difference is in the rhythm. Linux took years to seriously impact the market. R1 has caused immediate shaking even in stock assets of giants. Perhaps the market understands that AI is too valuable and promising to afford as closed and expensive development as the current one. Whether it is a Sputnik as if it is a Linux, R1 marks a before and after: it shows that the toe can be created more openly and efficiently. Yes, but. There are certain persistent doubts about the real costs of Deepseek. THE CEO OF SCALE AI He has suggested They have more hardware access than they say, but hide it for violating commercial restrictions. Even so, even if their costs were greater, R1 would continue to have a strong impact on the industry to choose the open model for such a powerful model. The time of truth. The American’s industry now has a dilemma: or maintain its current model, based on very high investments and a strong secrecy, risking being displaced by Chinese efficiency … or pivoting towards a greater opening to compete on the board raised by Deepseek . Deepen. The analogy with Sputnik beyond technological surprise. As then, this moment can catalyze a transformation in how technology develops and markets. The difference is that this time disruption does not come from the state apparatus but from the Chinese private sector, and its weapon is not the space race but the open source for AI. China, frequently criticized for its control over information, is using the openness and transparency to stand up to Western technological domain. In Xataka | Two years and 60 million users later: how Luzia has become the greatest success produced by Spain Outstanding image | Deepseek, Xataka with Mockuuuups Studio

What is it, how does this artificial intelligence have

by

Let’s explain What is Deepseekthe artificial intelligence chat created by the company of the same name. It is an AI that at the beginning of 2025 is calling a lot of attention, and that even in mobile stores of applications already managed to overcome Chatgpt. And not only that, but also in performance and reasoning tests with its model R1. In recent months we have talked a lot about this company and its models, especially because They are open sourceallowing it to be more accessible to everyone, and because it shows how much China is progressing in artificial intelligence technologies. What is Deepseek Deepseek is the name of a Chinese company specialized in artificial intelligence. It is also The name of your ia chatan alternative of Chatgpt, Gemini, Copilot and company. Therefore, you can listen and read Deepseek mentions referring to both. In this article we are going to focus on the artificial intelligence chatbot, which is a Great linguistic model o Large Language Model (LLM), created to help with software development, natural language processing and business automation. Deepseek stands out for open source. This means that anyone can see how it works internally, It is totally transparentand anyone can be installed locally or use it with total freedom. A developer or researcher can Download the github and modify it for various scenarios, including some commercials .. This totally contrasts with the secrecy of private models. There is usually the false perception that one of the things in favor of the private and opaque code of the majority of developers is that the quality of their products is superior. However, Depseek is proof that open source can match and even exceed these companies in some aspects. Another of Deepseek’s main characteristics is that its native chatbot, that of its official website, It is totally free and does not require any subscription to use its most advanced model. If we compare it with Chatgpt asking the same questions, Depseek can be a bit more concise when answering, it goes more to the grain. However, it has the same flexibility as other models, and you can ask you to explain things more widely or adapt them to you. Two models of AI as spearhead Deepseek has given much to talk at the end of 2024 and early 2025 for two specific models. First it is Deepseek v3a LLM long scale model that exceeds most of AI, and also surpasses some owners. In evidence such as programming, this model managed to overcome Llama 3.1 405B, A GPT-4O and QWEN 2.5 72B, although they all have much less parameters and that can influence performance and comparisons. But what has really surprised this model is that “only” needed 2,788 million hours of training, and this process cost only 5.5 million dollars. To give an example, with the OpenAI GPT-4 about 80 million dollars were invested. Come on, what cost much less to trainwhich makes it a very efficient model, and 60% larger that calls 3.1 405b, of finish line. The other great model has been Deepseek R1which specializes in reasoning, and has been able to achieve or overcome the performance of OpenAI’s most advanced models in key mathematics and programming tests. With him, it is the first time that an open and free Chinese model Equal Western leadersbreaking the monopoly of Silicon Valley. To put some figures, this R1 model has cost to develop between 90% and 95% less than its competitors, and has 671,000 million parameters. In addition, in mathematical tests Math-500 has achieved 97.3% hits, compared to 96.4% OpenAi. However, not everything is good. As with Gemini or Chatgpt, Deepseek It is also subject to the values ​​of the country where it has been developed. In this case it has been developed in China, and adapts to its regulation, and must “embody fundamental socialist values.” To give an example, this does not answer sensitive questions about politics and geopolitics, such as those about Tiananmen or Taiwan’s autonomy. In this regard, in China they are more ferrous when controlling the answers, there is greater state control. Therefore, while Western models can have more liebrtad to discuss controversial issues for the values ​​with which it has been trained, in Chinese models They must always align with ideology and narrative of the party that governs the country. How Deepseek works Deepseek has been developed using pure reinforcement learningwithout previously labeled data. This is a technical feat that until now was considered impossible, and that opens new doors to training this type of systems. This referez learning allows the model to be able to Learn for yourself by proof and errorjust like you can learn to ride by bike or perform some tasks. This makes the initial result more erratic and inaccurate, but the model itself discovers and develops unique reasoning strategies to continue improving. Once this training was done, Depseek refined the model combining it with other supervised training to polish it and create the final version of R1, which maintains this component but adding consistency and polished. Facing the user its operation is the same than that of other models. You have a home screen with a writing field and several operating options, and then when you see the results you have a left column with a history of all your chats. To use it you have to Write a question in natural languagejust as you would do it with a question. Then, Depseek will analyze the words of your question to find the result, and then look in the database with which he was trained or on the Internet to obtain data with which to compose an answer also in natural language. Thanks to the way it has been created, this model can understand complex contexts of very long and elaborate questions. In addition, you can also learn and improve. Its application can be used for free online or lowering your mobile app, and has no fee. What can you do with Depseek … 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.