point

3,500 missiles point to Taiwan

by

In the month of July he gave himself A situation which could pass perfectly through the stage of the filming of an action movie. In the early hours of a Monday, the stations of one meter began to fill, but instead of passengers loaded with wallets and suitcases, military, soldiers and more soldiers armed with anti -tanks. The place gave us an idea that this was not a movie, it was a simulation for what can happen: Taiwan. In fact, that possibility seems closer than ever from space. A huge display. I told this week The New York Times. Satellite images show that China is transforming its coast against Taiwan into a vast missile launch platform that constitutes the cornerstone of Xi Jinping strategy to force reunification and, at the same time, challenge US military power in Asia. He Pentagon estimates that the arsenal of the missile force, responsible for nuclear and conventional vectors, has increased by 50% in just four years until reaching about 3,500 units. Although the exact number of missiles located directly in front of the Strait is not known, the images leave no doubt: expanded bases, with tens of additional launch ramps and new facilities in key provinces such as Anhui or Jiangxi. New generation missiles. Among the displayed systems are the Dongfeng-17a hypersonic missile with great maneuvering capacity and difficult to interceptand the Dongfeng-26, known as The “Guam Express” due to its reach to US bases in the Pacific. These missiles can carry both conventional and nuclear heads and are transportable by road, which increases their mobility and complicates their tracking by enemy intelligence. The pentagon calculates that Beijing already has half a thousand df-26which makes this weapon a pillar of its strategy of denial of access against US forces in Guam, Japan or aircraft carrier deployed in the region. Exercises, deployments and messages. Practices on the Oriental China Coast They include simulated releases From agricultural fields, hidden or esplanade valleys next to highwayswhich demonstrates an effort to integrate missiles into the field in a flexible and dispersed way. The Times told that maneuvers have a double function: operational preparation and political signal. For Taiwan, the message is that resisting would be useless before an overwhelming arsenal; For Washington, that intervening would be too expensive. It is not, therefore, only military capacity: missiles They are the starting point of any Chinese coercion strategy and, in peace times, function as an instrument of intimidation through parades, tests and public exhibitions, such as The recent in Beijing in which missiles were Hypersonic and new ICBM. Expansion of Brigades 611 and 616. Brigade 611, in Anhui, ha Duplicate its size With training facilities, simulated tunnels and up to three dozen launch ramps, an unusually dense concentration that underlines the importance of the enclave. There, Xi Jinping It appeared Personally in 2024 to encourage troops to maintain a “crisis and combat mentality.” Further south, in Jiangxi, Brigade 616 is He has prepared To receive to DF-17with Adapted hangars to the dimensions of the new hypersonic missile. These deployments show how the Chinese oriental coast becomes a network of offensive nodes designed to saturate Taiwanese defenses and threaten United States facilities in minutes. Nuclear dimension and risks. The DF-26 Encarna the strategic ambiguity of Beijin, being able to equip yourself indistinctly With nuclear eyelets or conventional. Although American satellites could detect the transfer of nuclear heads from central deposits to these brigades, experts warn that the process It is not infallible And leave a very dangerous margin of uncertainty. A conflict around Taiwan would have, from the first moment, a Latent nuclear dimension. This ambiguity increases the risk of calculation errors and rapid climbing, especially if Washington decides to attack mobile pitchers in Chinese continental territory, which would mean a serious political and military decision. Drills and objectives. The Desert construction West of China of models of US warships and warships, some mounted on rails To simulate movement, confirms that Chinese missiles are not prepared only for static scenarios on land, but also to beat the US Navy In open sea. Chinese military plans plan to disperse mobile pitchers in caves and camouflaged locations, shoot and retreat, in a wear set to saturate and exhaust enemy antimisile defenses. Recent studies They conclude that the US air bases in Asia, many without sufficient reinforced shelters could be devastated in the first bars of a contest. Limits and vulnerabilities. The brilliant increase in capacities has not been exempt from problems. Scandals of corruption and purges internal have splashed to the missile force, and the Pentagon reports They alert that the quality of some nuclear infrastructure could be compromised. In addition, despite advances in radars and satellites, doubts persist about the efficacy of missiles in real combat conditions, especially against moving ships, where tactical uncertainty is much greater than against fixed objectives on land. The missile as a nucleus of power. If you want also, the accumulation of thousands of missiles on the coast Oriental reflects the conviction of Xi Jinping that the future of the Taiwanese issue goes through the coercion and threat of force use. That so -called as missile strength It is, in words of analyststhe “Crown Jewel” of the Popular Liberation Army: an instrument that expands the Chinese projection to the entire Western Pacific complicates any calculation of US intervention and places Taiwan under constant pressure. The “but” is that, the same strategy that seeks to ensure success, also contains the risk of a calculation error by precipitating a escalation nuclear unprecedented from the Cold War. Image | Maxar Technologies, Nara In Xataka | Taiwan has had an idea if Beijing invades her: surprise China underground In Xataka | While China debate about Taiwan, Europe does not waste time. Its greatest port has left a hole for war

Levante’s floods point out that we are not prepared before an increasingly dangerous Atlantic

by

Corted roads, trains delays, canceled flights rescue in three provinces. “The neighbors are panic“, said the mayor of Aldaia.” A few weeks after one year after the Dana tragedy, it seems that the phenomenon will be repeated, ” We could read In social networks. Gabrielle has been more than the remains of a hurricane, has been a reminder of all the pain of recent months and a promise: it will not be the last time. But are we prepared? A fact that seems curious, but it is something else. The 2025 hurricanes season has been very quiet, but something that has not happened for 90 years has happened. As Philip Klotzbach explainedwith “Humberto (…) the Atlantic would have a record of 3 of 3 hurricanes that became important in 2025 (Erin and Gabrielle were the others). The last time the first 3 hurricanes of a hurricane season in the Atlantic were important was in 1935”. But, in addition, none has touched earth (Gabrielle has already done it in the form of a postropical storm). That is, are the great Atlantic hurricane factory changing? Martín León has a good summary of the situation. The three cyclones “(1) have been formed from tropical waves of the east coming from Africa, (2) have moved through warm open waters, (3) have quickly intensified over very warm waters in the western Atlantic, (4) have recurred, or resort to the east to experience an extroatropical transition and (5) they will reach or reach the European coasts (transformed into the European coast postropical). It is true that it is early to draw conclusions. Until now (and despite the forecasts that it was going to be much stronger) “the 2025 hurricane station It is still close of the normal. “This was commented on by Martín León is curious, but is far from becoming a trend. And none of that changes the real problem. What problem? The current situation has taught us three things: the first one, a year ago, we were not prepared to support a blow like that of the Dana; The second, during this year, is that our institutional system does not seem capable of preparing quickly; And the third, these days, is that Valencia was not an isolated event, but a systemic risk in dozens of points of the country. Whether or not the change in tendency in the Atlantic, the situation is clear: climate change It exposes us to increasingly extreme meteorological phenomena. And our approach is the same as that of the last decades: nothing. But is this really new? Yes and no. As Emilio Rey explained to us“This type of phenomena has a period of recurrence of a certain time. Some occur every 20 years, another 50 or 100 years. But we know that it will happen again. It has always passed and will continue to happen in the future because our situation on the planet and the circumstances of this time of the year allow it. It will not pass every year but it will happen.” In any case, with climate change the frequency with which the strongest phenomena affects us will be modified. What the infrastructure of Castellón, TarragonaValencia or Saragossa There is much to do. Image | Via Stormyalert In Xataka | Google has demonstrated with its AI that the prediction of storms and hurricanes is outdated. This is how your new model works

Alibaba has presented its largest AI model, with a billion parameters. The question is whether at this point that means something

by

The Chinese giant Alibaba has announced a new language model, the largest they have announced to date. It is called Qwen-3-Max and presumes that it has more than 1 billion parameters. The biggest. It is the last model within the series Qwen3 which was launched in May of this year and, as its name ‘Max’ indicates, it is the largest to date. Its size is given by the parameters, 1 billion to be exact, while the previous models of its series reached a maximum of 235,000 million. According to South China Morning Post (Which owner Alibaba), his model stands out in understanding of language, reasoning and text generation. Benchmarks. The results of the benchmarks place QWen3-Max ahead of competitors such as Claude Opus 4, Deepseek v3.1 and Kimi K2. If Gemini 2.5 Pro or GPT-5 does not appear, it is because they are models of reasoning and have only compared rapid response models. As they point out in Dev.toboth Gemini 2.5 Pro and GPT-5 obtain higher scores in mathematics and code, so reasoning models continue to have advantage in those areas. Qwen3-max-preview can already be tested free of charge. Benchmarks shared by Alibaba. Parameters. The parameters are all the internal variables that a model learns during training. In other words, it is the knowledge that the model has obtained from the data with which it has trained and allows it to interpret our requests and generate their answers. In theory, the more parameters, the model will have more and better capabilities. It also implies that it needs more computational power both to train and to execute the model. More does not mean better. The speech of the parameters remembers that of the megapixels with the first cameras. A 100 megapixel sensor will take larger photos than a 10 sensor, but there are other crucial factors that affect image quality such as sensor size or lens luminosity. Quality data. More parameters can be translated into more learning capacity and more resolution of complex tasks, as long as quality training data has been used. It is obvious: a language model that has been trained with redundant, incorrect or biased data will learn and continue to reproduce those errors in their operation. There are more. In 2022, the laboratory Deepmind from Google, discovered that many models were oversized in parameters but underlined in data. To demonstrate it they created the Chinchilla model with “only” 70,000 million parameters, but four times more data. The result was that it beat Gopher, a model with four times more parameters. Architecture. The architecture of the model is another decisive factor in order to achieve an efficient model; A standard architecture is not the same that forces the model to use its entire neuronal network, than one like Mixture of experts which consists of many smaller networks. It would be something like having an expert committee each with a specialty. In this way, the model can choose your expert for each query and not have to use the entire network. For example, with this technique, Mistral manages to use only a fraction of his parameters And so it is faster and cheap to execute. Image | Markus Winkler, via Pexels In Xataka | The ASML-Mistral alliance reveals the European plan B: if we cannot manufacture chips, we will at least control how they are manufactured

We believed that the end of Windows 10 support would be a nightmare for Microsoft. There are those who point out that it will be a gold mine

by

Windows 10 He has already marked in red on October 14, 2025, day you will stop receiving updates If you don’t hire The ESU Plan. For Microsoft, what seemed like a problem is Also a business. An analysis firm estimates that the program could leave billions of dollars only in the corporate sector. Before the imminent end of Windows 10 support, the company insists that the best option is to update to Windows 11buy new equipment or use Windows 365 to access the cloud system. The change will affect both companies and individuals, who will have to decide how to continue. ESU: The proposal for Windows 10 and reinforces Microsoft accounts ESU is the official Microsoft offer for those who cannot leave Windows 10 in 2025. In exchange for an annual subscription, the teams receive only critical and important security updates. There is no standard technical support or new functions. The requirement is clear: to have version 22H2 installed. Microsoft describes it as a temporary continuity tool, not as a substitute for migration to Windows 11. The ESU business scheme is designed as a price ladder: $ 61 per device the first year, 122 the second and 244 the third, always with a three -year limit. Microsoft clarifies that The subscription is cumulative: Entering later does not reduce the cost. Access will be managed by volume licenses and activation keys will only be operational after October 14, 2025, the end of the free Windows 10 support. For private users, Microsoft has raised a different approach: a unique quota of 30 dollars to access security updates for 12 months, with alternative options for not paying. It will be possible to activate ESU by redeeming 1,000 Microsoft Rewards points or using the Windows Backup application to make a backupwhich will unlock access without cost. Microsoft indicates that it will be enabled soon. According to Windows Centralwhich cites a detailed analysis of Nexthink, Microsoft could enter until 7,300 million dollars Only in the business segment thanks to ESU. The consultant starts from the official Microsoft data, which in July 2025 estimated at more than 1.4 billion the PC World Park with Windows. Of that total, it is estimated that around 30% corresponds to public and private organizations, which is equivalent to about 420 million devices. The report also projects that, even with the impulse of Windows 11, about 121 million equipment will continue to run with Windows 10 after October 14, 2025. With the cumulative cost of the program, which doubles the price every year for a maximum of three periods, this user base would generate significant income. They are independent calculations that illustrate ESU’s economic potential, but Microsoft has not confirmed or commented on these figures. Nexthink figures draw a billionaire business, but reality is yet to be seen: we will have to wait to check how many companies and users pay for prolonging Windows 10. until you install Linuxchange Mac or move to a Chromebook. There is also the possibility of renewing equipment and moving on to Windows 11. Images | Windows | Arnav Singhal In Xataka | Send files among all my devices was a roll. Then I found this free application, Open Source and Multiplatform

We believed that at this point the PS5 and Xbox would be cheaper than ever. The hallucinating thing is that it is happening just the opposite

by

Sony has just announced that all models of their PlayStation 5 consoles will rise $ 50 in the United States. The argument that Sony has used in its statement To explain the decision is that there is a “difficult economic environment.” Or what is the same: they blame tariffs. But that is already a generalized trend … and very unreasonable. PS5 rises around the world. It is the same that happened in April in Europe, Middle East, Africa, Australia and New Zealand, when this company He also announced that prices went up for some models. In Europe, for example, the PS5 digital edition cost 499.99 euros. In the US, these increases affect all models, and the new prices will be activated as of Thursday in that country, being like this: PlayStation 5: 549.99 dollars (before $ 499.99) PlayStation 5 Digital Edition: 499.99 dollars (before $ 449.99) PlayStation 5 Pro: $ 749.99 (before $ 699.99) Prices for accessories, yes, remain unchanged. Microsoft already did the same. This movement is the same as Microsoft also with its Xbox Series S/X consoles, which They also rose in price at 50 euros In Europe and the US. Even the controls and some accessories saw their prices increased. And of course, Nintendo. The Japanese company took a long time to end up lowering the price of the original Nintendo Switch, but the Switch 2 has given it the excuse for uprise even more Even in the games. Not only that: the economic situation has caused Nintendo that Prices upload of the original Nintendo Switch, the Lite edition and the accessories. These data have not been met for the latest Sony and Microsoft models. In fact, the thing is totally the other way around. Damn tariffs. Most video game consoles are manufactured in China, and tariffs have especially affected the Asian giant. That certainly makes manufacturers more expensive to manufacture and export them to be able to sell them in the US or Europe. And if they want to keep their profit margins, they end up doing something logical for them and terrible for consumers, because we are the ones who end up paying the duck. But some manufacturing costs have dropped. The amazing thing about the situation is that both the PS5 and the Xbox Series S/X are now cheaper to manufacture (let’s not say the original switch). After five years in the market (nine in the case of the Nintendo console), the components of these consoles have dropped in price: their cost was then very high, but after all this time those components have ended up becoming almost obsolete compared to the new chips that have appeared in all this time. So what is happening? Not everything has come down. In fact, these consoles are based on GDDR6 memory chips and NAND chips for their solid state units, and there the prices They rose strongly in 2025 and 2025 for the relative shortage of this type of chips and the focus that manufacturers are putting in HBM memories for AI accelerators such as those of Nvidia. To this are added other costs not directly related to the material invoice (BOM), such as logistics, manufacturing energy, or marketing. The dollar loses bellows. But it is also the green ticket has lost value in front of the euro and other currencies. After 80 years being the pillar of the world economy, US tariffs They are starting to destroy it And that also translates into the price increases of everything. Including consoles. Less “boxes” are sold. Consoles manufacturers used to lose money with their models in the initial periods. The idea was always to recover those losses of plenty with the sale of physical games, but the rise of the sale of digital editions, the free-to-play games and the subscriptions have made the distribution of value change. For Sony or Microsoft the solution is simple: if we no longer recover so much value with the games, we upload the price of the hardware. But all that already happened before. The truth is that the situation is in many cases analogous to which it was lived with the PS4 and the Xbox One for example. These consoles ended up falling in a remarkable price over the years, but here a perfect storm has been created. One in which inflation, dollar weakness, tariffs and new video game access models have caused the worst result for users: more expensive consoles. In Xataka | 60 euros for a camera, 90 for a game. The worst (and best) of Nintendo prices for switch 2

GPT-5 will be better than its predecessors. The point is that for Openai that is no longer enough

by

It smells in the environment. GPT-5 is falling. Well, not only is he smells: it is also Sam Altman already shown that He is using it. There are many indications that point to an imminent launch, and it is therefore good time to assess what we expect from this new foundational model and its impact for the industry. Downward expectations. In general, the sensation of media and analysts coincides: the quality jump that we will see from GPT-4 to GPT-5 will not be much less that we saw between GPT-3 and GPT-4 when the latter He threw himself In March 2023. The reason is simple: climbing (dedicate more computation capacity and more training data) It no longer raises such notable improvements like those raised. Expected deceleration. That “Founder of AI“It was already expected – engineers knew that this better rhythm was not sustainable – and precisely the rescue came things like modes of reasoning. To make the model “reason” work: That allowed not to respond immediately, and instead the model explores several paths and ends up deciding the best. The problem? If “reasons” it takes longer to answer, and if it takes longer to answer conversing with it it is no longer useful. And that’s where GPT-5 enters. The off -road model. If we expect something from GPT-5, it is something like “the model to dominate them all.” So far there was a real chaos in OpenAi and its models: we have chatgpt as a unifying layer, but from this chatbot you can choose different models and different ways (reasoning, research, etc). GPT-5 is expected to automatically adapt to user’s needs (although being able to select modes manually does not seem a bad idea for advanced users) and becomes an off-road model that Unify everything. Better in everything, but maybe not too much. Of course, GPT-5 is expected to represent a sensitive leap in the capacities of this founding model. There are many rumors about what will be especially trained to program and to solve mathematical and even scientific problems. Advanced reasoning. It is also expected to have an advanced reasoning capacity, and here there may be two psible striking improvements. The first, that the model reasons better but also does it faster and more efficiently. Or that reason better (and slower) but reason as well as now, but at the same time that does not prevent chatting with it in that way. In both cases the result will be the same: more precise and detailed responses. Deep Think. The second, which perhaps Openai implements the same idea that Google has already implemented in Gemini: a Model “Deep Think” that combines the power of several instances of the same model so that they all reason and end up reaching the best conclusion. This second option, of course, raises the arrival of an even more expensive subscription plan, because this mode of reasoning is very expensive and using it intensively will be oriented only to very specific scenarios in areas such as the scientist or academic. Agentic. OpenAi already recently presented Your chatgpt agentand GPT-5 is expected to have a key participation in the future of that option. Right now, the agthic capacity of that service is surprising – it takes the control of the browser and you ask you to do something – but also limited (it takes a long time in everything, it is often wrong). With the new foundational model its performance should end up improving in a sensitive way too. A model of AI that threatens more jobs. THE IMPACT IN THE INDUSTRY It is certainly uncertain, but if these capabilities are really notable, there will be more threatened jobs. GPT-5 could, for example, being a strong candidate to even replace people in customer service and technical support: their ability to talk in real time-and to remember more and more details of users-is very striking in this regard. Sectors such as health, with improvement of diagnostic prelminar, or education adapting more and better to each student, such as They have already profiled with their mode of studyThey are clear candidates to be affected. After the failure of GPT-4.5, OpenAi needs a total success. A few months ago OpenAi suffered The greatest failure in its history with GPT-4.5a model that cost a lot to launch and did not show really striking improvements. He ended up going back and now he needs GPT-5 to represent a decent (at least) jump. Keep burning money. The role of GPT-5 is also key to maintaining the good reputation of the company. They live in much of it, and demonstrate the gigantic investment rounds that it has achieved over the last years. In Openai they continue to burn money as if there were no tomorrow, and although it already generates respectable income – it is estimated that will enter 12,000 million dollars In 2025— he is far from profitable. Fierce competition. The threat is also multiplied against rivals like Google and now also a goal. The first is Your most direct competitor (together with Anthropic) in terms of quality of commercial foundational models. The second has become a real toothache, especially since it has so much money (and generate both thanks to Instagram and Facebook) that does not need external investment and can waste money without apparent problems. For example, stealing talent without stopping. Please subscribe to chatgpt. Chatgpt’s popularity is undeniable, and today many users associate “AI” with “Chatgpt” when using any of the available models. That has allowed its number of weekly unique users either 700 million “Four times the one who was a year ago,” exceptional growth. Subscriptions, yes, are slower: at this time OpenAI has five million business business users, while the figure was three million in June. He needs many more to deal with his astronomical expenses, and GPT-5 should listen to get it. In Xataka | The AI war intensifies: Anthropic cuts Claude’s access to OpenAi. He has done it before the launch of GPT-5

How to create a Power Point presentation with artificial intelligence

by

Let’s explain How to create a Power Point presentation with artificial intelligence. In this way, you can create your slides quickly and easier, without having to have knowledge to design them manually. We are going to start the article with some previous tips that you must take into account before starting to create your slides. And then, we will tell you two or three ways to do it using different free tools. Before starting, some tips Before putting your PowerPoint slide designs, you must be clear of the presentation you are going to create. You will have to know the general idea of what you mean, the ideas you want to convey, and the tone you want to do with. Secondly, you must also Be clear about what you are going to put in each vignette that you generate for your presentation. It is not enough to know the objective and what you want to convey, you also have to order the ideas and plan the basic structure. Think of the classic introduction, development and conclusion schemes. An additional advice is that You can also use AI for planning. You can tell a generative AI as Chatgptalthough you can also use COPILOT, Deepseek, Gemini or similar that you want to make a presentation on a subject, and that helps you organize ideas on different slides telling you what you can put in each of them. It is also important Be clear about visual style What do you want to use. Think if you want a formal, creative, academic presentation or with a concrete personality, and then think of the colors or the type of images you want to use, so that you can then go to the AI in the creation process and that each vignette is not completely different from the previous one. Make the slides in chatgpt One of the things you can do is compose the presentation through images created with chatgpt or others capable of generating images. For example, you can use a Prompt like this: “I want you to create a PowerPoint slide. In it, the presentation title has to appear, which is” Xataka Basics “. Below, as a subtitle put” How to make your PowerPoint presentation with AI. “The slide must have the image of the face of a robot to one of the sides, and it must be the size of the PowerPoint slides.” Here, as you see, it is important to say what you want it to appear in the image. It is also important Specify size and dimensions. We, not to complicate much we have said that you use the size of the PowerPoint slides, so that I looked for and knows the exact dimensions. Although we have not included it, too You can mention a visual style with a description that you can then repeat in the other slides. Or if not, in the same AI chat that you are using asks to use the same style as in the previous slide to take into account. And finally, in the case of Chatgpt You can ask you to offer it in PowerPoint format. This will make the AI pass this slide to the PPTX format, which is the Microsoft tool officer. Copilot and Gemini can give you additional help Copilot is from Microsoft, the same PowerPoint company, and Google has another alternative called Slides, and Gemini’s AI. This means that They will give you extra facilities to create presentations or slides for these own tools. Even if you don’t know what to do, you can start with say you want to create a template or slide PowerPoint in Gemini, for example. Then, AI will guide you for the process and ask you questions to design everything to your liking. Make the slides with gamma Gamma is An application specialized in creating powerpoints using artificial intelligence. Your website is Gamma.appand after registering you will go to a creation screen where you should Choose between three creation options. With a free account you can make a presentation of up to 10 modules, and APRA make up to 60 you will have to go to the payment plan. If you choose the option of Paste textyou can create your presentation from a note, a scheme, or text content that you want to use. First hit the text, and then choose what you want to do. You can generate vignettes or loose ideas from notes or schemes, summarize the extensive text that you have put in several vignettes, or keep the text as it is and distribute it in vignettes. After configuring that, in the section of Paste textyou will go to A second screen to edit instructions. Here you can choose the subject and language, also the tone to be used. You can also choose the amount of text per vignette and many other elements to configure everything to your liking. If you choose the option Triggereverything is simplified to the maximum. First you have to choose the amount of modules, size and language, and then Describe the idea you want. When you have all that click on Generate scheme. Now, you will go to A second step with a scheme of what each of the slides will include. This scheme will be able to edit to your liking. When you have it click on the button Trigger that appears below, and then the slides will be generated. And if you choose the option of Import file or URLthen you will only have to choose a PowerPoint, Word, PDF or slides file, or directly a web page where you point to information, and will be used to create the slides from this material. Slides with Canva and its AI options Canva is possibly the simplest solution To make all kinds of designs, including presentations. The first thing you have to do is enter Canva.comand on the main page choose the option to create a presentation that will appear below. This will create … Read more

1,800 years ago China had one of the most intriguing technologies. Now is the point of resuscitating the dragon that detects earthquakes

by

Almost two thousand years ago, during the Han dynastythe Chinese scholar ZHANG HENG He designed a device that, according to historical chronicles, could detect distant earthquakes and even point out his address. That invention, called HouFeng Didong Yiit was a mechanical gadget that had the dragon as the main actor. Now China is about to resurrect what a legend was believed. A erased prodigy of time. We talk about a mechanical system (an ornate vessel surrounded by eight dragons with suspended bronze balls, oriented towards mouths of toads) that would have been able to register imperceptible earthquakes in Luoyang, the imperial capital, with a precision that “touched the divine”, according to THE BOOK OF THE Subsections. However, its sudden disappearance of historical records and the impossibility of replicating it exactly led to its total elimination of the Chinese educational curriculum in 2017, relegating it to the field of legend. Today, a team led by Professor Xu Guodongfrom the Hebei Disaster Prevention Institute, seeks to recover not only its operation, but also Your place in history of science. The rebirth of a prodigious machine. To the question: How the hell will they replicate it? The researchers They explain that from ancient literary fragments and principles of modern structural dynamics. Thus, Xu and his team have proposed a functional model of the earthquakescomposed of three key subsystems: structure of excitementtransmission and closure. In the heart of the device he was A “Capital Pilar” that should not be interpreted as an unstable column, but as a pendulum -like arm (a kind of gigantic stick anchored to the ground) that amplified the seismic vibrations. With just 1 mm of displacement at the base, the tip of the pendulum moved up to five times more, activating a “L” levers system that released a ball in the mouth of the toad corresponding to the direction of the epicenter. A blocking mechanism prevented other dragons from reacting, thus respecting the Original description of “a dragon who speaks and seven that shut up.” Mathematical Wisdom. Team simulations indicate that the system responded reliably to displacements just 0.5 mmwithout emitting false alarms. Although modern knowledge of propagation of seismic waves suggest that a single instrument cannot determine with total precision the direction of the epicenter, Xu argues that the Historical records coincide with optimal geological alignments. As proof, appointment The Longxi earthquake From the year 138 AD, when the instrument would have detected a tremor 850 kilometers away, without feeling in Luoyang. The initial skepticism of the officials vanished when messengers on horseback confirmed The shaking days later. Even more revealing, he explains, is the jump in the frequency of earthquakes registered in the capital after the implementation of the artifact: in the previous 8 years they were only documented three local earthquakes; In the 58 years later, there were 23, in a region considered low seismicity. Zhang HENG: astronomer and political victim. ZHANG HENG It was not any inventor. His appointment as a great imperial astrologer in 115 (a position equivalent to the director of a modern national observatory) and his creation of an armilar sphere Able to map the sky with precision, they prove their domain in mathematics, astronomy and mechanics. But his invention could have been politically uncomfortable. In a context where natural disasters were interpreted as signals from heaven and threats to the mandate of the emperor, an instrument that “predicts” earthquakes could have been seen as subversive. Some scholars suggest that HENG’s abrupt retirement in 138 and his death the following year They were not casual. Xu duck that the loss of the original earthquakes (along with its technical diagrams) could be due to wars, political chaos or even the greed of powerful aristocratic families that would have hidden its existence. Recovered legacy. In a gesture loaded with symbolism, xu He recalled that only two bronze objects have been defined in Chinese history: Nine cauldrons of the Xia dynasty and this earthquakes. Now, the objective is ambitious: rebuilding the instrument using only materials and technical Advanced knowledge reached in ancient China. Beyond the material restoration, the project aspires to reinsert this jewel of engineering in the global narrative of science, as proof that humanity had already tried, long before satellites or artificial intelligence, unravel the Mysteries of the tremor From Earth. Along the way, the Heng’s featerased for centuries of oblivion, it could be closer to recovering its place among the great milestones of human thought. Image | KowlooneseSSPL In Xataka | The most revolutionary and rare writing machine was lost in 1940. Until someone received a message In Xataka | Of the Wright brothers to Wu Zhongyuan, a homemade helicopter

Half of all the oil we need passes through two runners, and one is right now the hottest point of the planet

by

We can think that overall supply chain It is solid due to the number of transport options. However, commercial networks in huge airplanes, the rail transport that China wants to boost With his new ‘Silk routeor the road transport They remain in the background if we compare it with The reliable old man: the ship. And, when we talk about oil transport worldwide, this is something that is much more evidenced. This graph prepared by Visual Capitalist It reflects it perfectly. The data. Talking about oil is talking about certain countries that brings together the bulk of reserves. However, not always the most oil is the one who produces the most (to tell Venzuela), and world trade in crude passes through a few points between Asia and the Middle East. If these funnels have any problem, the entire world will also have it because the estimated amount that moves every day is an absolute barbarity: Millions of barrels a day in 2023 Strait of Malaca 23.7 Ormuz narrow 20.9 Suez Canal 8.8 Bab el-Mandeb 8.6 Cape of Good Hope 6 Strait of Denmark 4.9 Strait of Türkiye 3.4 Panama Canal 2.1 Malaca is the big funnel. With 23.7 million barrels per day moved in 2023, you have to talk about the first name of the list: the Strait of Malaca. It is a narrow complex due to its depth and wide, but it has become the main oil transport channel in the Middle East to Japan, South Korea and, above all, China. The Asian giant is the main oil importer (Although the batteries are being put to become one of the main producers) and it is estimated that 25% of the maritime transport of crude passes through this corridor. Its location between the Peninsula of Malaysia and Sumatra makes it the shortest and, therefore economic road between the Indian and the Pacific. And is responsible for Singapore is the largest gas station in the ocean. Ormuz, the hot spot. If Malaca’s is vital for the crude that goes to Asia, that of Ormuz It is essential for the one from the great producers such as Saudi Arabia, Iraq, Arab Emirates, Kuwait or will go to the rest of the West (although a lot of production also goes to Asia). The Strait has about 167 km long and about 40 wide, but navigation routes are much narrower. Now, its peculiarity is that it has sufficient depth to allow the transit of the great oil tankers and entrepreneurships. The problem is its location. This is the only significant maritime exit (for trade) that the Persian Gulf has and, like Malaca’s, it is a strategic point that Recently it has been in the foreground due to the War between Iran, Israel and the United States actions. Fragile balance. Being the 20% output point of the oil worldwide, and also moving much of the Liquefied natural gas which is consumed in the world, a block would have deep effects on the economy and global energy dependence. The estimate is that about 20 million barrels are moved daily and, if that flow is compromised, prices HE They would shoot at levels similar to the right views after Start of the Ukraine War or in the 70’s oil crisis. In addition, other points such as Suez or Malaca They would receive more ship influxwhich would carry overloads, shipping delays and would not serve to relax the barrel price increase. And the most important thing that lets us see this map is that, although there are several critical points for oil, two of them bring together almost half of the crude oil movement of the entire planet, which implies a balance that the Geopolitical tensions They can destabilize easily. Saudi Arabia and Arab Emirates are promoting the creation of pipelines to diversify the supply routes, and in the aforementioned ‘New Silk route’ are also raised alternatives to reduce the dependence of the narrows in the oil that moves to China. But, for the moment, Malaca and Ormuz are the hottest points of oil worldwide. In Xataka | The world capital of rare earth is being made of gold thanks to them. And it is also poisoning

Deepseek marked a turning point in the AI race. Now another Chinese company wants to imitate its success: Kimi K2 is born

by

The Chinese startup Monshot AI has presented Kimi K2, an open -source artificial intelligence model that arrives with outstanding programming capabilities and autonomous tasks that, according to The published benchmarksThey spray competition in several of their models. Its launch occurs at a key moment for the sector, when Chinese companies seek to replicate the disruptive success of Deepseek with potential height models and much cheaper than market alternatives. Kimi does not come from nothing. MoNshot ai was one of the most promising startups in the Chinese ecosystem of AI and that giants like Alibaba have invested greatly. His Kimi chatbot reached third place in monthly active users in August 2024, but fell to the seventh in June After the emergence of Deepseek R1 in January. Now try to recover ground with a strategy that combines open source and aggressive prices, following the formula that catapulted Deepseek. Image: MoNshot AI What Kimi K2 offers. The model has 1 billion total parameters and 32,000 million activated parameters, using The well-known Mixture-Of-Experts architecture to optimize computational costs. It is presented in two versions: a base for researchers and developers, and another optimized for conversation and autonomous tasks. Kimi K2 thus becomes Moonshot AI’s proposal with the ability to act as an intelligent agent to use tools, write code, complete workflows or talk, among other tasks. Kimi K2 explained in numbers. In performance testsKimi K2 has achieved 65.8% precision at Swe-Bench Verified, one of the most demanding benchmarks for software engineering. In LivecodeBench it reached 53.7%, exceeding 46.9% of Deepseek-V3 and 44.7% of GPT-4.1. In mathematics, its 97.4% score in Math-500 exceeds 92.4% of GPT-4.1, suggesting significant advances in mathematical reasoning. The price factor. MoNshot is charging $ 0.15 per million input tokens and $ 2.50 per million tokens out of the developers who use their API. Compared, Claude Opus 4 It charges 100 times more for the entrance (15 dollars) and 30 times more for the output ($ 75), while GPT-4.1 charges 2 dollars per entrance and 8 per exit. In addition, the model is available for free in Web applications and Kimi mobile, without monthly subscriptions that require chatgpt or Claude for their most advanced models. Technical innovation. MoNshot has developed the MuCanclip optimizer, which allows train models of one billion parameters “With zero training instability.” This technology could drastically reduce the training costs of large models, a problem that has limited the development of AI to companies with greater resources. Double channel strategy. The company offers so much Free access to the source code as payment API at a very competitive price. This strategy allows companies to start with the API for immediate implementation and then migrate to self -healing versions either by regulatory cost or compliance. And it is that each developer who downloads Kimi K2 becomes a potential business client. Moment of inflection. Kimi K2 represents a convergence point where open source models and proprietary alternatives shake hands. MoNshot AI intends to turn Kimi into a tool for everything, while offering its open source model and is reserved to charge for the use of its API for all types of implementations. And now what. The launch reaches a critical point in which both Openai, such as Google or Anthropic, must respond to this wave of cheap and high quality language models. The issue is no longer whether open source models can match the owners, but if large technological ones can adapt their business models fast enough to compete in this new scenario. The looks are put in GPT-5 And in the next movements of the industry at a rate, as always, accelerated. Cover image | Xataka with Mockuuuups Studio and Kimi AI In Xataka | Grok 4 destroys the tests and aims to be the most advanced AI model. The problem is that Elon Musk continues to sabotage his answers

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.