the Great Trade Route through the Mongolian Desert

At the end of the 19th century, as Japan emerged as an imperial power after the Meiji erahis army undertook an ambitious cartographic project to precisely know the territories beyond its borders. Those maps, prepared by the Imperial Japanese Army with methods that combined espionage, foreign sources and field work, were classified as state secrets and for decades remained hidden in military and university archives. Today, those maps have revealed a fascinating route. A forgotten runner. For centuries, the so-called Great Mongolian Route was a key artery of Eurasian tradean east-west route that crossed southern Mongolia connecting northern China with Central Asia and beyond, serving as a northern alternative to the better-known routes of the Silk Road. Despite its historical importance, it had been blurred between travelers’ stories and scattered references, without precise cartography that would allow it to be reconstructed in detail. That gap is what is now filled through a historical work, a published study in the Journal of Historical Geography by Chris McCarthy and his colleagues demonstrating for the first time that the Great Mongolian Route was not a literary abstraction, but a perfectly structured corridor, designed to enable the regular transit of camel caravans through some of the most arid and hostile landscapes on the continent. Military maps as secrets of the past. The researchers behind the discovery say that the key to rediscovering the Great Mongolian Route has been in the gaihōzuthose maps prepared by cartographers of the Imperial Japanese Army between the end of the 19th century and the Second World War, which systematically covered vast regions of Eastern and Inner Asia. Conceived with strategic and classified purposes For decades, many were on the verge of disappearing after the war (there were instructions to destroy them), but some were saved quietly and ended up in university archives that little by little became accessible to the public. Gaihōzu W6N2N map panel: Explaining the maps. The maps were not simple military sketches: they synthesized information from Chinese records, ancient Russian uprisings, and, in some cases, Japanese field work, resulting in a surprisingly accurate representation of routes, wells, monasteries, oases and geographical features key to survival in the Gobi Desert. Owen Lattimore’s map of several Inner Asian caravan routes, including the Great Mongolian Trail, whose name appears next to the location of Gurbun Saikhan Confirm the map on the ground. Recent work has gone beyond the archive, touring more than 1,200 kilometers on the ground to verify to what extent those sheets coincided with current reality. The verification has confirmed about fifty nodes (from water sources to settlements, caves and sacred places) spaced at intervals of about 24 kilometers, a distance that fits exactly with the average day of a camel caravan. Plus: the oral traditions of local shepherds, the physical traces of secular transit and the persistence of toponyms have reinforced the idea that these maps captured a refined logistical system, one in which each stop was essential to making the journey possible. Gaihōzu W9N2N map panel Caravans, tea and benefits. Although the main objective has been to document the infrastructure of the route, everything indicates that was parte of the historic tea trade, with Chinese goods traveling west and steppe products returning east. Inscriptions found in caves and oases speak of journeys of up to 120 days for heavy caravans and faster journeys, of about 90 days, for urgent transports seeking extraordinary benefits. The harshness of the route did not deter the merchants, moved by the promise of “triple benefits,” a reminder that these routes were not only avenues of cultural exchange, but high-risk economic gambles. From stories to cartography. For decades, knowledge of the Great Mongolian Route depended almost exclusively on the descriptions of the explorer and scholar Owen Lattimorewhose diagrams offered a conceptual vision of the corridor. Now, the combination of his stories with the millimeter detail of the gaihōzu transforms that diffuse image into a concrete and verifiable layout, where each lake, well or monastery has a clear function. The result not only recovers a lost route, but shows to what extent these military maps constitute an exceptional file of landscapes, economies and ways of life just before modern transportation erased centuries of mobility caravan in the interior of Asia. Image | McCarthy et al. 2026 In Xataka | The entire history of the Iberian Peninsula year by year, summarized in six minutes of interactive map In Xataka | Our conception of the world has changed a lot during history. This map illustrates all its forms

AI is starting to change that dynamic

For years, Apple was more than just a mobile phone manufacturer: it was the customer that everyone wanted to keep happy. The company that could negotiate with suppliers and reserve capacity in advance. But that stage is beginning to break down for a very specific reason: the industry has started buying hardware for artificial intelligence on an enormous scale, and that new appetite is reordering priorities. AI companies are willing to pay more and secure supplies up front, a shift that is beginning to put pressure on something Apple has protected like a treasure: its margins. Memory, the bottleneck. The easiest example to understand is in something as everyday as the storage and speed of the iPhone. It’s no secret that memory chips are in short supply due to the explosion of AI, and that is pushing prices up. Tim Cook dropped it in the last earnings call acknowledging limitations in chip supply and that memory prices were rising “significantly.” Not so comfortable terrain. The Wall Street Journal points out that AI giants are willing to close agreements with very attractive conditions for suppliers, including the possibility of securing supply with firm commitments and advance payments. This context gives room for companies like Samsung Electronics and SK Hynix to raise prices on certain DRAM chips destined for Apple. Even on the less visible plane there is friction: many engineers who previously worked on improving displays for smartphones now also spend time on specialized glass for packaging advanced AI chips. It’s a silent fight for capacity and attention. Alternative to TSMC. The newspaper says that TSMC is doing more business with NVIDIA and other AI companies. Consequently, according to anonymous sources with knowledge of the plans, Apple would be exploring the option of manufacturing some less advanced processors with another supplier. No name has yet been released, but it would be a fairly important change in the Cupertino company’s supply chain. How it affects us. In the short term, the blow is taken by the income statement: if components rise, margins suffer, even in a company used to working with ease. For the consumer the scenario is more ambiguous. The well-known analyst Ming-chi Kuo estimates that Apple would not expect to raise the price of the next iPhone if they are equipped in a similar way to the iPhone 17. That doesn’t take the pressure off, but it suggests that adjustment could come through other avenues, from configurations to tighter margins. Images | Apple In Xataka | The AI ​​bill for Meta has grown by 400% since 2023: Zuckerberg wants to lead the sector at any cost in 2026

It is the key step in the “rebirth” of a company on the tightrope

The United States is immersed in the war for technological independence. They aim to be sovereign and produce most of the key components of their technology, but they remain so dependent on the rare earth from china like from Taiwan for advanced chips. At the same time, the strongest Taiwanese company, is planting its flag on North American lands with TSMC. In this scenario, Intel has become the great hope of the American foundry. And they just announced that their Intel 18A plan is ready for action. Self-imposed goal. Intel has been going through the deepest crisis in its history for years. Unlike NVIDIA, Qualcomm or AMD, which design chips but are manufactured by others –TSMC mostly-, Intel designs and manufactures semiconductors (although it also outsources part of its production). It is, as it is known in the industry, a foundry, and after being devastated in the mid-2000s, they have seen how his rivals ate his toast. Both in semiconductor production and in their design and in the market. For this reason, in 2021 they set a goal: to develop five nodes in four years. This strategy, baptized as 5N4Ywas a move to restore the company’s position at the forefront of semiconductor manufacturing. In between, they have acquired ASML’s latest generation machines, they have positioned themselves as the local foundry for the US to achieve technological sovereignty… and they have needed an unprecedented injection of money from the US government. Intel 18A. But well, the plans seem to be coming out and, if in September 2024 Ben Sell, vice president of technology development at Intel, commented that Large-scale production of node 18A would begin in 2025now it is the company that has declared that they are ready to start mass manufacturing products based on that technology. The result is two processors with a very different approach. Panther Lake – It is the architecture of the Intel Core Ultra Series 3the first System on Chip from the American company created with this photolithography. They are chips created in a smaller size, which allows for greater density (30% according to Intel) and 15% greater performance per watt. It is focused on mobile devices and integrates both CPU and GPU. Clearwater Forest – It is the other current leg of Intel 18A, a processor for data centers hyperscale, cloud computing and AI training. It will be the heart of the Xeon 6+ processors and shares those characteristics of higher density, more performance and lower consumption. Technological avant-garde. Things seem to be starting to go well at Intel and the interesting thing is not that they are already preparing for large-scale manufacturing of these processors, but rather the technologies which, for the first time in a long time, will allow Intel to be at the forefront in its sector. RibbonFET – It is Intel’s first new architecture in more than a decade and is what allows improved performance per watt compared to the previous node used by Intel. It is an improvement over the Classic FinFETs. PowerVia – It is the true revolution: it is an architecture that separates the power supply from the processor lines to deliver it through the rear. Power flow is improved and delivery is optimized, allowing better power flow that increases the processor clock frequency while consuming less power. American foundry. This last technology is a pioneer in the sector and, in fact, it is expected that its main competitor, TSMC, will not have a response until the end of 2026 and Samsung its GAA in 2027. And that is precisely what is positioning Intel as a good option for the big technology whales. Because there is no point in having technology if you don’t attract attention, and here Intel has an advantage. Fab 52, the Intel plant in Arizona where these new chips are manufactured On the one hand, and obviously, technology. But on the other hand, and just as important, being an American manufacturing in the United States, with what this implies when it comes to get government favors. I know point that NVIDIA and Apple are in conversations with Intel to have a certain range of their GPUs and SoCs manufactured by them instead of TSMC. Because that commercial success is the last frontier and attracting whales is what will mark Intel’s destiny in a war that is no longer just about having the best technology, but also about where you manufacture it. Images | Intel In Xataka | The world’s technology industry practically depends on a single road: the one that leads to the Spruce Prine mine

SpaceX, founded by Elon Musk, has just announced the purchase of xAI, founded by Elon Musk

Until very recently, this was just a rumor. Today, SpaceX just told it as a fact. The aerospace company has published an official statement in which it states that it has acquired xAI, the artificial intelligence company founded by Elon Musk. The text does not go into details of the operation, but it does set the tone: it talks about integrating AI, rockets and space connectivity as part of the same strategy. And, although the announcement is forceful, it leaves many important questions in the air that still have no answers. SpaceX frames the operation as part of a vertical integration. The official statement is signed by Musk and has a more ideological than corporate tone, with references to “freedom of expression” and an almost existential mission. But, beyond the story, the document leaves out some elements that are important to understand this movement: there are no figures or details of how the agreement materializes. In development. In Xataka | Genie 3 is awesome at creating worlds for video games. But the problem with video games was never creating worlds

“We are not going to launch a flagship a year just for the sake of it”

If there is an immovable rule in the technological world, it is that every year a new generation of any product must be launched. If not more. The RAM market may explodebut what is certain is that every year we will have a new samsung galaxy and a new iPhone. And to the question of whether it is necessary to have a new high end mobile Every year, someone has answered that, perhaps, it is not necessary. It has been Nothing, and curiously it could do with a high-end. But… it makes sense. In short. Carl Pei is not only one of the founders of OnePlus: is the mind behind the launch of the Nothing brand. After landing in 2022the British company has relied on different marketing, but also on a CEO who is very active in networks, as well as open about the future of the company. Faced with the opacity of colleagues/rivals, Pei has always been quite ‘playful’ with the opinions of both the industry and his brand and the segment in general. In a recent self interview published on his YouTube channel, he has given an interesting key. “We’re not going to launch a new flagship every year just for the sake of it.” There are two melons here: one is that we won’t have the Nothing Phone (4) in 2026. The other is that you are quite right considering how things are. Rampant crisis in the background. Although companies like Micron, Samsung either NVIDIA It is coming in handy, we have been immersed in a -new- components crisis for weeks. RAM was the first product whose price turned this basic component into a luxury one, but the graphics cards and SSDs have followed the same path. And things do not look like they will improve in the short and medium term. This RAM crisis has already resonated in the smartphone segment. There are two options: either -much- more expensive mobile phones or mobile phones with -much- less RAM. Goodbye to the crazy 24 GB of memory on a mobile, welcome 4 GB. Pei himself already commented: if things continued like this, and they have remained the same, the user will have to choose between pay 30% more for a new mobile or settle for a new mobile phone at the same price as the previous one, but with less RAM (and the storage we would see). Come on, Pei has said, without saying it directly, what the decision would be with a Nothing Phone (4). Software > Hardware. Now, as they point out from Xataka Mobilethe fact that there is no new Nothing ‘flagship’ for 2026 does not mean that they are not going to launch a mobile phone. It is estimated that they are working on a mid-range Phone (4a) that will take up the baton of the notable Nothing Phone (3a) at the same time they keep the components short so that the price does not go down. And, furthermore, it gives meaning to an industry strategy, one in which the greatest advances that we have seen in recent generations have more to do with cloud services, artificial intelligence and everything that software encompasses… more than hardware. Yes, more powerful, faster and capable hardware is important to perform tasks within the device, but the cloud is also a pillar in this software. Qualcomm pushes another narrative. On the other hand, there is the strategy of some hardware companies that are obliged to keep the wheel turning. NVIDIA and Qualcomm are two examples, with more capable graphics cards not so much in terms of raw power performance, but rather better processing of artificial intelligence tasks such as DLSS. And also Qualcomm, which every year-end presents its new chips for mobile devicesand those are the ones we see in the new launches of the most premium range. Because each new generation is more powerful than the previous one and -may- have better cameras and more generous batteriesbut it is also true that from year to year we are not seeing considerable jumps between devices from the same brand. And that is when it makes complete sense for a company like Nothing to point out that, perhaps, this annualization of the ‘flagship’ is not necessary. It would be necessary to see what would happen in a context other than a component crisis, of course, but Pei himself has said on occasion that software is the future. Image | Xataka In Xataka | I had no idea what the future of the smartphone is. Until I spoke to Carl Pei

Microsoft has finally realized what the community has been shouting at it for months: we don’t want so much AI

The people are fed up with the avalanche of AI that has flooded Windows and Microsoft turned a deaf ear to the numerous community complaints. They have put AI even in Notepadwhich is saying something. Microsoft’s obsession has caused the Windows image to sufferbut finally it seems that they are listening to the users. We are still passing. One of the things Microsoft has been doing in its pro-AI crusade is add Copilot buttons everywhere. It’s in Paint, in Notepad and even they want to put it in file explorer. Although Microsoft has not commented, according to Windows Central fontsthe company is rethinking its AI strategy and one of the things that is under review are these buttons that they have been adding almost indiscriminately. Maybe they end up eliminating some or just being more selective from now on. Windows Recall. “It’s like having a photographic memory.” This is how Microsoft sold what aimed to be the PC+Copilot star feature. What followed were many doubts about your safety and so many criticisms that Microsoft had to delay the project for more than a year. Recall is already implemented, but according to Windows Central the company is not satisfied with how it is working and wants to correct course. How they will do it at the moment is unknown. There will still be AI. Microsoft still has a lot of AI features in the works and nothing indicates that they will stop, so if you were rubbing your hands at the idea of ​​a Windows 11 without AI, that is not the case. Some of the initiatives they have underway are: agentic functions which they announced in November of last year (to which The community flatly refused.by the way) and developer features like Windows ML or semantic search. The complaints have been heard. There will probably still be more AI features than the community would like, but it seems that Microsoft has heard the feedback and they are going to take their foot off the accelerator. The obsession with AI has not been the only reason for discontent, there have also been highly criticized decisions such as force to use an online account to upgrade to Windows 11 or the stability problems after updating. Despite everything, Windows 11 is advancing unstoppably and It is already on more than 1 billion devices. Image | Microsoft, edited In Xataka | I have decided to become independent from all US technology and embrace European technology. This is how I’m getting it

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

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

the problem is how you want to do it

Mexico is about to make a profound change in the organization of its work week, reducing the maximum working day from the current 48 hours. 40 hours per week The proposal presented by President Sheinbaum is to bring the working day closer to OECD standards. Currently, Mexico is one of the countries with the longest working day in the world. The problem is that the path to reaching those 40 hours a week is not going to be as simple as it was proposed. in the first proposals and the result is a reform that is born with many nuances, with a calendar that extends until 2030 and that already generates doubts among businessmen and accusations of “simulation” from the opposition. How the working day will be reduced. The opinion circulating in the Constitutional Points and Legislative Studies commissions that is being studied for its approval in the Senateestablishes a progressive reduction of the working day at a rate of two hours less each year until 2030. That is to say: 2026 will end with a 48-hour work week. In 2027 it will be reduced to 46 hours. In 2028 it will be reduced to 44 hours. In 2029 it will go to 42 hours. In 2030 the goal of 40 hours per week applies. This design seeks to give companies room to adapt so that salaries are not affected and productivity adapts progressively. Congress will have 90 days to adjust the Federal Labor Law and align it with the new 40-hour limit, which includes redefining new regulations to regulate overtime and working hours control mechanisms. A single day of rest. One of the first changes that have occurred with respect to President Sheinbaum’s first proposal is that the obligation of two days of rest will not be applied, but rather the current text of the article 123 of the Constitutionwhich states that for every six days of work there must be “at least one day of rest.” In practice, this means that even with a 40-hour week, the ruling does not explicitly establish a work schedule with Saturdays and Sundays off, but rather preserves the minimum of a single day of rest with full salary. Eight hours that will be extra. On the other hand, in the modifications with which the Senate is working, it is considered that the eight hours that are going to be cut from the ordinary day may be covered as overtime. The new framework states that overtime may not exceed 12 hours per week, compared to the current eight, and may be distributed in up to four hours a day for a maximum of four days. Compared to the three hours a day with a limit of three consecutive days that regulates the current Federal Labor Law. That is, the working day is reduced, but the limits for overtime are increased. Of course, when this limit is exceeded, each extra hour must be paid at 200% of the usual salary. Controversy over the hours. This extended extra work design has triggered criticism from the opposition and some specialists, who consider that the reduction in working hours is distorted if a relevant part of those hours ends up returning to the calendar as overtime. The national leader of the Citizen Movement, Jorge Álvarez Máynez, summed it up harshly from your Facebook profile stating: “Alert, Morena wants to make a fool of themselves in reducing the working day,” and maintained that the ruling “does not reduce the working day to 40 hours until 2030 and extends overtime hours to 12 per day, so that the working day does not actually decrease.” Less hours, less rest. For their part, business associations demand that the new regulations establish a reduction in rest timesso that the eight hours a day are effective work. Employers are asking to compact the day to reduce rest times during the day, so that it can be done uninterrupted. The Government has already positioned itself against it since its commitment was not to cut the rights already acquired by workers. 30 million Mexicans will work fewer hours. It is estimated that with the reduction of the working day, some 30 million Mexican employees will work fewer hours, better reconciling their work and family life. Mexico is currently one of the countries with the longest working hours in the world with an annual working day of about 2,124 hours per year, compared to an average of 1,687 hours. in OECD countriesthat is, around 23% more. In Xataka | Airbnb and digital nomads brought dollars to Mexico City: they have also brought the biggest housing crisis in years Image | Unsplash (Clayton Cardinalli, Denise Jans)

This is what we know about the alleged ‘hacking’ of the Treasury

From names and ID numbers to names and bank account codes of more than 47 million taxpayers in Spain. That is, at least, the information that a cybercriminal claims to have obtained from the Ministry of Finance, and that he would be offering in at least one of the many forums dedicated to trading stolen data. The question is inevitable: are we facing a real leak? And, whatever happens, what should be done in a scenario like this? Let’s go in parts. The alleged leak. As Hackmanac points outon January 31, a publication appeared in a forum frequented by cybercriminals in which a user called ‘HaciendaSec’ stated the following: “Today I am selling the updated Treasury database that contains the information of 47.3 million citizens.” In that same message, he detailed that the file supposedly included information associated with taxpayers such as: ID DNI/NIF Full name Address (type of street, address, address details, postal code, province, municipality, town) Telephone (country code, telephone) IBAN Email Last collected Total collection Capture of the forum with part of the information hidden for security The usual goal: make money. It’s no secret: the main driver of cybercrime is usually money. And this case fits perfectly into that pattern. ‘HaciendaSec’ is offering these supposed data in exchange for financial compensation. We do not know the price of the database, but we do know the terrain in which these actors usually operate: payments in cryptocurrencies, a channel that allows them to receive illicit funds reducing, at least in theory, the possibilities of being tracked and identified. The big question: has Treasury data been leaked? Here comes the key point. If we stick to the story of ‘HaciendaSec’ itself, it would be “an updated database.” The problem is obvious: to what extent is the word of a criminal reliable? The user includes a supposed “sample” of the data, but this type of evidence says little on its own and does not confirm the real scope of the incident, or even if it exists as such. In these types of cases it is advisable to maintain skepticism: sometimes these are recent and legitimate leaks, but other times we are dealing with compilations of other breaches, data that has already been circulating for a long time, or outright hoaxes designed to sell smoke (and make money). What the Treasury says. From Xataka we have contacted the Treasury to request comments on this matter. The official response, for now, is that they have no indication that their systems have been compromised. Of course: they also tell us that those responsible for security are working to rule out any type of intrusion. So, for now, we have to wait for more clarity about what happened. But in the meantime, it is advisable not to sit idly by. What to do about the alleged ‘hacking’ of the Treasury. As things stand, we have two different levels: what the author of the publication claims and what the Treasury maintains, which maintains an internal investigation without conclusive signs of compromise. In such a scenario, the most sensible thing is to act prudently and take preventive measures. In short: be attentive, distrust by default and verify before accepting anything as good. Anti-spoofing tools. We should be attentive to the campaigns phishing and impersonation, which is where many people end up falling. The Tax Agency remembers that it never requests confidential, economic or personal information, account numbers or card numbers from taxpayers by email, SMS or Bizum, nor does it attach annexes with invoice information or other types of data. Additionally, he recommends: Do not open messages from unknown or unsolicited users, delete them directly. Do not respond under any circumstances to these messages. Be careful when following links in emails even if they are from known contacts. Be careful when downloading email attachments, even from known contacts. Images | Treasury | freepik In Xataka | How often should we change ALL our passwords according to three cybersecurity experts

The eternal battle between whether you are from Apple/iOS or Windows/Android intensifies because the AI ​​wants to take advantage: Crossover 1×36

In the world of technology there are usually two main types of users: those who choose Apple’s closed ecosystem, and those who prefer to bet on (somewhat) more open alternatives such as those proposed by Microsoft with Windows and Google with Android. Is there a way to know which of the alternatives is better? Into that mess we get into this new episode of Crossover in which we analyze what a technological ecosystem is and the evolution of this concept. Thus, we review how Microsoft began to implement that idea without still using the word “ecosystem.” He did it with Windows because with it he had that central element on which to sell us other applications like Office or Internet Explorer in those beginnings. But with smartphones and the cloud, the ecosystem concept ended up making complete sense, and If there is someone who has exploited it in an extraordinary way, it has been Apple. It has done so, however, with a closed focussomething that has clear advantages, but also disadvantages. Faced with this conception, Microsoft first on desktop computers and then Google on mobile phones continued to promote open ecosystems, which gave much more choice but also posed their own problems. Added to all this now the rise of AIwhich these companies will undoubtedly try to use as a new argument to strengthen their ecosystems. They are all doing it already, and it remains to be seen whether or not this reinforces these ecosystems, whether open or closed. On YouTube | Crossover In Xataka | With Cowork, Anthropic has opened the doors to one of the most promising revolutions in AI: our computer

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.