Mobile Tech & Gadgets

Claude Code is being the big favorite among programmers. So much so that he already signs 4% of everything that is uploaded to GitHub

by

It is worth taking a look at how generative AI It is transforming the daily lives of many programmers. And little by little these tools are conquering the environments of millions of developers. The achievement in this aspect is for Claude CodeAnthropic tool, which already represents 4% of all public commits uploaded to GitHub, according to a report by SemiAnalysis. The media says that, if it maintains its current pace of adoption, it is very possible that it will reach 20% of all daily contributions before the end of 2026. Although there are nuances that should be highlighted. Why is it important. Claude Code is slowly gaining the reputation of being the favorite tool for programming with AI. The tool works radically differently than traditional code wizards. It is not a chatbot integrated into an editor like Cursorbut rather a terminal tool that reads entire code bases, schedules multi-step tasks, and executes them with full access to the developer’s computer. You can start from spreadsheets, entire repositories, or web links, understand context, verify details, and complete complex objectives iteratively. The interesting thing is that, by default, Claude Code includes a co-authorship note if the user has used this tool in their program and uploads it to Github. But the user can also decide not to include that signature if modify the parameters by Claude Code, so that 4% could remain small. In March of last year, a month after its launch in private beta, Claude Code already had the co-authorship of about 15,000 Github commits in a period of 48 hours. Things have ended up escalating quickly. Opinions. The newsletter stands out the comments of some industry professionals regarding the vibe codding. Andrej Karpathy, one of the first to coin the term vibe codding, recognized in a post that he is “starting to lose the ability to write code manually.” Ryan Dahl, creator of Node.js, counted directly that “the era of humans writing code is over.” Boris Cherny, creator of Claude Code, assures that “practically 100% of our code is written by Claude Code + Opus 4.5“. Even Linus Torvalds, creator of Linux, has fooled around with vibe codding for some of his personal projects. It should be noted that, despite all the benefits of Claude Code, it is not perfect. Already we pointed out some time ago the words of Kelsey PiperAmerican journalist for The Argument, who explained that 99% of the time using Claude Code is like having a magical, tireless genie, but 1% of the time it’s like yelling at a pet for peeing on the couch. He can and does make mistakes. It also gets stuck. Hence, the expertise of the person who uses it also plays a very important role. Beyond programming. There is an increasingly latent threat with the use of AI tools (well there are a few that accumulate already). And according to account SemiAnalysis, any information work that follows the READ-THINK-WRITE-CHECK pattern can be automated with this technology. The report mentions sectors such as financial services, legal, consulting and data analysis, which add up to billions of workers globally. Anthropic has already taken the next step with coworkreleased a few weeks ago, which is basically Claude Code applied to general office work. According to the company itself, Cowork was developed by four engineers in ten days, mostly with code generated by Claude Code himself. The tool can create spreadsheets from receipts, organize files by content, write reports from scattered notes… And all with access to your computer. The big consultancies and AI. In December, Accenture signed an agreement to train 30,000 professionals on Claude, the largest deployment of Claude Code to date. OpenAI, for its part, Frontier has launched focused on business adoption so as not to lose steam in the field of corporate use of AI, a business that can end up being very lucrative for startups. Cover image | Anthropic and Mohammad Rahmani In Xataka | Programming is the new board of AI. OpenAI and Anthropic have made it clear with GPT-5.3-Codex and Claude Opus 4.6

They have found a bacteria capable of increasing your risk

by

We often think of the health of our mouth as something completely isolated that has no more significance than the odious cavities that we get. forced to go to the dentist or the bad breath. However, science has been warning for years that the mouth is the gateway to much more complex systems, such as the possibility that a bacteria from our gums travel to the breast tissue and may accelerate tumor growth. An unwanted traveler. The protagonist of this new discovery is the bacteria Fusobacterium nucleatum, an old acquaintance of dentists. We are talking about an opportunistic bacteria that thrives in dental plaque and is one of the main culprits of periodontitis, which is undoubtedly one of the most recognized gum diseases. What the team led by Dipalo Sharma has recently demonstrated is that this bacteria does not stay still on the gumsbut it has the ability to travel through the body to the breast tissue or even also is already linked to colon cancer. Its effect. The study In this case, he used mice to simulate two different scenarios in order to see how this very common bacteria behaved. The first of them was to inject the bacteria into the breasts of healthy mice, where precancerous inflammatory lesions began to be seen. In the case of injecting into existing tumors is where the alarms go off, since in these mice the presence of the bacteria tripled the size of the cancer and caused lung metastases in 100% of the cases observed. How he does it. It’s the million-dollar question: how does a bacteria from the mouth know that it has to go to the chest and how does it manage to do so much damage? Science has found an explanation at a molecular level that begins with inflammation of the gums in periodontal disease, since this causes the bacteria to enter the bloodstream. Once in the stream, the bacteria begins to travel and takes advantage of a very specific protein, called Fap2, which acts like a key that searches for a specific lock: a sugar called Gal-GalNAc, which turns out to be very abundant on the surface of breast cancer cells. Creating a shield. Once the bacteria adheres to the tissue thanks to this specificity, it begins to colonize, but it also has the ability to suppress the cells in charge of our defense. And specifically those that defend us from cancer cells that bypass the body’s checkpoints. Furthermore, it induces direct DNA damage and preferentially colonizes cells that have mutations in the BRCA1 gene, exacerbating the risk in genetically predisposed people. Dental hygiene. The result of this research leads us to a very clear question: does not brushing your teeth cause cancer? Logically not. In the field of health, causality is not as simple as ‘do this and that happens’, but rather it works as an accumulation of risks that increase the chances of generating a problem such as cancer. A risk factor. In this case, science suggests that having periodontitis, due to poor hygiene sustained over time, is associated with an increase of around 22% in the risk of suffering from breast cancer. And it is not the first time that dental disease is a risk factor of this type. A well documented case is in the relationship between deep dental caries and bacterial endocarditisan infection of the inner lining of the heart. That is why the recommendation here is always to maintain good oral hygiene and always treat cavities as soon as possible when they appear. Images | Caroline L.M. In Xataka | AI is no longer a promise in breast cancer: the largest clinical trial confirms that it detects more and reduces the burden on the radiologist

with his latest Rosco, he has achieved audiences typical of the last century

by

Rosa Rodríguez has entered the history of Spanish television, thanks to the 2,716 million euros she won in the legendary final rosco of ‘Pasapalabra’. The nightly special gathered 4.1 million viewers at its peak (a 45.3% share), figures that seem taken from another decade. But what is notable is not the prize or the specific audience, but rather that this contest, broadcast outside the prime timehas maintained for 25 years a capacity for convening that defies all logic. The fragmentation of audiences pulverizes formats each season, but ‘Pasapalabra’ grows. What happened. After 307 duel programs with Manu Pascual, Rodríguez, an Argentine teacher living in Galicia, completed the Rosco that gave her access to the largest jackpot ever awarded by ‘Pasapalabra’. Rodríguez and Pascual starred in the longest duel in the history of the program, 307 broadcasts faced in the final Rosco, a mark that far exceeds any other confrontation in the format. Pascual accumulated the absolute record for individual participations with 437 programs, and on six occasions he was one letter away from completing the rosco. On Thursday night, Rosa correctly resolved the 25 definitions (from “cruiser” to “Earl Morrall”, the American football player who closed her victory – not without controversy, since her pronunciation was not entirely correct, which triggered the inevitable on social networks. tongo accusations-) and dethroned Rafa Castañountil then holder of the largest jackpot with 2,272 million euros obtained in March 2023. Pascual left the contest with 270,600 euros accumulated, a considerable figure that does not mitigate the frustration of having touched the jackpot on half a dozen occasions. The audience. The data turns Pasapalabra into a statistical anomaly. This season, the contest registers a daily average of 1,928 million viewers with a 20.3% screen share. These figures correspond to its usual evening broadcast, in a time slot, eight in the afternoon, which the industry does not consider prime time and which competes with the end of work days, commuting and family routines. The tentacles of Pasapalabra. The impact of the program transcends its own broadcast. On Thursday Antena 3 reached 18.9% daily averagedouble its usual performance. ‘The Anthill’with Rosa and Manu before Rosco, scored a spectacular 23.5% of shareits best figure since March 2023. Vicente Vallés’ nightly news program reached close to 3.1 million viewers, its highest in three years thanks to the audience awaiting the outcome. The evening magazine ‘Y Ahora Sonsoles’ and the daily series ‘Sueños de Libertad’ also recorded season highs dragged by the ‘Pasapalabra’ effect. It was the day with highest television consumption of the entire seasonwith 10 million Spaniards in front of the television after 11:00 p.m., 20% more than the previous week. A revealing fact: the final Rosco segment alone reaches a 25% screen share and 2.6 million viewers on average, surpassing the global audience of the entire program. It is the moment of maximum tension, when the secondary screens turn off. 25 years. The permanence of ‘Pasapalabra’ in the Spanish television ecosystem for a quarter of a century is complex to explain. Since its debut on Antena 3 in July 2000, the contest has aired on three different networks, several judicial stoppages and presenter changes. It has not lost cultural relevance. Its format maintains a deliberately simple structure: two contestants accumulate seconds in verbal agility tests that they then invest in the final Rosco, 25 definitions whose answers correspond to the alphabet. That invariability is, paradoxically, one of its greatest attractions. An anomaly. The context in which Pasapalabra thrives makes its success even more surprising. At the end of 2024 63.3% of Spanish households with Internet access used at least one paid audiovisual platform. Netflix, Amazon Prime Video, Disney+ and the rest of the on-demand services have radically reconfigured consumption patterns: viewers decide what to watch, when to watch it and on what device to play it. The rigidity of the traditional television schedule should be an obstacle, but it is not However, the numbers refute the supposed obsolescence of linear television. Digital platforms accumulate 16.7% of total audiovisual consumption in Spain, and traditional television maintains 83.3% according to a Kantar analysis from July 2024. Among those over 50, free-to-air television continues to be the dominant medium, with consumption exceeding three hours a day on weekends. ‘Pasapalabra’ capitalizes on that type of audience. The unique touch. What distinguishes ‘Pasapalabra’ from extinct formats like ‘Who wants to be a millionaire?’? Its ability to generate events within the routine. Each Rosco is not just another episode: it is a potentially historic event, a unique opportunity to witness a record. The architecture of the growing jackpot transforms the daily broadcast into a series with no pre-established end. The suspense builds up for months until it explodes on nights like this Thursday. At prime time. Atresmedia’s decision to move the delivery of the boat to prime time generates a recurring debate among Pasapalabra’s loyal audience. Miguel Aparicio, director of the program, recognized that initially the team resisted this strategy: they preferred to “reward that follower who pays attention day after day” while maintaining the surprise factor. When Rosa completed the Rosco, Antena 3 built an entire programming architecture around the event, with a special prior to 8:00 p.m. in the program’s usual time slot, the appearance of the contestants on ‘El Hormiguero’ and finally the broadcast of the decisive Rosco after 11:00 p.m. This tactic has been repeated with the last big jackpots: Pablo Díaz in July 2021 (30.8% and 4.3 million), Rafa Castaño in March 2023 (37.4% and 4.5 million) and Óscar Díaz in May 2024 (30.1% and 3.2 million). They were all moved to nighttime hours after weeks of building expectations. The strategy works. Not only for the contest, which multiplies its usual audience, but for the entire grid. In Xataka | Telecinco’s crisis is so great that it is leading it to extreme measures: merging sets to save costs

The Spain we know is not prepared for today’s world

by

“The worst thing has been the explosions, we thought the houses were going to collapse.” María José Díaz, from Diario Sur, spent last night talking to the neighbors from Grazalema that have been relocated to Ronda. That phrase perfectly sums up the terror that ran through the people of Cadiz. The evacuation of Grazalema. It has rained a lot in the mountains of Cádiz, that is not new. And they are not onlyalmost 600 l/m² on the rainiest dayis that in recent weeks more than 2,000 l/m² have been accumulated. That is what has turned the streets of the town into a continuous river. That is what has caused water to flow from the floors of the houses, from the baseboards, from the wall sockets. That is what has finally caused the ‘explosions’ (the noises or cracks) that at first seemed like storms, but were quickly identified as hydroseisms. Grazalema is in an environment of karst limestone rock. This suggests that beneath the ground there is a whole network of microcavities, conduits and small sinkholes. When the system becomes saturated and the water table rises, water can escape everywhere. So he has done it. What is reported in the press is calling ‘hydrosisms’ They can be understood as the response of the soil to that enormous amount of water. In Grazalema, the creaks are interpreted more as a form of rearrangement of the terrain. It may seem strange, but there is clear evidence of the process (also in Spain). Why has it been evacuated? A priori, the evacuation makes technical sense (the state of the clogged aquifer is being studied), but also psychological (the situation among the population – as evidenced by testimonies – was becoming a nightmare). What lessons can we learn from all this? As González Alemán recalled, we can’t say for sure that all this has something to do with climate change. It will have to be studied in detail, but what is certain is that yes has assumed (is assuming) a stress test of the water infrastructures of Andalusia and, by extension, of Spain. And that should lead us to reflect on the enormous urban reconversion that will have to be undertaken if this follows what the trends indicate. It is not just towns like Grazalema, nor the coasts of the country. It is not only the buildings built in flood zones, nor the retaining walls that appear insufficient. It is the system as a whole. A system that it is not clear that we can change in time. Image | Heparin1985 In Xataka | If the question is how the Andalusian water system is holding up all this water, we have an answer: they are going to evacuate Grazalema completely

that of “50,000 Russians a month” or giving Moscow what it wants

by

Throughout history, the cold has acted as a silent weapon that has changed the course of entire wars: in 1812, the Russian winter destroyed Napoleon’s army during their retreat from Moscow, causing more casualties than many battles. In the Winter War From 1939-1940, Finland used extreme temperatures and frozen terrain to hold back a vastly superior Soviet Union force, and in World War II, the winter of 1941 paralyzed the German troops at the gates of Moscow. In all cases, the cold accelerated defeats, collapsed logistics and forced decisions that were not in the original plan. Something similar is starting to happen in Ukraine. The cold as an accelerator of war. Winter has turned war into a race against time because extreme temperatures amplify the impact of each Russian attack against energy infrastructureforcing entire cities to live without heat, electricity or water for days or weeks. With minimal close to −20 °C In many enclaves, each damaged power plant, each destroyed substation or each prolonged blackout is no longer just a technical problem but is a military and political factor that shortens the margins of resistance and pushes us to make decisions that are increasingly harsh and unthinkable until recently. Energy as a goal. Since winter began in the war, Moscow has had clear your objective. Russia has systematically hit power plants, thermal plants and distribution networks again, knowing that the damage is cumulative and that repairing under constant bombing is almost as expensive as rebuilding. Ukraine, for its part, has avoided a total collapse of the system thanks to quick repairs, generators and management increasingly flexiblebut the price is enormous: buildings without heat for weeks, networks saturated when the power returns and an exhausted population that lives pending of blackout and shelter schedules improvised. Kamikaze logic. In this context, an unprecedented idea appears strongly, kyiv’s most extreme bet: to accelerate the war by attrition until it becomes unbearable for Moscow. The government has explained that the idea of ​​causing up to 50,000 Russian casualties per month It is not proposed as a slogan, but as an explicit attrition strategy to force a negotiation based on the opponent’s weakness. If you will, it is a flight forward that assumes that, if the war cannot be slowed down and winter multiplies the suffering, the only way out is to drastically raise the human cost for Russiaeven knowing that Ukraine will also pay a very high price. The limits of the war of attrition. This strategy clashes with clear structural problems: lack of infantry, shortage of drone operators and a technological competition in which Russia has cut advantagesespecially in electronic warfare and fiber optic drones. As many analysts point out, prioritizing the constant elimination of enemy soldiers can give tactical results, but it does not always solve the key problem of operating depththat is, the Russian ability to continue moving troops, ammunition and drones from the rear while the front remains stable. The invisible front. In Insider told that the cutting off of Russian access to satellite communications systems via Starlink has shown the extent to which modern warfare depends on connectivity. The interruption has generated specific disorganization in Russian units and has been celebrated in Ukraine as a key advantage, although it has also affected its own and civilian users, demonstrating that each technological gain is very fragile and requires constant management. In the middle of winter, any added failure in communications or coordination translates directly into more casualties and more chaos. The unthinkable idea. As military and climate pressure accumulates wildly, I told a few days ago the new york times that a growing part of Ukrainian society start to contemplate through surveys what was previously little more than a taboo: accepting territorial concessions in exchange for firm security guarantees. It is not yet a majority, nor even a decision made by the leadership, but the simple fact that it is being discussed reflects the extent to which the cold, blackouts and a war with no clear end are forcing a profound rethinking about what it means to win or simply survive. A dilemma pushed by winter. What seems abundantly clear is that the scheme that emerges is hard and lacking in epic some: winter is literally freezing the population Ukrainian, and its effect is accelerating the war and narrowing the options. Thus, Ukraine seems pushed to choose between maximally intensifying the kamikaze logic of the “50,000 Russians a month” to force a quick outcome or accept territorial concessions to stop the destruction before another winter just as bad or even worse. The cold does not decide on its own, there is no doubt, but it does act as the factor that has turned an already long and exhausting war into an urgent decision. Image | armyinform.com.ua, 7th Army Training Command In Xataka | “A human safari”: going outside in a Ukrainian city is now equivalent to being a shooting target for drones In Xataka | The war in Ukraine has become something absurd: there are drones shooting at Russian soldiers dressed as “penguins”

We believed Amazon was already spending too much on AI. Your answer to Wall Street: spend even more

by

The honeymoon between AI and Wall Street is over. Amazon knows this very well, having just received that dreaded “we have to talk” message from investors with a drop of more than 10% in its shares yesterday. It seemed that the stock markets rewarded the fact that companies They invested absurd amounts of money in AI. It is just what Amazon announced yesterday, but that strategy has had a totally negative response in the markets. what has happened. Amazon presented yesterday financial results for the last quarter of 2025. Revenue grew by 14% and net profit by 6%, modest figures that were not very popular. But above all, I did not like that Amazon announced that it estimated a capex (capital expenditure) of $200 billion in 2026 in AI. Amazing. Wall Street used to reward, now it punishes. In 2025, that capex was $131 billion, and Amazon is determined to continue betting everything on AI. Before, investors rewarded that audacity. Now they are punishing her: the shares plummeted 11% “after hours“, and it will be today when those actions start with that reflected fall. We want return on investment. That market reaction is not an isolated event. Amazon’s fall comes just hours after Microsoft or Google suffered similar falls. The market before valued the potential of AIbut now he demands return on investment more than ever and has become impatient. Big Tech had operated with a blank check, but when revenue forecasts fall short of estimates, optimism evaporates. Income grows, yes, but not that much. The real problem is the imbalance between capex and revenue growth. AWS grew a spectacular 24% in revenue, but spending is growing at an even greater rate. Google, Amazon and Microsoft are trapped in a kind of infrastructure “arms race”: the first one to stop spending loses, and that is a big problem. He who does not risk, does not gain. Amazon CEO Andy Jassy explained that “this is an extraordinarily rare opportunity to forever change the size of AWS and Amazon as a whole. (…) We are going to invest aggressively to be the leaders.” It is a speech identical to that Mark Zuckerberg said a few months ago when he said he was willing to lose hundreds of billions on AI: not investing them would be worse for Meta. But Amazon is much more than AI. There is another disturbing element in this huge bet by Amazon. The reality is that the company has many expensive fronts. From the Kuiper satellite network to compete with Starlink to the robotization of its Whole Foods logistics and other areas. When adding AI to the equation, the math doesn’t seem to work out. Optimism ends. Historically, large technology companies have taken advantage of the optimism of the market and investors to justify spending forecasts completely unrelated to their income. In 2026, with the macroeconomic situation of “we no longer like risk” —tell it to bitcoin— and the pressure for profitability, “free optimism” has disappeared. If you are going to spend like crazy, you have to raise like crazy too. Amazon is doing well, AI is not. This total commitment to AI is preventing us from seeing that the rest of Amazon’s businesses are doing very well. Online sales grew by 10% and advertising grew by a notable 23%. E-commerce, the cornerstone on which Amazon was built and operates, is funding the AI ​​party, but it is turning into a bottomless pit. Like Qatar’s GDP. According to the world bankQatar’s GDP in 2024 was $219 billion. That Amazon invests almost the same in AI data centers alone is dizzying. It is the same thing that we said yesterday about Google, which also projected a capex of 135 billion dollars by 2026. The figures are no longer dizzying: they are crazy. Beware, obsolescence. And all that investment can end up wasted, especially because there is an implicit risk in the data centers that are built: in three or five years they could become obsolete if the architecture of AI chips changes radically. It is bread for today, and hunger for tomorrow… without counting the energy factor or the water consumption. Xataka | While Silicon Valley seeks electricity, China subsidizes it: this is how it wants to win the AI ​​war

attack the trident that dominates the market

by

One more day, new bad news related to the RAM memory crisis. If you were expecting a Steam Machinenow you can expect it to be more expensive. The rise of AI is causing a component crisis that has no clear end. The SSDs have gone up in price a lot and have 32 GB of RAM on your PC It is the new “I have land.” All Big Tech needs more and, in the absence of it, Intel has chosen to get into it like an elephant in a china shop. Hand in hand with SoftBank to create its own memory chips for data centers and, in the process, take a bite out of the South Korean industry that controls the scene. No end in sight. Intel has been covered several times in recent days. Like a phoenix, it seems that the crisis is behind us and, after years of promises, realities begin. They are ready to start producing its new generation of processors, but also its CEO, Lip-Bu Tan, has commented that They will start producing GPUs for data centers. With that they want to take a bite of the cake that NVIDIA is eating almost alone, but also, the executive let a more bitter pill: “there is no relief in sight for the end of the AMR crisis.” So it points out that does not see a horizon for this price escalation before 2028, and it makes sense if we take into account recent forecasts or possible ‘strange’ movements by some companies. Intel 🤝 SoftBank. Market estimates such as those of TrendForce They point to a memory price increase of between 90 and 95% quarter-on-quarter in this first quarter of the year, but it is not the only thing: SSDs will also rise between 55 and 60% due to one of their components, NAND memory. It is a bad time to build a PC, although companies are moving to open RAM factories, but not for you: for the AI. And here Intel, as we already saidhe doesn’t want to be left behind. A few months ago, Intel partnered with Japan’s SoftBank to find a replacement for HBM memory for data centers. Fix a problem. HBM memory (high bandwidth memory) is ideal for data centers. They allow a large amount of data to be temporarily stored, and they do so at high speed. The problem is that they get very hot and are complex to manufacture. It is one of the key components of GPUs and requires both large amounts of power and optimal heat dissipation. What Intel and SoftBank are looking for is to create an alternative based on stacked DRAM chips. They are not that optimal, but the idea is to find a way to wire them as efficiently as possible so that they are a threat to the HBM memory monopoly. “Monopoly“The idea is to have a prototype in the short term with the aim of starting to market it by 2030. We will see if AI fever lasts so long. But anyway, if the experiment goes well, it could be a significant blow to the South Korean companies Samsung and SK Hynix, as well as the American company. Micron. They are the three companies that they practically control the market for both RAM like HBM chips. And something as important as it is symbolic: it will be the first time that Japan aspires to return to the throne of memory chip manufacturers that it ruled in the 80s and that it lost to South Korea. Although be careful: Samsung is also investigating these stacked DRAM memories. NVIDIA and its demands. In the end, they are all moving. The big producers have already detailed its roadmap for the development of RAM for the next five years. And the whales that are taking over the product -NVIDIA-, are ‘encouraging’ the production companies let’s get the batteries. Without going any further, a few days ago Jensen Huang, boss of NVIDIA, met with representatives of the Taiwanese industry (TSMC, Asus or Foxconn) and told them that this year he needed a lot of memory and a lot of wafers. Also I know it has told Samsung. As we said in the first lines, if you were hoping for a quick solution to the crisis or wanted to build a PC, the news is not good. And all the movements that we are seeing in the industry to expand production and find solutions point to the same direction: continue powering data centers. Images | Intel In Xataka | TSMC’s only problem was that it was in Taiwan. So the United States has decided to get her out of there

Demolishing prices on Google and Xiaomi mobiles, the Nintendo Switch 2 with free game and more bargains, Bargain Hunting

by

There’s nothing left for Valentine’s Day, so many stores have begun to lower the price of many devices, whether or not they are related to Valentine’s Day. In the new Bargain Hunting we are going to review the best deals we have seen all weekwith pretty good offers on cell phones, sound bars and even Meta glasses. Google Pixel 10 by 584.10 euros with coupon, a high-end mobile phone with an excellent photography section. Ray-Ban Meta by 246 eurosperfect glasses if you are looking for more than just glasses. Hisense AX5125H by 199 eurosa great sound bar with subwoofer and rear speakers. nintendo switch 2 by 459 eurosthe latest console launched by the brand with totally free games. Xiaomi 15T Pro by 799 eurosa mobile phone that has no less than 1 TB of internal storage. Google Pixel 10 If there is a mobile phone that we should pay close attention to, that is the Google Pixel 10. During these last few months it has dropped a lot in price, its current offer on MediaMarkt being the best: with the coupon 10PIXELMMFEBRUARY stays for 584.10 euros. It is a fairly small mobile phone that stands out mainly for its photographic quality, but also for its design and because its software will receive updates for many years. The price could vary. We earn commission from these links Ray-Ban Meta Also at MediaMarkt we can find at a very good price (in fact, it is the minimum in the store) the Ray-Ban Metavery practical glasses that with the discount are worth 246 euros. They accept prescription lenses and allow record audio or video and translate in real timemake video calls and even listen to music, among other things. The price could vary. We earn commission from these links Hisense AX5125H The best deal we’ve seen this week on sound bars has fallen on the Hisense AX5125Hwhich can be found right now on Amazon for a price of 199 euros. We are talking about an interesting bar because it comes with a subwoofer and wireless rear speakers, but it is also compatible with Dolby Atmos and offers a power of up to 500W at 5.1.2 channels. The price could vary. We earn commission from these links nintendo switch 2 The nintendo switch 2 It has received many offers since its launch and now MediaMarkt has a very interesting one. Not only is it slightly lowered, but 459 euros Plus we can take the video game ‘Donkey Kong Bananza‘ totally free. And we are talking about a title that was nominated for Best Game of the Year in 2025. Nintendo Switch 2 + Donkey Kong Bananza The price could vary. We earn commission from these links Xiaomi 15T Pro As an alternative to the Google Pixel 10, this week we have found a quite interesting offer on the Xiaomi 15T Pro. Amazon (like other stores) has it for 799 euros in your 1TB configurationso it is ideal if you want to have a lot of storage. It has the MediaTek Dimensity 9400+ processor, its screen offers a 1.5K resolution and its cameras are signed by Leica. The price could vary. We earn commission from these links Some of the links in this article are affiliated and may provide a benefit to Xataka. In case of non-availability, offers may vary. Images | Google, Meta, Hisense, Nintendo, Xiaomi In Xataka | The best mobile phones, we have tested them and here are their analyzes In Xataka | Best sound bars in quality price (2026). Which one to buy and seven recommended models from 99 euros

The US and Mexico have just taken the step to treat them as such

by

The digital economy, the energy transition and a good part of advanced industry depends on a set of materials whose importance is only perceived when they are scarce. Governments classify them as critical minerals precisely because of their essential role and the fragility of their supply chains. This change in perception, from industrial resource to strategic asset, is reordering commercial decisions and international alliances. The step taken now by the United States and Mexico is part of that deeper transformation in how the materials that support contemporary technology are managed. A clear objective. Both governments have announced the development of a bilateral Action Plan that will explore commercial and coordination tools aimed at mitigating risks in the supply of critical minerals. Beyond the technical content that has yet to be defined, the announcement itself indicates that the management of these raw materials has come to occupy an explicit place on the bilateral agenda between the two countries. The detail. The announced framework describes an intention for cooperation, but does not yet establish its operational content. Both the minerals that will be included and the trade mechanisms that could be applied remain to be specified. This lack of precision is relevant: usual lists of strategic materials, like lithium or copperare part of the industrial and energy context in which the plan is discussed, but we will have to wait to know what elements will end up making up the pact. Price floors. The proposal introduces an unusual instrument in the public debate: setting minimum values ​​for certain imports to respond to “global market distortions” and reduce vulnerabilities in the supply chain. The idea appears linked to the resilience of these chains and considerations of economic and national security in the argument that accompanies critical minerals. Of course, its eventual application would be subject to subsequent agreements and its fit into international trade frameworks. The agreement also emerges under the shadow of an unavoidable trade event: the review of the North American treaty shared by the United States, Mexico and Canada (T-MEC). The proximity of this process gives the plan additional meaning within the regional economic architecture. However, the information released about this bilateral initiative does not include mention of Canadian participation, a detail pointed out by Reuters that delimits the immediate scope of the ad. It is not a static list. The concept of critical mineral describes a condition more than a closed catalog. United States energy legislation links them to economic or national security, the vulnerability of their supply chains and their indispensable role in the manufacture of products, as explained by the USGS. But that classification changes over time as technology, demand or external dependence evolve. Therefore, rather than a fixed list of materials, what is really at stake is the capacity of each economy to secure resources considered strategic at each industrial stage. The board is moving. The bilateral agreement appears in parallel to a broader international deployment to reinforce supply chains considered strategic, with new frameworks and memoranda. But its reading does not end in that dimension. For Mexico, coordination opens a way to consolidate its role within the North American industry and attract projects linked to mining, processing or advanced manufacturing. The result is a two-way movement: an expanding global strategy and, at the same time, a redefinition of the place that Mexico can occupy in it. Images | Dominic Vanyi + Nano Banana In Xataka | Greenland has 1.5 million tons of rare earths. The problem is that there are no roads to get to them.

Anthropic wanted to secretly scan and then destroy millions of books to train its AI. It hasn’t been so secret

by

A language model for AI needs input if it is to be trained to be more accurate and effective. The issue is how the information is obtained and whether there is an ethical way to do it that is profitable for the technology company in power. There is no doubt that the preferred option for companies has been to use all possible physical and digital content without anyone’s permission. There is also evidence. A judicial leak reveals that Anthropic invested tens of millions of dollars in acquiring and digitizing literary works without permission from the authors. According to account Washington Post, the project, internally called “Panama”, was part of a frenetic race among big technology companies to accumulate massive data to train their artificial intelligence models. How it all started. The Panama Project was launched by Anthropic in early 2024. According to internal documents revealed per the Washington Post, the goal was to “destructively scan every book in the world.” Furthermore, these documents also explicitly state that the company did not want anyone to know that they were working on it. In about a year, the company spent tens of millions of dollars buying millions of books, cutting their spines with hydraulic machines and scanning their pages to feed the AI ​​models that power Claudeits star chatbot. According to the media, the books, once digitized, ended up being recycled. Because has come to light. The details of the project have been revealed in a lawsuit for infringement of rights copyright filed by literary authors against Anthropic. Although the company agreed to pay $1.5 billion to close the case in August 2025, a district judge decided to make more than 4,000 pages of internal documents public last week, exposing the entire operation. They are not the only ones. Court documents reveal that other technology companies such as Meta, Google and OpenAI had also participated in this race to obtain massive information to train their models. According to revealed According to the documents, an Anthropic co-founder theorized in January 2023 that training AI models with books could teach them “how to write well” instead of imitating “low-quality internet slang.” On the other hand, an internal Meta email from 2024 described access to a digital library of books as “essential” to be competitive with rivals in the race to dominate AI. However, the documents revealed by the media also show how Meta employees expressed concern on several occasions about the legality of downloading millions of books without permission. An internal email from December 2023 indicates that the practice had been approved after being “escalated to MZ,” apparently referring to CEO Mark Zuckerberg. According to court records to which the media has had access, the companies did not consider it “practical” to obtain direct permission from publishers and authors. Instead, they found ways to mass-acquire books without the writers’ knowledge, including downloading unauthorized copies from third-party sites. Chat logs from April 2024 show an employee asking why they were using servers rented from Amazon to download torrents instead of Facebook’s own. The answer: “Avoid the risk of tracing” the activity back to the company. Data torrent. The documents to which the Washington Post has had access also they test that Ben Mann, co-founder of Anthropic, personally downloaded over 11 days in June 2021 a collection of books from LibGen, a gigantic library of copyrighted content. The outlet further revealed that, a year later, in July 2022, Mann celebrated the launch of the ‘Pirate Library Mirror’ website, which boasts a massive database of books and openly claims to violate copyright laws. “Just in time!!!” Mann wrote to other Anthropic employees, according to the outlet. Anthropic stated in legal documents that it never trained a revenue-generating business model using LibGen data nor did it use Pirate Library Mirror to train any full model. Anthropic’s legal solution. According to point the medium in its article, faced with the legal risk, Anthropic changed its strategy. The company hired Tom Turvey, a Silicon Valley veteran who had helped create the project Google Books two decades earlier. Under his direction, Anthropic considered purchasing books from libraries or secondhand bookstores, including New York’s iconic Strand bookstore. The company ultimately ended up buying millions of books and stacking them in a giant warehouse, often in batches of tens of thousands, according to court filings. The Washington Post assures In addition, the company worked with used book sellers in the United Kingdom. A project proposal mentions that Anthropic sought to “convert between 500,000 and two million books in a six-month period.” What the law says. Most legal cases against AI companies are still ongoing, but the media mention two court rulings that have considered that the use of books to train AI models without permission from the author or publisher may be legal under the “fair use” doctrine of copyright. In June 2025, District Judge William Alsup determined that Anthropic had the right to use books to train AI models because they process them in a “transformative” way. He compared the process to teachers “teaching schoolchildren to write well.” That same month, Judge Vince Chhabria ruled in the Meta case that the authors had not shown that the company’s AI models could harm the sales of their books. In the Anthropic case, the physical book scanning project was considered legal, but the judge determined that the company may have infringed copyright by downloading millions of books without authorization before launching Project Panama. The final agreement. Instead of facing a trial, Anthropic agreed to pay $1.5 billion to publishers and authors without admitting guilt. According to point According to the media, authors whose books were downloaded can claim their share of the settlement, estimated at about $3,000 per title. Cover image | Emil Widlund and Anthropic In Xataka | If AI is going to leave us without jobs, in the United Kingdom they are already seriously discussing the solution: a universal basic income

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.