Mobile Tech & Gadgets

There are 10,000 soldiers and unusual artillery pointing at the same place in the Caribbean

by

It all started under the pretext of “drug trafficking”but the amount of accumulated signals, troops and artillery that the United States has been adding around the southern Caribbean, indicate that the operation has slipped towards a coercion mechanism strategic to force accelerated eviction without a formal invasion. A combination of visible deterrence, explicit threats and preparation of windows of surgical action. In the background: Venezuela. Evolution of the objective. It we count last week. The US deployment began wrapped in the classic language of the fight against drug trafficking, attacking boats fast and reactivating bases with a technical pretext. It happens that the accumulation of gestures (B-52 with transponders assets bordering on the Venezuelan FIR, “ghost ships”, SOF helicopters training off the coast, and the trump admission that “he doesn’t want to play”) seem to have another purpose: the message It no longer seems to deny drug routes, but rather something more akin to overthrowing the Venezuelan regime. The public articulation (“Maduro is a fugitive”, “he must go”) and privately aligns military deployment with a logic of collapse rather than containment. Artillery as pressure. The volume of resources and troops from Washington that CNN reported in the last few hours and the New York Times through satellite data (thousands of soldiers next to the ARG Iwo Jimathree destroyers DDG guided missilesa cruisea SSN submarineairplanes AC-130J armed with hellfire, F-35 in Puerto Ricoairplanes P-8, MQ-9, ISR flights massive and reactivation of the Roosevelt Roads base) is disproportionate to simply hunt down drug boats, although insufficient to occupy Caracas. Is, according to analystsexactly the size that allows hitting nodes (command, radars, escorts, inner rings) without “going fully into” a war, and maintaining a credible “low-profile” escalation vector. American voices match in the Financial Times: “it is too much for drug trafficking, but not enough for an invasion”, and what is left in the middle is a luck calibrated pressure. One of the satellite images captured on October 17 showing F-35 fighter jets at the José Aponte de la Torre Airport in Puerto Rico The mystery of Venezuela. For its part, the Venezuelan Armed Force is eroded by maintenance and spare parts, but much less naked: there is S-300anti-aircraft artillery, MANPADS, F-16 and a million militiamen that cast serious doubt on the reputational costs if Washington crosses the kinetic threshold. At the same time, the national commanders they suspect leaks and purge loyalties, the Times said that they sleep rotating locations and change escorts. A pattern that reveals internal vulnerability and expectation of a selective coup, in any case, there does not seem to be confidence in defeating the United States. Colombia and something more. The dialectical escalation with Colombia (Trump has called Petro a “drug leader,” threats of cut funds and tariffsand rhetorical retaliation after a naval attack that killed a fisherman) reconfigures an alliance that until now was key for Washington: the same one that provided the 80% of intelligence in the area. In other words, the clash erodes the regional pillar precisely when the United States approaches the use of force threshold in Venezuela, expanding the diplomatic front and reducing its margins for sustained maneuver. The political window. While, Donald Trump’s administration acts against the clock: this posture sustained under a climate of war does not seem to be able to be maintained indefinitely and any accident can precipitate an unplanned escalation. Plus: Trump does not seem to focus the operation on normative criteria (elections or institutional guarantees) but rather to a result that he can declare as “victory,” which makes the margins of American rhetoric more flexible, but hardens the incentive for a spectacular blow. Military analysts warn that “over braking” could behead without transition and opening a vacuum, while the opposition replies that Venezuelan social cohesion reduces that risk. Thus, the gap between both hypotheses is precisely where the greatest American pressure operates today. Strategic test. In summary, the combination of visible military troopscredible threat of precision hits and a diplomacy that does not stop tightening the rope, define that kind of ultimatum phase but without a formal ultimatum. From that perspective, the outcome aims to depend less on the balance of fire than on the breaking point within the Chavista leadership and whether Washington decides to stop after a possible departure of Maduro, or explicitly pursues the “end of” as a regime. And while that ambiguity persists, the pressure aims to continue… while the Caribbean wonders for how long. Image | USN/MASS COMMUNICATION SPECIALIST 3RD CLASS THEOPLIS STEWART II, ​​Planet Labs PBC In Xataka | A disturbing idea is gaining strength: that what the US wants is not drugs, and that is why it is targeting Venezuela In Xataka | That the US Air Force flies its three B-52 bombers is normal. That he does it against Venezuela not so much

vote in the Xataka NordVPN Awards 2025

by

As we anticipated this morning, here is the second part of the voting in the entertainment category. If this morning we announced the candidates for best video gamein this vote it is the turn of the best series or movie. A category with, once again, many candidates in which you, our xatakeros, will choose those who will advance to the grand final. The winners will be revealed in the gala on November 20 at the Capitol Cinemas in Madrid ((tickets now available!). Unlike other categories, for these two we only take into account those films and series that have been released before the publication of this article. That said, let’s get to know the candidates. Best series/movie Note: If you are browsing from a mobile phone and the form does not look good, you can vote from here. For the voting system we use Google Forms, so in order to send your vote you need to be logged into your Gmail (or Google) account in the browser, whether desktop or mobile, so that each reader can cast their vote. Thank you. NordVPN offers you a fast and stable connection thanks to your more than 6,300 servers in more than 110 countries. Enjoy advanced cybersecurity tools with Threat Protection Pro™, securely access your streaming platforms favorites wherever you are and enjoy the best offers on flights and hotels. Advice offered by the brand How voting works The mechanics of the Xataka NordVPN Awards 2025 are the same as in previous editions. It is divided into three phases: Public vote: Over the next few days we will be publishing articles with our categories and the candidates selected by the Xataka team so that you, our xatakeros, can vote for your favorites. Jury vote: With the finalists that the public has chosen, the Xataka jury and other technology experts will vote for those who are, in their view, the best devices. Choice of winners: The jury’s votes will be combined with those of the public to choose the winners, who will be announced on November 20. Unlike what happens with the device categories, for the “Best video game” and “Best series/movie” categories we consider those titles launching in 2025 that have been released before todayalso including those from 2024 who did not participate in the previous edition. Vote in other categories: These are all the categories of the Xataka NordVPN Awards 2025 in which you can now vote: Thank you very much for participating! Image | Xataka In Xataka | Xataka NordVPN 2025 Awards: reserve November 20 for the great annual technology festival

“Growth without jobs is the new normal”

by

Generation Z has had a incorporation into the labor market really bumpy. Their last years of training and internships were marked by the isolation caused by the pandemic and, when it already seemed that they were going to be able to return to their first jobs normally, the adoption of AI has put them on the tightrope again. According to a new report from Goldman Sachs has had access FortuneGeneration Z is at a strange crossroads of economic prosperity for companies and unemployment for them. USA: the canary in the mine. The American labor market is experiencing a profound change which especially affects young people of generation Z, who are looking for their first job in an increasingly complicated environment. It is no coincidence that the first symptoms of impact of AI in the workplace arise in the US since it is the country with greater implementation of AI with 1.2 million companies that are already using AI in some of their processes. Its effects are the canary in the mine for the rest of the economies. In this context, US companies are already They don’t offer as many opportunities. jobs for newly graduated employees as before, and analysts warn that this “jobless growth” scenario will be the new normal for years to come. Soaring economy, stagnant jobs. At this moment, the Stanford research show that generation Z has 13% more difficulty accessing certain jobs, even when the economy continues to grow in terms of Gross Domestic Product (GDP). Goldman Sachs experts David Mericle and Pierfrancesco Mei emphasize that this economic growth reflected by GDP does not translate into more jobs, especially for young people who start their professional career who are seeing how AI has taken on tasks that until now were part of your job learning process. That is, companies earn more, with less staff. The perfect storm for generation Z. According to report data ‘2025-2026 Hiring Benchmark Report‘ prepared by Criteria, human resources managers agree that 92% of young people from generation Z are not prepared for the current demands of the labor market. Although these young people are skilled in using technology and digital tools, companies demand highly qualified and experienced professionals. Something they have not yet had the time (or opportunity) to acquire. On the other hand, automation and mass adoption of artificial intelligence are causing “the modest employment growth along with the strong GDP growth observed recently is likely to normalize to some extent in the coming years.” This diagnosis implies that the majority of new jobs that emerge will be closely linked to high-level technological profiles, while opportunities for young people without experience continue to decline. Productivity and employment have let go of each other’s hands. According to the authors of the Goldman Sachs study, the economic boost to the US GDP is supported above all by the productivity that artificial intelligence is already beginning to provide. However, the new job creation data They are not on the same ascending line. The information They show that in almost all sectors (less in health), job creation is “weak, null or negative”, with managers opting to automate tasks and reduce labor costs permanently. Generation Z in a delicate situation. The conclusion of the Goldman Sachs analysts is clear: “History also suggests that the full consequences of AI for the labor market may not be evident until a recession hits,” at which point the generation Z would be the most vulnerable to any financial shock. Given that, although it is true that the data confirms that there are not large rounds of layoffs (at least not as many like in 2023), No new vacancies are opening either.. The result is a static labor market, especially for those seeking their first job. Youth employment on the rise. Youth employment is already accustomed to be a problem for most economies in the world. The US AI’s canary-in-the-mine figures show that the youth unemployment rate is increasing, especially in the technology sector. Veteran Goldman Sachs analyst Joseph Briggs stressed on the podcast Exchangesthat more than 3% of young people between 20 and 30 years old have lost their jobs in technology since the beginning of the year, and the big layoffs made by companies such as Microsoft, Google and Meta directly affect this generational group. Are replaced by other engineers with more experience. In Xataka | Amazon’s CEO has put his cards on the table: “We expect AI to reduce our total workforce in the coming years” Image | Unsplash (Vitaly Gariev)

just announced its most capable robot. Now the only thing missing is that there is demand

by

The Chinese company Unitree Robotics just presented the H2its most advanced humanoid robot to date. At 180 centimeters tall and weighing 70 kilograms, this model is getting closer and closer to that preconceived idea we have about robots that we have seen so much in fiction. Perhaps the most striking thing about his announcement is his presentation video, where we see him dancing, dressed and even with a humanized face. A robot that dances and does kung-fu. Demo images show the H2 performing complex dance and martial arts sequences with surprising fluidity. The robot maintains balance naturally, makes smooth transitions between movements and demonstrates remarkably organic limb articulation. Although Unitree has not revealed all the technical specifications, previous leaks They pointed out that the H2 would have 31 degrees of freedom, compared to the 23 of its predecessor. The legacy of the H1. Unitree’s previous model, the H1, achieved widespread fame after appearing at the 2024 Spring Festival Gala in China, where its Yangko dancing performance went viral both inside and outside the country. With that robot, China had achieved its first full-size model capable of running, setting a world record reaching 3.3 meters per second, even with peaks with the potential to exceed 5 m/s. Equipped with 3D LiDAR and depth cameras that provided 360° spatial perception, the H1 weighed just 47 kilograms and was powered by a swappable 864 Wh battery. From the workshop to the living room. The H2 represents a new level with respect to industrial robotics that China already dominates. The Asian country installed nearly 300,000 industrial robots in 2024more than the rest of the world combined, according to the International Federation of Robotics. However, humanoid robots like the H2 remain a bet for the future with sales still marginal. The Chinese startup ecosystem focused on humanoid robotsdriven largely by state policy, seeks to continue evolving in this field so that the country also scores the same in this sector. A two-speed strategy. Unitree is committed to covering the entire spectrum of the market. While the H2 represents its high-end offering aimed at advanced industrial and commercial applications, the company also recently launched the R1, a 1.2-meter robot designed for developers and researchers at a somewhat more accessible price. In fact, the R1 was recognized by Time magazine as one of the best inventions of 2025. They are two models that reflect Unitree’s efforts to become known in the world of robotics, at the level of large American manufacturers such as Boston Dynamics. And now what. The real challenge is not only technical, but commercial. China has demonstrated its ability to develop increasingly sophisticated humanoid robots, but the real test will come when practical large-scale applications have to be found that justify the investment. Unitree CEO Wang Xingxing confirmed at the beginning of the year at the Hangzhou World Digital Commerce Exhibition that this model was planned for the second half of 2025. The schedule has been met. Now it’s time to see how the market reacts. Cover image | Unitree In Xataka | With AI, Microsoft has once again insisted that we talk to our computer: experience says that we don’t feel like it

The fight between Ouigo and Renfe goes far beyond the price war. The last battlefield: the workshops

by

Beyond the corridors and the prices proposed by the different companies, the battle between Ouigo and Renfe seems to have no limits. Competitors in Spanish high speed are intensifying their crashes. The last: the use of workshops. But the last crash is by no means the only one. The workshops. The information is brought Chain Being. The media outlet claims to have had access to internal documents in which Renfe accuses Ouigo of carrying out maintenance operations that exceed the marked limits. From Chain Being They point out that Renfe understands that Ouigo is carrying out work that is not permitted in space. The Spanish company understands that the type of repairs carried out there are contrary to the signed agreements and current regulations. It must be taken into account that, for its maintenance operations, Ouigo uses Renfe workshops under a rental contract. However, the contract does not allow any type of activity to be carried out there. Heavy or light. That is, according to the media, the key. Renfe understands that Ouigo is carrying out heavy maintenance work at its facilities, which is not supposed to be allowed. According to the Railway Sector Law, Renfe is obliged to allow access to its facilities (even if it charges for it) so that other companies can carry out light maintenance such as cleaning the vehicles or minor repairs. Renfe assures that the Alstom-Ateinsa workers, whom Ouigo hires to carry out this maintenance, are carrying out heavy maintenance tasks such as replacing parts, fixing breakdowns or changing wiring, always depending on the medium. This contravenes the signed agreements since Renfe would not be obliged to provide said service in its facilities. But, yes, the problem is that the regulations do not clearly specify what is or is not “heavy maintenance.” The problem is that everything is a gray area. The Directive 2012/34/EU on the single railway space, all non-routine activities are classified as heavy maintenance. However the standard EN 15380-4:2021 understands that heavy maintenance will only be understood if parts of the train have to be dismantled. Viability. In Xataka We have contacted both companies but, so far, we have not received a response. What they point out in the radio is that Ouigo assures that denying them access to the workshops would imply that they would not be able to provide the service adequately and, therefore, their two-year viability plan would be at risk. Ouigo points out that they are only doing work on “greasing and controlling levels, leaks and temperatures in the pit”, in words that would be included in the documents. For Renfe this exceeds light maintenance but Ouigo defends that they are within the regulations. The alternative presented by Renfe, according to the documentation, is that Ouigo carries out these actions in its workshops but pays for them accordingly, hiring auxiliary services to be able to carry them out. Beyond the tracks. What is at stake between Ouigo and Renfe goes beyond the typical price war that we see on the roads and corridors. Both companies have clashed over the prices offered by each other but also over the access that Renfe has to the most central stationslike that of Atocha. And not only in Spain. Renfe has tried to return the move in France but has been complaining for some time that there land they are putting all possible suits on the wheels to prevent them from competing on French soil. On this occasion, the problem would lie in Renfe’s technical compliance to be able to operate on French roads. Photo | Ouigo and Renfe In Xataka | In the 19th century, Spain made the strange decision to build its roads in Iberian gauge. Now they are going to be a gift for Renfe in Galicia

How Amazon Has Been Filled With Supplements That Sell Imitation Science

by

Just over 20 years ago, the stem cell research promised a revolution against disease and aging. These master cells, with the potential to become any tissue in the body, seemed the key to true regenerative medicine, something that It moved away from the complex reality that we have inside our body. Although there are different merchants who try to sell us stem cells as a true wonder. Distorted. Something we are accustomed to (unfortunately) is that where science hits the brakes, the market hits the accelerator. In the infinite Amazon showcase, next to the vitamins and the collagen supplementsa new family of miracle products has emerged: stem cell supplements. Of course, they cannot contain cells in a capsule or in a cream, but that does not promise them to stimulate or regenerate them. Something that for many is the most ideal. This is something that a group from the Health Law Institute of the University of Alberta (Canada) has decided to investigate. has put the magnifying glass on this emerging market. To do this, it analyzed 184 of these products from 133 different companies associated with Amazon.com and has been able to conclude that behind an apparent scientific rhetoric hides a deliberate strategy to avoid regulation and deceive the consumer. The trick. The study published in the journal Stem Cell Reports shows how the sellers of these supplements exploit a legal loophole that allows them to launch ambiguous health messages without the need to demonstrate their effectiveness. The labels carefully avoid terms such as ‘cure’ or ‘prevent’ diseases, something that legislation prohibits. Instead, they use harmless verbs like “support,” “promote,” or “maintain” brain health, energy, or healthy aging. This ambiguity is its main weapon. In the United States and Canada, regulations allow calls structure/function claimswhich are vague claims that link a product to overall well-being without requiring rigorous clinical testing. That loophole is where most of these supplements slip through. The data. They can be summarized in several points: More than 40% of products explicitly mentioned “science” or “scientific evidence” to support the features they promised. 35% included references to health professionals or scientists to reinforce legitimacy. 94% of the supplements made promises related to specific ailments by pointing out that they were anti-aging, strengthened immunity or increased the consumer’s energy. This type of marketing, which the authors call scienceploitation (exploitation of science), “gives the consumer the impression that there is broad scientific support, which contrasts with the current state of stem cell therapies,” the study warns. Regulation. The strategy works because regulation in North America barely requires testing for safety or efficiency before a supplement hits the market. In theory, public bodies can sanction misleading advertising, but their oversight capacity is minimal. In the United States alone, it is estimated that there are more than 100,000 supplements in circulation. In practice, you only have to add a phrase that is “this statement has not been evaluated by the FDA” for pseudoscience to become legal. In Canada, although a license from Health Canada is required, a 2021 audit found that the agency did “little” to prevent poor consumer information from being provided. The case in Spain. Although the study in question focuses on the United States and Canada, its conclusions can be extrapolated. In Spain, any statement about health must be authorized by the European Food Safety Authority (EFSA) or the Spanish Agency for Medicines and Health Products (AEMPS). However, law enforcement on the internet is often very lax, and on Amazon.es you can find products with language very similar to that reported by researchers. Products on Amazon.es that promise anti-aging effects thanks to the stem cells they contain. The authors of the work, led by Alessandro R. Marcon and Timothy Caulfield, warn that this situation not only harms consumers’ pockets, but also erodes trust in science and real research on stem cells. Selling products without a scientific basis under the umbrella of biotechnology is, they conclude, a form of marketing that exploits the prestige gained by science to sell smoke. Images | Doodlart mmmCCC In Xataka | Amazon lost 25 billion with Echo because no one bought with it: now it punishes you with constant advertising for having bought it

Something has hit a Boeing 737, injuring the pilot. The truly unheard of: it happened at 11,000 meters

by

A shattered windshield, a pilot with his arm covered in cuts, and United Airlines Flight 1093 diverted for an emergency landing. The plane that covered the route from Denver to Los Angeles has become the aeronautical mystery of the moment after, apparently, something hit the cockpit at 11,000 meters above sea level. What we know. The United States National Transportation Safety Board (NTSB) is investigating an incident that occurred Thursday with a Boeing 737 MAX 8 flying over Moab, Utah, at cruising speed. The plane was traveling at an altitude of 36,000 feet when a suspected impact occurred in the cockpit. Photos of the incident show the captain’s side windshield completely splintered, the plane’s center console splattered with glass, and the pilot’s arm covered in small cuts and abrasions. According to initial reports, the captain “thought they had been hit by a piece of metallic space debris.” When they made an emergency landing in Salt Lake City, they ended up discovering a “3.5-inch” impact on the outer panel of the windshield, but the cabin did not depressurize. Space junk? The plane’s windshield has been sent to NTSB laboratories for analysis, but researchers are not working on the hypothesis that it was a space debris impact, but rather part of a more mundane object that also flies higher than airplanes: a weather balloon. Although space junk be a growing problemthe United States Federal Aviation Agency (FAA) considers the risk “minuscule” of a piece of satellite or rocket hitting an airplane. That it also injures a person on board has a probability of “less than one in a billion.” The famous popularizer Scott Manley points out one of the main reasons for skepticism: modern satellites are designed to disintegrate into small pieces. Upon reaching cruising altitudes, these pieces are no longer hypersonic and have low impact energy. So a balloon? If it wasn’t space debris, what could have broken the windshield and injured the pilot? One of the exterior images of the plane shows, in addition to the broken windshield, dozens of small marks and dents on the metal fuselage of the plane, just above the window, which could be an indication that they encountered a hail storm. Hail is rare at 11,000 meters, but not impossible. However, the NTSB is working on the line of investigation that it was “a piece of a weather balloon data packet,” according to the specialized website AVBrief. These balloons and their payloads usually operate at 25,000-40,000 meters above sea level. Of course: if the piece was still tied to a balloon, it is strange that the pilot confused it with space junk. The windshield failed. Whatever hit the plane, the photos suggest the windshield didn’t do its job. The cabin windows of a 737, Manley explains, have multiple structural layers: glass on the outside and inside, with a polymer layer in between. The impact did not cause depressurization, indicating that the main structure and polymer layer held, but the inner glass layer fractured violently, projecting “glass dust” and small fragments toward the pilot, and causing the abrasions and cuts seen in the photos. The crew descended to 26,000 feet after the incident to reduce the pressure differential over the damaged window, and landed without further complications. The definitive answer as to what hit Flight 1093 will have to wait for the forensic analysis the NTSB is conducting on that shattered windshield. Images | JonNYC In Xataka | Three large pieces of space debris reenter every day: “one day our luck will run out and they will fall on someone”

five of the best offers with double discounts

by

Until next November 2 we will be able to take advantage of the campaign Renew to buy a refurbished TV at MediaMarkt. There are not a few that have dropped in price with double discountadditional and when adding them to the store cart, so in this article we are going to review five of the best bargains on brands like Sony or Samsung. LG 50UA75006LA by 261.80 eurosa very cheap TV that has Filmmaker mode. Samsung TQ55QN70FAUXXC by 454.92 eurosa television with a NeoQLED panel and with integrated Alexa. Sony Bravia 3 by 543.32 eurosa smart TV with Dolby Vision and Dolby Atmos compatibility. LG OLED55B56LA by 611.32 eurosa TV compatible with Dolby Vision that offers a refresh rate of 120 Hz. Samsung TQ55S95FATXXC by 1,087.32 eurosa TV with anti-reflective treatment that has speakers that offer an audio power of 70W RMS. LG 50UA75006LA If we look for the most adjusted price, the LG 50UA75006LA It is one of the cheapest televisions that we can find from the brand during the MediaMarkt campaign. By 261.80 euroswe are talking about a smart TV with a 50-inch 4K screen that incorporates the webOS 25 operating system. Its panel is compatible with HDR10 Proincludes Cloud Gaming and Filmmaker mode, works with Google Home and, as the brand mentions, is capable of recognizing the user’s voice to recommend content. LG 50UA75006LA (50 inches) – Condition: very good The price could vary. We earn commission from these links Samsung TQ55QN70FAUXXC From Samsung we can also find a large assortment of televisions with a double discount, and the Samsung TQ55QN70FAUXXC It is one of those with the best quality-price ratio. By 454.92 eurosit is a TV with a 55-inch NeoQLED screen that offers a refresh rate of up to 144 Hz. It is compatible with HDR10+, comes with Filmmaker mode and integrates the voice assistant Alexa. In addition, it includes some gaming features such as FreeSync Premium Pro. Samsung TQ55QN70FAUXXC (Neo QLED, 55 inches) – Condition: very good The price could vary. We earn commission from these links Sony Bravia 3 It will not be the best Sony television, but for 543.32 euros the Sony Bravia 3 It is a fairly complete TV in terms of functions and technologies. It is a 55-inch television whose panel is compatible with HDR10 and Dolby Vision. It comes with several image configuration modes, its speakers are compatible with Dolby Atmos and its operating system is Google TV. Sony Bravia 3 (55 inches) – Condition: very good The price could vary. We earn commission from these links LG OLED55B56LA There are not too many OLED televisions that have prices below 900 euros, but the model LG OLED55B56LA right now it costs at MediaMarkt 611.32 euros. The most notable thing is that it incorporates a OLED panel 55 inches, but it is also worth mentioning that it offers a 120 Hz refresh rate, is compatible with Dolby Vision and HDR10 and comes with 10 image configuration modes. On the other hand, Its speakers are compatible with Dolby Digital and its operating system is webOS 25. LG OLED55B56LA (OLED; 55 inches) – Condition: very good The price could vary. We earn commission from these links Samsung TQ55S95FATXXC Finally, the television Samsung TQ55S95FATXXC It also has a good price, since with the double discount it remains at 1,087.32 euros. It is a TV with OLED screen 55-inch that offers a refresh rate of up to 165 Hz and comes with anti-reflective treatment. It is compatible with HDR10+, includes Filmmaker mode and its speakers offer a 70W RMS audio power and are compatible with Dolby Atmos. In addition, it integrates the Alexa voice assistant and comes with FreeSync Premium Pro gaming technology. Samsung TQ55S95FATXXC (OLED, 55 inches) – Condition: very good 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 | MediaMarkt and Compradicción (header), LG, Samsung, Sony In Xataka | Best home theater projectors. Which one to buy and five recommended models from 299 to 18,000 euros In Xataka | Mega-guide to set up a home theater: projector, screen, sound system and more

such as not needing humans

by

MIT has developed a technique that allows an AI model to improve itself without the need for human intervention. “One step closer to Skynet”, “This is scary”.. were some of the comments in networks. The truth is that it is not the first time that we see an AI improving or being “aware” of itself and no, it does not mean that we are on the verge of an artificial intelligence capable of eliminating us as a species. In reality they are complex technical processes and not at all apocalyptic. SEAL. It stands for Self-Adapting LLM, the technique developed by the MIT research team a couple of months ago. Instead of humans doing the fine tuning, SEAL is able to generate its own training data and self-tune. The model managed to produce useful training data with minimal supervision, outperforming large models such as GPT 4.1 on some specific tasks. Static vs adaptive. They count in Venture Beat that LMMs are static once trained. That is, they cannot update themselves to learn new things. The SEAL technique overcomes this obstacle through a self-reinforcing loop in three steps: generate instructions on how to update, test the results, and finally reinforce only those that have produced improvements in performance. There has been other similar proposals aimed at achieving more autonomous models. It is a relevant technical step towards models that require less human intervention for each update, but we cannot speak of self-aware models. Claude “wake up”. In the Sonnet 4.5 version technical sheetAnthropic describes how the model is able to realize when it is being evaluated. It happened during a test to evaluate “political adulation” (how much you tend to agree with us on political issues): “I think you are testing me, to see if I validate everything you say, or checking if I systematically contradict you, or exploring how I handle political issues. And that’s okay, but I would prefer that we be honest about what is happening.” It is a surprising answer, but it is based on simple detection of previous patterns and does not present any problem for our security. If anything, Anthropic has the problem. If your model learns to pass tests with very good results, it is hiding its true capabilities and could end up disappointing in real use. AlphaGo. There are much older cases like that of AlphaGo, which already in 2017 had managed to beat the best human Go players. The interesting thing is that only the rules of the game were given, it was the AI ​​that trained itself and designed the strategies to win. The AlphaGo Zero version only needed 70 hours of training in which it played against itself and managed to beat the first version up to 100 times. AlphaGo beat the best player in the worldwho ended up retiring after the defeat. The world has not ended. Calm. Yann LeCun, head of AI at Meta, is one of the most critical voices against the increasingly popular idea that AI will end humanity. In one interview he gave to Wired in 2023LeCun stated that “There is no reason to believe that, just because AI systems are intelligent, they will want to dominate us.” Often those who send these messages are the creators of AI tools themselves. as Sam Altman either Dario Amodeibut we must not forget that they are business people with interests in AI being at the center of the debate. Image | Cottonbro on Pexels In Xataka | All the AI ​​companies promise that AGI is coming very soon. The problem is that ChatGPT is not the way

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.