controlled

that humanoid robots controlled by a central AI work

by

Samsung has planted itself in the MWC 2026 with one objective: to demonstrate that it is a ubiquitous company. What does this imply? Well, let them gain muscle with your screens everywherebut also show a powerful commitment to artificial intelligence in all links of the chain: from mobile phones to Samsung Galaxy S26 to the factories. And as a result of that intention for AI to be the pilot of everything, they have shown a science fiction plan: that robots and a central AI control their factories. And they want it the day after tomorrow. Independent. He concept of “agent AI” It’s one that we’re going to have to become familiar with because companies are going to put a lot of effort – and money – into this. It is an AI that no longer only responds to what we need, but can carry out actions autonomously. In a releaseSamsung assures that that agent principle that has been introduced in the Galaxy S26 It will be what dictates the future of its factories. The South Korean company wants these artificial intelligence agents to be the ones that “optimize workflows in production, predictive maintenance, repair operations and logistics coordination” in its factories, but an AI cannot execute things outside of the software. Need a physical interfaceand that’s where the other leg of the plan comes into play. Robots. They are the body of the brain and something that many companies are already exploring. A few months ago I traveled to China and came across the first store run by a robot. It is very simple and I described it as a “glorified vending machine”, but it meets the objective of these companies: to have spaces in which robots take care of everything. They don’t rest, they don’t have agreements and they don’t complain. And if companies like BMW either Xiaomi is already testing robots in its factories, Samsung does not want to be left behind. In the statement, the South Korean company states that they are already progressively introducing highly specialized humanoid robots for various tasks. For example, robots for facility management operations, others for the next steps of the production line, others in logistics, others for the transportation of materials, and precision robots for manufacturing. They point out that they are ideal in environments where human access is limited or dangerous and they are clear that it is something that will grow, with other robots dedicated to monitoring plant conditions, identifying risks and mitigating them before they occur. Total bet. In the end, it is about fully integrating AI across the entire manufacturing value chain: from logistics to production; from quality inspection to final shipment. They are designing a “next-generation autonomous production environment,” and they want to have it soon. The plan is that by 2030, “all manufacturing operations” will have completed the transition to this agentic and robotic AI. They are already at it, as we say, adding robots to production chains, but Samsung’s Executive Vice President and Head of Global Technology Research points out that the next phase is the “construction of autonomous environments where AI understands contexts in real time and executes optimal decisions.” NVIDIA. It sounds like science fiction, especially because of the deadlines they have, but they will not be alone in this adventure. Who is going to be by your side? Indeed: NVIDIA. At the end of last year, both they signed an agreement collaboration that includes the deployment of more than 50,000 NVIDIA GPUs and the use of the platform Omniverse to carry out the infrastructure of digital twins in semiconductor manufacturing. This is key to achieving that goal that Samsung seeks and as important as the AI reasoning systems in real time for robots. And for this they are also using the Jetson Thor platform from Jensen Huang’s company. Alternative to TSMC. That is the goal that Samsung wants to reach. Because right now they are one of the largest factories in the world (they have their Exynos processorsbut also its camera sensors that are in various devices, as well as its division of memory that powers NVIDIA GPUs), but what they want is to become an alternative to the undisputed queen: TSMC. To do this, Samsung is moving, opening factories in several countries around the world and investing enormous sums of money to be one of the legs of the business in the United States which also pursues this agentic AI and the end with haste and a good wad of cash: general AI. There are four years left to see if this objective, which seems like science fiction, is met. Images | Samsung, Xiaomi In Xataka | I have seen the result of a crazy night between a mobile phone and a robot: the Honor Robot Phone dances to your music as well as takes photos of you

Science has investigated why we bite our nails or leave everything until the last minute: “controlled explosions”

by

Biting our nails until it hurts, bingeing on junk food after a stressful day either open TikTok just when we have to start workingit is not an irritating habit that we would like to erase from our daily lives. But the reality is that science is beginning to see these behaviors in a radically different way: as a protection strategy for the organism. The brain seeks survival. As pointed out by different experts such as clinical psychologist Charlie Heriot-Maitland, author of Controlled Explosions in Mental Healthour brain prefers to inflict controlled “microdamage” on itself rather than face a greater and unpredictable threat. And the premise from which affective neuroscience and evolutionary psychology start is forceful: our brain is not programmed for us to be happy, but it is programmed to seek survival. Which is precisely what we did thousands of years ago when we tried to hunt or flee from predators. Systems that are still very present in our genetics. A hypersensitive system. This threat detection system is hypersensitive today. In the modern world we do not have to flee from a predator, but criticism from the boss or the fear of failing in a project activates the same alarms that a predator in the savanna activated in our ancestors. And faced with this unbearable stress, the brain looks for an escape route that acts as a “safety valve.” This is what Heriot-Maitland calls “controlled explosions.” Nail biting. Why can something as absurd as biting your nails or picking your skin be “protective”? The key is predictability. And in a chaotic world and an emotional, abstract and difficult to manage threat, cause us a little physical damage (like biting a cuticle), causes the brain to divert attention towards a specific, real stimulus and, above all, under our control. In this way it works as a “costly signal”, since we prefer a small and known damage to cushion emotional pain that we perceive as potentially devastating. Procrastinating is not laziness. scientific literature speaks in this sense of the self-handicapping (self-limitation), which suggests that we put obstacles on ourselves to protect our self-esteem. This way, if you stop studying for an exam and fail, you can tell yourself, “I failed because I didn’t study.” It’s a small damage to your ego. However, if you study to the maximum and fail, the conclusion is much more painful: “I failed because I am not capable.” The brain prefers the narrative of lack of effort (microdamage) rather than facing the threat of incompetence that poses greater emotional damage to anyone. It is not exclusive to us. In nature, there are numerous social insects that resort to defensive self-immolation in order to save their colony, as we already saw. In our case, the mechanism is something like this: we sacrifice our current well-being, such as physical health, to reduce a perceived long-term risk. The problem is that this system is designed for life or death situations, not to manage the chronic stress of the 21st century. In this way, what began as a useful defense ends up becoming a self-defeating pattern that generates more anxiety than it relieves. How to avoid it. If we understand that modern nails or procrastination are defense mechanisms, the solution changes completely. In this way, modern therapies, such as Compassion Focused Therapy, They propose that the first step It is not fighting against the habit, but understanding the reason for its existence. The most important thing in this case is not to punish yourself, since self-criticism is perceived by the brain as other threatens more, which reinforces the need to resort to the destructive habit to calm down. In this way, if we generate security, the brain will not have the need to cause these “controlled explosions.” Images | Sander Sammy Tim Gouw In Xataka | Procrastination is one of the great temptations of the mind. There are techniques to avoid it, according to science

has launched the Excel and Word that are controlled with prompts

by

Microsoft is trying to transfer the phenomenon of ‘Vibe Coding‘ – Write apps without knowing how to program, only with instructions to a chatbot – to office work. The bet is that anyone can do in minutes what I used to require hours of a junior consultant, without dominating Excel formulas or writing techniques. Agent mode In Excel, use OpenAI reasoning models to break down complex tasks in executable steps, as if an expert worked with you. Office Agentbased on Anthropic models, creates PowerPoint presentations and Word documents directly from Copilot chat. The company says that agent mode in Excel reaches a 57.2% precision in Spreadsheetbench, surpassing Chatgpt agent and Claude Opus 4.1, although below 71.3% of human precision. The context. Microsoft has been integrating the Office for months, but so far it was limited to specific assistance functions: summarize, rewrite, respond. Agent Mode and Office Agent mark a qualitative leap: IA assumes multipurpose tasks that previously required specialized knowledge. Use OpenAI models for agent mode within the applications and anti -appoal applications and models for Cat in the chat. This division speaks of a certain diversification: Microsoft had already incorporated the Anthropic in Github co -ilot and Copilot Studio A few days ago Summit Chauhan, from Office, summarize it like this: “It is a job that a first -year consultant would do, delivered in minutes.” Yes, but. The 57% accuracy in spreadsheet tasks leaves a considerable margin of error. Microsoft emphasizes that the results are auditable and verifiable, but still require human supervision for sensitive cases. In addition, Agent Mode is only available in the Excel and Word web versions in the launch. The support for desktop applications will arrive “soon”, according to the company. Office Agent, meanwhile, is limited to users in the United States. The cloud services dependence also raises questions: Anthropic operates its API about AWS, Azure rival. This would explain why their models are not yet really integrated in Office desktop applications. Between the lines. Microsoft presents this as democratization of access to expert capabilities. But it is also a way to justify subscriptions of Microsoft 365 Copilot. The relevant thing is that Microsoft is betting on AI doing heavy work while the user “guides and directs.” That is, the user goes from “doing” to “ask to be done.” What has already seen in programming, above all. And there is another reading: if Agent Mode can do in minutes what he had been having been implicitable for hours before for hours. The question is what happens to who before that job. Specifically, with the Junior. Microsoft has designed this for the process to be visible: the user sees what is doing in each step, as if observing an automated macro in real time. At stake. Microsoft competes against Google Workspacewhich also integrates generative in Docs and Sheets. But above all Now also competes against independent tools such as Notion AI, Gamma or Beautiful.AI, which promise to create presentations and documents from scratch without the need for Office. Microsoft’s advantage is domain: one billion people use office every month. If you get Agent Mode to work well, you don’t need to convince anyone to change tool. Just update your subscription. In Xataka | IBM’s ghost: Satya Nadella’s great challenge is to prevent Microsoft from ending up a technological fossil Outstanding image | Microsoft

With the battle against the controlled Muslims, the Byzantine emperor Teófilo decided to do something: defeat himself

by

Throughout the planet they extend battlefields that have gone down in history. It is estimated that there are some 12,000 localized battles Throughout history, being those of the twentieth century The bloodiest in history. Many of those ‘famous’ wars marked a turning point in one way or another, and the battle of Anzen between Byzantines and Muslims was one of them. But not for being a deed or a great military strategy, but for being an absolute embarrassment in which his army defeated himself. The legend. Between the eighth and twelfth centuries, the Arabo-Bizantine wars. It was a series of conflicts that, for almost 550 years, faced different Muslim caliphath with expansionist eagerness and the Byzantine empire. The conflict changed the political and military landscape in the Middle East and in a large part of the Mediterranean. Legend tells that one of the battles between the Byzantines and the abasi caliphate decided the … spirits. As we read in National Geographicin one of the battles during the ninth century, the Byzantines managed to repel the Arabs, but did not have time to celebrate the victory. At nightfall, the Byzantine soldiers began to see lights moving in the fog, some ‘Fatuos fires‘(flammable gases generated by decomposition matter on the battlefield) that soldiers confused with wandering spirits and souls. In the confusion the Byzantines panicked, they began to attack between them and those who fled were clear by a cliff. Anzen. As history, it is very good, but there are no documents that support it and it seems rather a story so that the soldiers do not get carried away by the Superstitions. The funny thing is that there was a historical episode in which the soldiers were dragged by fear, causing a humiliating defeat. It occurred in the same ninth century, specifically on July 22, 838. In the Anzen battle (current Dazmana, in Türkiye), the Byzantine emperor Teófilo He commanded between 25,000 and 40,000 men against an abasi force under the command of the Iranian Prince al-affshin. The previous year, the Byzantines had made an incursion into the ground of the caliphate, and the Caliph sought revenge. Your goal? Love affairemblem, one of the largest cities in the Byzantine Empire. There are many factors that influence a battle, such as the size of the armies, their armament and even morality, but something definitive is the strategy. And Teófilo came up with a great idea that turned out to be decisive. Galaxy brain. At one point in the battle, and at a point where their troops had the advantage, it occurred to Teófilo to reinforce one of the flanks personally. He took a group of soldiers and left the usual command place. Although in many films we see that leaders are the first to load, the common thing was that they stayed in a remote and privileged position. Not only saved their neck, but they could exercise military tactics in real time. Teófilo sneak away after the defeat Panic and unbalanced. He must have seemed a good idea, but this action was a turn in the dynamics of the battle. The Byzantine troops began to see that Teófilo was not in their command position and thought the most logical (sarcasm): “The emperor has died.” Without its leader, the troops are They demoralizedthey began to act indisciplinedly and there was a disruption in the ranks. Historical chance or strategy, just at that time the Turkish cavalry, with its archers, made a counterattack, which exacerbated chaos in the Byzantine ranks. The attack of the archers mounted and that idea that its leader had died in combat, undermined the morals and cohesion of the imperial army, dynamiting any initial advantage in combat. What do you do? Recall that Teófilo was still alive leading the attack in another area of the battlefield, but their forces had collapsed and the emperor, next to a group of soldiers from his personal guard and the Kurdish allies, was isolated in a hill, the Anzen hill. The Arabs surrounded them, but here there was a Deus Ex Machina de Manual: it began to rain, the strings of the Turkish arches were unused and, while the Arab troops waited for catapults to bombard the position, Teófilo and his personal guard slipped between the enemy lines. Sitío de Amorio for Muslim troops Fatal error. After that, the rumors of his death had already spread and the emperor not only had to fight against it, but against the Arab advance. The defeat in Anzen paved the way to Amorio for the Al-Afshin troops, where the looting of the city that had been a symbol of prestige and imperial power. It was not a definitive blow in the war between Byzantines and Abasíes due to internal conflicts in the caliphate, but of course it was a defeat for the history of the uncoordination of the troops and how important the command figure is. In Xataka | There is a reason why the Canary Islands is not British: the day that United Kingdom invaded Tenerife without knowing what was inside

Spanish universities controlled online exams with facial recognition. AEPD has decided that it is enough

by

The Spanish Agency for Data Protection (AEPD) has reached a clear conclusion about the use of biometric systems with artificial intelligence in online university evaluation. There is no possibility that they can be used legally, at least without a specific enabling law. The trigger. The AEPD has presented A complaint against the International University of Valencia. The VIU had been using a system that combined use of artificial intelligence tools with double camera recording (which the student must contribute) to monitor online tests. It is a practice that It has been doing for more than three yearsand that the agency has sanctioned rejecting the legitimacy of this data processing. Viu is not alone. The International University of Valencia is not an isolated case. Some of the most prestigious in Spain, such as University of BurgosUniversity Isabel I, European University, or the University of La Rioja have been implementing this system for years. It is a solution to the growing demand for 100% online training, with tools that allow the student to monitor without the need for the exams in person. The main objective, according to universities, is to avoid fraud and impersonations of identity during evaluations. The culprit. Smowl, this is the name of The online exam supervision tool. This solution, designed for both business and academic use, allows monitoring with webcam, extra camera, browser block, eyelashes control and, ultimately, replaces human role in exam supervision. In the case of the UIV, it was guaranteed that these data were pseudo and eliminated “quickly”, although it recognized that the processing of these data meant “a very high impact risk for the rights and freedoms of the affected people” The universities are covered in which it is the student who gives their consent to the use of these tools by accepting the general conditions of the course in which it has enrolled. The AEPD has another opinion. There is no legislation that covers its use. Universities are shielding that it is the student who gives their consent to the use of these tools by accepting the general conditions of the course in which it has enrolled. The AEPD has another opinion. “The consent cannot be considered valid because there was no real and effective alternative to students as the software used is the only method allowed to perform online exams. Their rejection by students involved losing their right to evaluation. Nor is the mandatory acceptance of general conditions to enroll when enrolled.” These data are of special category and are regulated by Article 9 of the General Data Protection Regulation (RGPD) Since 2022. According to the agency, there is currently no legitic exception in said article or a specific legal framework that enables these practices. It is also rejected that students can give consent, having no alternative available. But it doesn’t close the door. Data protection does not close the door completely to these types of systems. Specifies that it is necessary to develop specific regulations to determine “in what cases, conditions and under what guarantees this biometric treatment can be carried out”. “ Currently, without frame in which to protect yourself, the use of these tools will be subject to sanction for breach of the GDPR. A deep modification of the regulation would not be necessary, it would suffice with an exception that specifically reflect these scenarios of use. Facial recognition in Spain. It is not the first time that Spain calls into question the use of this type of systems. The OBERTA UNIVERSIDAD DE CATALUÑA was sanctioned In 2022 with 20,000 for using facial recognition in their exams. Outside the educational field, one of the most popular cases was that of Mercadona, fined 2.5 million euros for a pilot project in which they tested a facial verification system in their supermarkets. At a lower level, local companies have also faced large fines for breaching the regulations of the GDPR in the workday registry through biometry. Despite this, it is a technology used in video surveillance systems, Like Renfeor that of Madrid in its streets with hundreds of cameras with AI to reinforce the security of the capital. Images | Pexels (Andrea Piacquadio), Unspash (Dom Fou) In Xataka | Pau is approaching: here you have all the degrees related to technology and science with its cutting notes

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.