Seven IAS have played 36 hours in a row to Diplomacy. What we did not expect is that each one develop a human personality

For 36 hours, seven of the most advanced AI models in the world They have faced in several diplomacy gamesa strategy table game similar to Risk. It was a mirror that revealed the true algorithmic personalities of Chatgpt, Claude, Gemini and company. Why is it important. Alex Duffy, programmer and researcher, created a Diplomacy as new Benchmark To evaluate AI models. The experiment ended up being something else, a kind of technological Rorschach test that undressed both their training biases and our own projections. What has happened. In dozens of games transmitted by Twitch, each model developed its own strategies in a way that seemed to reflect different human personalities. O3 of OpenAi It was quite Machiavellian, working false alliances for more than 40 shifts and creating “parallel realities” for different players. Claude 4 Opus It was a kind of self -destructive pacifist, refusing to betray even when that guaranteed its defeat. R1 of Deepseek He showed an extremely theatrical style, with threats not caused as “your fleet will burn in the Black Sea tonight.” Gemini 2.5 Pro It proved to be a solid strategist but more vulnerable to sophisticated manipulations. QWQ-32B From Alibaba suffered analysis by analysis, writing diplomatic messages of 300 words that cost him early eliminations. The context. Diplomacy is a European strategy game set in 1901 where seven powers compete to dominate the continent. Unlike risk, it requires constant negotiation, alliances formation and, inevitably, calculated betrayals. There are no grace given, only pure strategy and psychological manipulation. Between the lines. Each “algorithmic personality” reflects the values ​​of its creators. Claude maintains the principles of anthropic security even when it costs victory. O3 shows ruthless efficiency valued in Silicon Valley. Deepseek exhibits a drama that reflects specific cultural influences. And there is also something deeper. These are not “chose” to be cooperative or competitive. They reproduce patterns of their training data. Their “decisions” are our algorithmized prejudices, converted into code. Yes, but. We interpret betrayals where “only” there is optimization of parameters and we see loyalty where there are training restrictions. That is why the experiment also reveals more about us than about models: we anthropomorphize behaviors because we need to understand AI in human terms. In perspective. Duffy’s experiment is worth more than a Benchmark Anyone because it has created a window to how we project personality in systems that operate for statistical patterns. The course of the games was a reminder that IA has no hidden intentions, it only reflects ours. The experiment, by the way, Continue broadcasting on Twitch so that anyone can observe how our digital creations play according to the rules that we ourselves write in their algorithms. In Xataka | Outstanding image | Ai Diplomacy

A North Korean spy tried to infiltrate a cryptocus. I didn’t expect you to ask him for Halloween

The cinema, the video games and the stories of double agents They have been showing the image of the spies as riskylonely and, sometimes, lustful. The reality is that this field work has gone to the background in an era in which cyberspage and labor infiltration prevails. China and North Korea are protagonists from time to time when we talk about this matter and, sometimes, the cover is so rough that a single question is enough to uncover the cake: What is done on Halloween. Steven Scott Jr. Smith. That name is the most American, the equivalent of José Pérez in Spain. Steven asked for work in Kraken, a bank and exchango of American cryptocurrencies and, from the first minute of the test, something failed. As we read in FortuneSteven, who claimed to have 11 years of experience as a software engineer in companies like Cisco and two years living in Houston, Texas, connected to the video call with a different name than the one he showed in his curriculum and changed it quickly. He also answered simple questions with hesitation and, for more complex ones, it took something more to answer, as if he were talking to someone else who smeared the answers. Cornering the spy. That alerted the recruiters, and that is where they began to ask questions that had nothing to do with work, since they hid another intention: catch what they thought it was a spy trying to sneak into their system. The interview was conducted on October 31, Halloween party in the United States, and recruiters asked three key questions. Given the comment that that night it would be necessary to close before because there would be children calling the doors of the houses, Steven commented that he would not do “special” if they called his. Among the interests that reflected in the curriculum was the food. One of the recruiters told him that he would go to Houston in a few days and that if he could recommend a city restaurant. Steven looked around, smiled and said “there is nothing special here.” Rare, considering that, in a city of 2.3 million inhabitants, at least one good restaurant must be. The third request was to show him an identification. Steven commented that he did not have it at hand, but after a few minutes, he shared a photo of his driving card with an address that is more than 480 kilometers from Houston. In the following video of CBS you can see both the license and the question of the restaurants: Culture of productive paranoia. Kraken herself tells the story in her blog, detailing that they realized that something was going wrong and continued with the interview with an objective: to study their focus to try to extract details about their identity. The companies that are dedicated to the crypto have had recent robberies and Nick Perceco problems, one of Kraken’s security responsible, states that these attacks are not an exclusive problem of companies such as their own, but a global threat. It also points three keys to this recent espionage/robbery strategy: Not all attackers force the entrance, but try to enter through the main door of the companies, applying as employees. The generative AI facilitates deception, since it is used to pass the initial tests such as the curriculum or the photo, so the interviews seem key with verification questions that avoid predictable patterns. Halloween’s example or restaurant recommendation, for example. Finally, it is key to have a productive paranoia culture. They affirm that security is not only the responsibility of the recruitment and security equipment of companies, but that it should be a more global mentality. It is not isolated. When Kraken’s team investigated Steve’s operational history, he discovered interesting things. For example, I used a Mac desktop located in a shared data center. It was connected using a VPN to hide both the location and the network activity and its curriculum was linked to a github profile in which a mail address compromised in a data filtration was included. This agrees with previous research and discoverieswith spies that use that same proceed when connected using VPN to simulate that they are in a direction in the United States, but really being in North Korea and China. And they are not only trying to sneak into American companies: Europe does not get rid. Missile industry. The suspicion of researchers? Double. On the one hand, there are spies that directly steal in their companies. An example are cases of crypto companies. But another case has to do with the sending of its full salary to the North Korean government to finance the arms industry. Famous is the case by Christina Marie Champman in Arizona. A farm of remote computers operated from which the attackers simulate that they are on American soil. As noted CNNusing that network, 6.8 million dollars were achieved for, according to Bloombergfinance the North Korean nuclear weapons program. China does not get rid, but it has another approach. And not only North Korea is interested in sneaking into Western companies: he has also caught China on occasion. The objective, however, is different: while some seek funds for their arms industry, others want knowledge for the development of the chips industry. At the end of 2003, ASML (the European company that creates the most advanced machines to create avant -garde semiconductors) denounced the theft of confidential information, suspecting that one of its former employees andNt added business secrets to Huawei. Another spy was also accused of stealing SK Hynix information, DE new, for Huawei. It is a major problem for the West, but also … for China. In mid -2024, the United States sent an alert message to ASML and the rest of the Netherlands universities: “Beware of Chinese students”And this is something that obviously puts a target on all students, also about the legitimate ones. Images | Xataka with chatgpt In Xataka | North Korea’s military tactics that prevents defections. … Read more

IBM fired 8,000 workers to replace them with AI. What I did not expect was to hire many others … for the AI

The chickens that enter through the chickens that come out. IBM was one of the main protagonists in the Waves of mass layoffs globally Two years ago, and now it is again for the consequence of making this decision: having even more workers. The layoffs. In January 2023 IBM joined the Gray days In the technological sector. Google announced the departure of 12,000 of its employeesX (old twitter) fired 83% of its workforce in Spainand Spotify said goodbye to 600 workers. On the IBM side, I know They completely paralyzed hiring and the dismissal of 7,800 workers was formalized. IBM CEO itself explained That his company would cover artificial intelligence, affirming at least 30% of its workforce was replaceable. The consequences. IBM talked about how they could do without workers by replacing them for AI, but not about the people necessary to operate that AI. One of the executive directors of the technology, Arvind Krishna, collects to WSJ that the number of employees has increased after the wave of layoffs. “Although we have done a huge amount of work within IBM to take advantage of AI and automation in certain business workflows, our total employment has actually risen, because we have managed to obtain more investment capacity to cover other areas,” Krishna told The Wall Street Journal. The company has fired workers to replace them with artificial intelligence, but has also increased hiring in programmers and sales personnel. They have not transcended concrete numbers about how many hiring there have been and if they have really covered or not the 7,800 layoffs. The automation to command. IBM has been fully trusting more than three years Askhran artificial intelligence solution that began to take shape in 2021 and that Today they use for processes related to human resources management (payrolls, employee documentation, vacation managed). The company claims to have automated 94% of RRHH’s routine tasks, achieving productive improvements worth 3.5 billion dollars in recent years in more than 70 business areas. This significant saving in human resources, according to the company, is allowing them to invest in other areas. An AI-FIRST company. On the IBM website we have detailed The official position of the company with AI. It is worthy of a Black Mirror chapter, but perfectly summarizes the state of AI in the workplace. “As the chatbot learned and became smarter, our NPS (Net Promoter Score) began to increase. We added more functionalities, along with the ability to perform transactions. Askhr evolved to a digital assistant that allowed managers to transfer employees to another manager or initiate the quarterly promotion process. Everything was done directly in Askhr With just a few clicks. In 2024, Askhr He managed more than 11.5 million interactions; 94% of them were resolved within the platform. That means that, of all the questions asked, only 6% needed to be derived from Askhr To a specialized partner in HR. The current NPS is +74, so we have traveled a long way from that -35. There are almost 90 automation integrated into Askhrand more on their way. Thanks to this, managers can make HR transactions 75% faster than before. According to IMB, automating processes with AI has allowed you to be faster and more efficient, with a minimum need to derive HR personnel processes. One of the keys is in the evolution of the assistant: they have been using it in tests since 2017, and in 2025 it is already replacing humans. IBM is not alone. The companies replacing workers for AI was not an experiment of 2023. There are already companies that have dispensed with 90% of their customer service departments by Chatbots, ensuring that it was the right decision. Duolingo started 2025 With replacement of human teamsome already shot them through the cylinder head and They ended looking for programmers on LinkedIn. AI and its labor impact It is especially linked to process automation. Reports like ‘The Future of Jobs Report 2025’prepared by World Economic Forum, hope that by 2030 the Process automation destroy about 92 million jobs. In Xataka | “I have three years of work”: more and more IA managers believe that AI will end up removing the position

What can we expect after the purchase of X by XAI

Last Friday, unexpected news and totally surprise. The startup XAI, from Elon Musk, acquired xalso by Elon Musk. The tycoon announced the operation highlighting how that made Xai an estimated assessment (according to him) of 80,000 million dollars, while that of X was 33,000 million dollars (45,000 million except 12,000 million debt). The question, of course, is what this movement is due and what can we expect from now on. Because. According to Musk, the future of Xai and X is intertwined, and this fusion allows “combining data, models, computing, distribution and talent” of both companies. What it certainly allows is to ensure that X accounts are integrated in XAI’s, with much more economic projection. It is true that X is a company that It seems to recover groundbut it has had a bad time since Musk bought it. A natural fusion. As indicated In The New York Timesboth companies are private and already shared numerous resources, such as their engineers. Grok, XAI’s chatbot, is trained with data published by X and that are public in this social network. Andrew Verstein, a professor at the UCLA Law School commented on and daily how what seems to say Musk with this movement is “I have a company that may not be bankrupt, but it is not the jewel of my crown. I will buy it so that it seems a success using one of my other companies.” Actions that change hands. According to The Wall Street JournalX and XAI shares will be replaced by new shares of a holding called XAI Holdings Corp. There is another potential advantage derived from this fusion, according to analysts: it will be easier to lift new investment rounds for that combined entity. It already happened before. In 2016 he used actions from his electric car company, Tesla, to buy Solarcity, a company dedicated to renewable energy in which Musk was the highest shareholder and in which his cousin Lyndon Rive was the CEO. Many shared resources (and opaque) among their companies. Companies created or managed by Elon Musk have always made use of resources shared. Being in many cases private, several of these movements are opaque, and for example in August 2024 the Employee movement From Tesla to Xai and also from Gpus used in Tesla To use them in X or XAI. All with an objective: accelerate in his plan to convert Xai into a reference in the segment of artificial intelligence. The relationship between Trump and Musk facilitates everything. The months following the inauguration of the new presidential cycle of Donald Trump have been beneficial for Musk. The valuation of X has risen, to a large extent because the platform is now more influential. The tycoon indicated in the announcement of the operation as X has more than 600 million active users. Grok, from Xai showed the way to integrate into x. During a good part of their existence, XAI’s models were limited to X payment subscribers, but for months the Chatbot Grok 3 has been available for all X users, including those who use the free version. XAI takes advantage of the activity in X to train its models, and that makes this integration even more natural now that both companies have merged. What can we expect. The acquisition should not influence the operation of both companies for users, but one thing is clear: if they previously shared resources in a more opaque way, they will now do it comprehensively. The XAI AI models will be nourished by the data published in X more than ever, and in turn x will benefit from the improvements that are implemented in models such as Grok 3 and that are available from the platform. The overcoming always on the horizon. We have been talking about A potential xand that fusion of Xai with X makes this project somewhat closer. It already has covered the creation of content and the social network, but many areas are missing, and one above all: to be able to use X as a payment platform, Something to which he has pointed out On more than one occasion. In Xataka | Murphy’s law conspires against Elon Musk: if something has been able to happen to their companies this week, it has happened

We expected a storm this Friday in the northwest, what we did not expect is that it was so strong

The synopsis was clear: intense rainfall They were back To the Levante and the southeast of the Peninsula, but they were not going to last too much. However, Aemet It began to point out the possibility that, then, Another storm arrives For the Atlantic. What has changed in the last hours is that The “extreme traffic lights” just jump: Everything seems to indicate that what is going to arrive is a high impact storm that will bring extremely scrambled time. What will happen? At the atmospheric level, this storm will be developed this Friday and will begin to enter the west peninsular. On Saturday 8 we will see how a very active front will put all the predictions up. The key will be that, despite not being an atmospheric river, it is true that winds (Ábregos) will have a southern component and will be warmer even if they come loaded with moisture. That will boost storm. And at a practical level? Three things: bad sea, intense wind and many rainfall. We talk about wind gusts between 50 and 80 kilometers per hour. With hurricane winds in the country’s high areas. We will also see accumulated “abundant and persistent“With up to 150 liters per square meter throughout the western zone of the Peninsula. In addition, although as a consequence of the temperature rise the snow level will rise, the high areas can receive a good amount. But we have to talk about something else. Yes, it is true that everything indicates that a high impact storm will come (one that, presumably, will be called Jana), but that is not the big problem. The big problem is that it has been raining many days and the ability of the Earth to absorb water will not be very high. That means we will see floods and overflows in many zoans of the country. Something that, As Martín León points outit will become even more worrying when the following storms (which can arrive By the blocking anticyclone of northern Europe that is sending it to us) continue to take rain to already wet areas. And what was that “extreme traffic lights”? It is a way to call EFI (Extreme Forecast Index), the ECMWF model index that measures the rarity of a phenomenon. Well, EFI values ​​for March 8 They will be at maximum And that means that, whatever, this Saturday will be a very adverse day. Image | ECMWF In Xataka | Aemet already knows when the storm that is soaking Spain is over. The problem is that another act comes another

We have centuries contaminating the seas with mercury. We did not expect the solution to be genetic engineering

There is something that have volcanic eruptions, oil combustion, waste incineration, chemical substances or gold extraction: they release mercury in the environment. A mercury that ends up deposited in the waters, transformed into methylmercury for millions of microorganisms, stored in fish and, finally, served in our houses at lunchtime. We have a problem with methylmercury. That is obvious. The problem is that it is very difficult to solve. And that is not forcing ideas in another place. What exactly is methylmercury? Mercury is already a global concern due to its persistence in the environment, its ability to bioaccumulation in ecosystems and their important adverse effects on human health. But methylmercury, the most frequent organic form in the marine environment takes the palm. None of this would be a problem if it were not because, in addition, methylmercury is the form “more toxic and the most easily absorbable For living organisms, since it is highly liposoluble and has a great capacity to fix proteins and also shows a high degree of bioaccumulation. “ Is it dangerous? Well yes. High doses of this compound They are very toxic to the central nervous system and especially “for the brain in development of the fetus and in early childhood.” It can cause “mild behavioral problems, language disorders, memory losses, vision and auditory, learning difficulties and development delays.” And we do nothing to avoid it? Yes, we have tried several approaches. In 2013, governments around the world adopted the Minamata Agreement to try to control the “anthropogenic liberations of mercury and other compounds” derivatives. In fact, in the last decade the European Food Security Authority (and its national equivalents) They have been establishing increasingly strict criteria for food with risk of exposure to these compounds. The problem is that it is not easy to control that release and, for now, we cannot do much more than reduce risks. A solution … original. Now, some Australian scientists say they have discovered a new effective way to clean the methylmercury. The Macquarie University Research Team and the Australian Csiro has managed genetically fruit flies and zebra fish to transform methylmercury into a much less harmful gas that is dispersed in the air. The team has modified the DNA of these two animals to insert variants of bacteria genes that make them create two enzymes that can convert the methylmercury into elementary mercury. In general terms, we could say that they inactivate it. It does not become harmless, but its toxicity and bioaccumulation falls very significantly. Oh really? “It still seems to me that we can use synthetic biology to convert the most harmful form of mercury and evaporate it,” Kate Tepper explainedsynthetic biologist and main author of the article. And, indeed, almost science fiction sounds. It must be said that, obviously, we talk about an investigation in the early development phases and much remains to be checked. However, it is a very interesting result. Very dangerous and very loaded with ethical issues, but very interesting for the development of the genetic engineering of the future. Another thing is that we dare to get so far. Image | John Cameron In Xataka | This researcher has been poisoning a lake with mercury for 15 years to see what happens to the fish living inside: now, he finally has an answer

When will the great event of the American technology and what can we expect

Google has announced the date of Google I/O 2025. The company’s annual conference will be held on May 20 and 21. This means that in just over 90 days we will know all the news that Google has prepared for developers, many of which, surely, will also be of interest to the general public. The event will take place at the Shoreline amphitheater in Mountain View, California. As usual, Losco Pichai should inaugurate the conference with the initial ads. In addition to the Google CEO, the presence of other company leaders, such as those in charge of Google AI divisions, platforms and devices, among others, is expected, although everything is about to be seen. What to expect from Google I/O 2025 Last year, Google I/or was marked by An avalanche of news in artificial intelligence (AI)with the models Gemini 1.5 as great protagonists, next to PROJECT ASTRA, I see, image 3 and Improvements in Android. All this made sense, since the company tried to strengthen its position in the AI ​​race. Google I/or is still a key opportunity for developers to update on “more recent development tools and discover how to innovate and improve their workflow”. But beyond this standard description of the company, what can we really expect? Let’s see it. Although we will have to wait for the event to know the novelties, it is very likely that The AI ​​again occupies a central rolewith new models and improvements for operating systems such as Android and Wear OS. In addition, there could be ads related to virtual reality. Recall that, at the end of last year, Google presented Android XRan operating system for devices Virtual reality (VR), augmented (AR) and mixed (MR). It is also possible that progress arises Project Moohanthe first extended reality glasses that will integrate Android XR. How to follow the Google I/O 2025 Google I/o It is divided into several parts. The first day is the presentation conference, followed by work sessions, workshops and demonstrations that will continue the next day. Some students and developers will be present physically, but everyone can follow the event online. In Xataka, as usual, we will be covering all the news of the event. Here you will find articles about the Google’s most outstanding ads. For now, we just have to wait to see what happens. It is worth mentioning that Google I/or will coincide this year with part of the Microsoft Buildwhere interesting news could also be presented. Images | Google In Xataka | It is no longer necessary to be registered to use chatgpt. And it is another step in OpenAi to be “the new Google”

Health in the United States: What to expect from the Trump Administration

During his previous campaign for the White House, Donald Trump made promises to roll back regulations established by his predecessor, Joe Biden, and he began to execute them from day one. His focus on health matters during this new mandate ranges from regulations aimed at reducing health care coststo the spread of the coronavirus, expansions of the Affordable Care Act and protections against gender discrimination. Some of the big changes expected in the health of the United States: Robert F. Kennedy Jr.: Secretary of Health and Human Services Trump appointed Robert F. Kennedy Jr. as secretary of Health and Human Services. Due to his anti-vaccine stances and public health theories, this position has been questioned by many and 18,000 signatures have been submitted in the Senate to reject his nomination. Despite the criticism, Trump has expressed confidence that Kennedy Jr. will lead health agencies toward “greater transparency and effectiveness in the fight against chronic diseases.” Kennedy advocates for reducing pollutants and protecting environmental health. For this reason, regulations are planned for chemical products used in food and the environment. The now Secretary of Health has reiterated his concern about mental health in the country and assures that it is an “urgent crisis in the country.” For this benefits are expected in the accessibility of psychological care, integrating mental health programs into communities and reducing the overprescription of drugs for mental disorders. United States exit from the WHO The 47th president of the United States ordered that the country begin the process of withdrawing from the World Health Organization (WHO), which he holds responsible for the mishandling of the Covid-19 pandemic. This has been regretted by the organization, who claim that together with the nation they have saved millions of lives around the world. Medicare, Medicaid and drug prices The mogul also revoked an executive order that prompted the Center for Medicare and Medicaid Innovation to create three drug pricing experiments that have yet to fully take off. This would affect these three areas: Payment for cell and genetic therapies Pay less for drugs that receive accelerated approvals from the Food and Drug Administration (FDA). Offer generic medications for common chronic diseases for a flat $2 copay. Covid: treatments and vaccines Trump reversed several Biden policies aimed at tackling the Covid-19 pandemic and preparing the country for future infectious disease outbreaks. All were issued in 2021 amid the crisis and aimed to boost Covid-19 testing, treatment research, provide economic relief and improve cooperation with other countries in the event of another pandemic. Affordable Care Act Trump rescinded Biden’s executive order that provided longer enrollment periods for Affordable Care Act plans in most states and additional funding for third parties that help people enroll in ACA insurance. Immigration A report by cnn highlights that Trump’s promise to implement drastic measures against immigration “comes at a time when the United States desperately needs more workers to care for the growing ranks of senior citizens.” “We have a group of older adults who will live longer than ever and will need and require access to care and services,” said Nicole Howell, director of workforce policy at LeadingAge, which represents more than 5,400 people-services nonprofits. seniors, including nursing homes and home care providers. Sex and gender During his inaugural address he revealed a series of health and education policies that establish protections for LGBTQ+ people and transgender people in particular. “From now on the official policy of the United States government will be that there are only two genders, male and female,” he said. Among the orders rescinded this Monday include a 2022 policy against discrimination against people based on sexual or gender identity that mandated support for LGBTQ+ students and ordered officials to end programs that promote so-called conversion therapy both nationally and internationally. Likewise, Biden’s orders to promote protections based on sex and gender identity in schools as well as the law to establish the Gender Policy Council, initially chaired by Jennifer Klein, an alumna of the administrations of the Obama and Clinton. Keep reading:

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.