Month: May 2026

The horror movie of the summer is ‘Backrooms’, and its origin is so surprising that there is a rumor that its director is not real

by

‘Backrooms’ premieres today in the United States (it arrives in Spain on June 5). The film, produced by the unstoppable indie A24is expected to be the horror bomb of the summer (hand in hand with the already tremendous ‘Obsession’, which is putting its hand in the face of ‘The Mandalorian and Grogu‘). It has 87% positive reviews on Rotten Tomatoes and is expected to open between $45 and $50 million, which would be the biggest debut in the studio’s history. Of course, it faces an unexpected controversy: there are those who say that its director, the very young Kane Parsons, has not really directed the film. What has happened? Days before the premiere, this unexpected rumor has circulated online: a more experienced director would have been working from the shadows. Osgood Perkins, producer of the project and director of the great ‘Longlegs‘. Mark Duplass, who stars in the film alongside Chiwetel Ejiofor, Renate Reinsve and Finn Bennett, responded in X: “I don’t remember seeing you on set. When I was there, Kane was 100% in control. More than many directors three times his age.” The origin of ‘Backrooms’. As we already explained in detailon May 12, 2019, an anonymous user posted on /x/, the paranormal board on 4chan, a photograph without a signature or context. It looked like a kind of abandoned office: yellowish carpet and walls, fluorescent lighting… It was ridiculously disturbing. The next day someone added a description that spoke of “not clipping out of reality” (a term taken from a glitch of video games in which the player falls into a geometric void beyond the mapping), and ending up trapped in a space that extends infinitely. The backrooms They are an extreme version of what the internet calls liminal spaces: hotel hallways at three in the morning, empty waiting rooms, closed shopping centers, underground parking lots without cars… Recognizable places but stripped of their function and of the people who normally inhabit them. Just like has been explainedthese types of environments activate the same response as the phenomenon of uncanny valleybut applied to physical places. The brain identifies these spaces as known and at the same time does not know how to read them logically. Jump to the cinema. Kane Parsons was 16 years old when he posted his The Backrooms (Found Footage): nine minutes in first person with a VHS filter, in which someone was chased by a strange presence in one of these spaces. The series that followed this first video, full of secret research institutes and dimensional experiments in the eighties, exceeded 197 million views. A24 bought the rights a year later. Youth, divine treasure. One of the apparent hooks of the A24 film, the extreme youth of its director, has worked against it. The press has underlined Parsons’ youthand some conspiracy theorists consider it to be a marketing strategy. In reality, what this talks about is the current situation in Hollywood, which has produced franchised cinema for two decades in which the director is, fundamentally, a technical executor under the creative supervision of the studio. The system of great sagas has normalized the idea that a good film cannot come from the criteria of a single person, a young person without credentials. We viewers are distrustful because that is what the industrial cinema of recent years has taught us. The explanation. Parsons was born in 2005, the year YouTube launched. “YouTube, more than a cultural reference for me, has been the way I know how to do everything I know how to do,” declared. Parsons doesn’t have the kind of resume that the traditional production circuit demands, but rather his only credential is a massive audience of followers who have been reacting to his work in real time for three years. And that is capable of arousing the suspicions of anyone who is buried by the industrial machine logic of modern Hollywood. In Xataka | When a town found a dead whale on its beaches, it decided to dynamite it. 55 years later they still celebrate it

The DGT has never imposed so many penalties for drugged driving

by

I’m going to make a confession: I’ve gotten hooked on Road Control. One day, YouTube put a 15-minute short video in front of my eyes. The algorithm hit home because right now I am bingeing on a product where the novelty is almost non-existent and is repeated like the worst fast food. And the surprise with each video is: none. Time and time again we attend a breathalyzer test where one, two or several drivers are “caught” under the influence of alcohol. “Have you had anything else,” the voice of the agent on duty sounds already tired. Then, drug controlnew positive and more problems for the unwary. If the premise has not yet convinced you, I will tell you that from time to time there are new features. Some driver has caused an accident and fled. Another skips the breathalyzer test and tries to flee. But the end is usually the same: positive for alcohol and/or drugs. Does it sound repetitive? Yes. But it is. At least that’s what the DGT data says, that in 2025 drug tests were carried out on more than 144,000 drivers. Of them, almost half took home a fine for having consumed some type of substance. More than 70,000 complaints The figure is rescued by our colleagues from Motorpassion. Last year the DGT punished 70,717 people for having tested positive in an anti-drug test. The figure comes in response to a parliamentary question from the Popular Party. It detailed that last year 144,346 roadside drug tests were carried out. That is, almost 50% of the people who faced these controls tested positive. The figure is much higher than in previous years. In 2025, 122,938 tests were carried out and then 64,314 positives were detected. Control on the roads has been increasing, they point out in Europa Press. According to the information detailed by the DGT, the data in the last five years are as follows: Year 2021: 123,211 tests and 41,067 sanctions. Year 2022: 58,126 tests and 42,103 sanctions. Year 2023: 101,927 tests and 50,002 sanctions. Year 2024: 122,938 tests and 64,314 sanctions. Year 2025: 144,346 tests and 70,717 sanctions. The fine for testing positive In a drug test it is 1,000 euros and the subtraction of six points on the driving license. In addition, the driver faces the withdrawal of his license from one to four years and a prison sentence of three to six months, a fine of six to 12 months and community service of 31 to 90 days if he commits a crime against traffic safety, is a repeat offender or if he has been involved in an accident. It must also be remembered that in a drug test, unlike an alcohol test, the amount of substance present in the body is not taken into account. If the control detects that it is present, the driver is sanctioned. This is important to take into account since some substances leave traces in the blood for days after being consumed. Photo | DGT In Xataka | Drunk driving is not enough to “arrest” someone: the Constitutional Court acquits a woman despite testing positive

now it goes against its users

by

Imagine enjoying all the football and premieres for just 40 euros a year, with the peace of mind that a sophisticated system and payments in cryptocurrencies make you completely invisible. That tranquility has just vanished suddenly for many users of the Cinemagoal network who has been dismantled by the Italian authorities. The biggest risk is no longer that the screen goes black in the middle of the game, but rather receiving a notification in your mailbox with a fine of up to four figures. what has happened. Through the baptized as Operation “Tutto Chiaro”, The Italian Guardia di Finanza has managed to dismantle the technological infrastructure behind Cinemagoal. More than a hundred searches have been carried out in Italy and key servers located in France and Germany have been seized. But what is striking about the case is that they are not only persecuting those responsible for the network, but they are tracking and identifying the subscribers of the service, who are receiving fines ranging from 154 to 5,000 euros. Why is it important. Italy is one of the toughest countries against the broadcast of matches and content without a license; It is not content with dismantling the infrastructure, but is going after the final link that feeds the business: the users. It’s a war they’ve been fighting for years, even leading to prison sentences. Although in Spain it has not reached that point, LaLiga has followed in the footsteps of Italy with the massive blocking of IPsalso affecting legitimate companies and services that have nothing to do with with the distribution of unauthorized content. This is how Cinemagoal worked. The network did not transmit video via IPTV, but used virtual machines that operated 24/7, maintaining open sessions on services such as Netflix, DAZN, Sky, Disney+ or Spotify through legal subscriptions registered in the name of false identities. Instead of copying and broadcasting the audiovisual content, these machines extracted the keys or authorization tokens from the official sessions every three minutes and sent them in real time to the application installed on the clients’ devices. Through this system, users downloaded the video directly from the platforms’ official servers, which completely hidden their IP addresses and made it Piracy Shieldthe Italian platform against the dissemination of unauthorized content, was blind. The service cost between 40 and 130 euros per year and was distributed by more than 70 resellers in Italy, who paid preferably in cryptocurrencies or through foreign accounts to evade tracking. For the users. Cinemagoal customers believed they were protected by anonymity, but the police have managed to identify at least 1,000 users by cross-referencing data. To find them, investigators analyzed customer records seized from more than 70 resellers operating in Italy, tracked payment histories and activity logs (logs) hosted on the application’s own servers. The minimum fine is 154 euros, which will be for users who were simply viewing content. Those who had the “fictitious subscriptions from which the authentication tokens were extracted”, that is, those who redistributed the content, will have to pay up to 5,000 euros. 300 million. The Italian authorities estimate that the damage caused by this platform reaches 300 million euros in subscriptions. It’s what It is known as lost profits, That is, the money that platforms stop earning due to unauthorized content. However, their way of doing the math assumes that the hacking user would have paid 100% if they did not have the illegal option: it is a fallacy and an unrealistic metric of losses. Image | Xataka with Gemini In Xataka | LaLiga wanted to fine VPNs that did not block IPs during matches. A court has been set up

We have found a hidden “switch” of Alzheimer’s. And the best thing is that we have promising candidates to put it out

by

Alzheimer’s is still one of the biggest medical challenges of our century, since we are facing a disease with a very important incidence and above all that entails a large number of social problems around it. Here research over the decade has focused on the accumulation of protein plaques beta-amyloid in the brain to explain it. However, the scientific community has begun to pay much more attention to an equally devastating factor: neuroinflammation. A new gene. Science continues to advance and one of the latest discoveries that has been made lies in the APOE4 genewhich is a known risk factor for Alzheimer’s disease. And it is no wonder, since people who inherit this variant have a much higher probability of developing the disease, and often do so at younger ages. But now a research team has been investigating exactly why having this genetic variant predisposes one to Alzheimer’s, and the answer appears to lie in chronic inflammation. More specifically, in APOE4 carriersthe brain’s immune system overreacts, creating a toxic environment that damages neurons and accelerates cognitive decline. And at the center of this inflammatory storm, researchers have indicated to the enzyme cPLA2 as the main culprit. It’s a challenge. Knowing that cPLA2 plays a crucial role in the inflammatory cascade associated with Alzheimer’s, the objective is logically set turn it off permanently. However, inhibiting enzymes in the brain is not an easy task, since the brain is very well protected by the blood-brain barrier, which acts as a true customs control that allows only some very selected elements to pass through. That is why creating a drug that passes through it without causing side effects in other parts of the body is a great challenge. The strategies. To reach this goal, science is now doing computer simulations of thousands of molecules to be able to find those with the exact shape and properties to “fit” into the cPLA2 enzyme and deactivate it. Once this ‘key’ that fits the enzyme that looks like a lock is identified, candidate compounds can be refined for testing in animal models. Until now, research already has several selective cPLA2 inhibitors that have proven to be powerful and capable of penetrating the brain, making it possible to reduce neuroinflammation in the models studied. Personalized medicine. The study, supported by multiple leading institutions such as the National Institute on Aging and the Alzheimer’s Drug Discovery Foundation, is not only relevant for the design of the new drugs, but also for its personalized medicine approach. Looking back, clinical trials for Alzheimer’s have treated all patients equally, often resulting in million-dollar failures. But now, by targeting these new cPLA2 inhibitors specifically at neuroinflammation fueled by the APOE4 gene, scientists are creating tailored treatments for the most biologically vulnerable patients. Although we are still in a very early phase of research, it may take years to see a tangible result. Images | Robina Weermeijer In Xataka | Alzheimer’s no longer seems irreversible: science allows brains with advanced damage to recover for the first time in animals

Metro de Madrid has insisted on being a pop icon like London’s “Tube.” His plan: retro football jerseys

by

There are cities that are recognized by their metro before you can place their flag. The iconic roundel red and blue Tube London has jumped from the tunnels and platforms to the closets of half the world, becoming a global emblem that people wear with pride. It is not a simple transportation logo: it is the city itself compressed in a circle. The Madrid Metro knows this well, which has been looking at its British counterpart for some time with the envy of someone who aspires to transcend their status as a mere public service. And he has just made his boldest move to close the gap: launching his own collection of retro football shirts. Nostalgia reaches underground. With its sights set on the 2026 FIFA World Cup, the Community of Madrid has launched Metro FC: five retro t-shirts inspired by world champion teams, at 54.95 euros each, sold in the Metro online store and at the Ópera, Sol and Plaza de Castilla stations. Each garment bears the Metro’s own embroidered shield – based on the historic rhombus of 1921 – and the seven stars of the Community, with a color that varies according to the World Cup of each honored team. Line 2 is the Spain of 2010; Line 1, Maradona’s Argentina in ’86; Line 3, Romario’s Brazil in ’94; Line 5, the Germany of Italy 90; and Line 10, Zidane’s France in ’98. The question that closes the campaign sums it all up: “And you, which line are you from?” It’s not just merchandising. It is a declaration of intentions about what the Madrid Metro wants to be. The model that Madrid envies. To understand what the Madrid Metro is after, you have to look at London, where this path has been taken for more than a century. He roundel —the red circle with a blue bar that identifies the London Underground— was born in 1908 as a simple stop sign. Today it is one of the most recognizable and emulated commercial symbols ever designed. The architect of that transformation was Frank Pickhead of communications for the Underground in the first half of the 20th century, whose philosophy was as simple as it was ambitious: design is not something optional that appears here and there. Design must enter everywhere. The result of this obsession is today an unparalleled brand machine in global public transport. According to a Nielsen studyhe roundel Londoner is more recognizable than Mickey Mouse, and the phrase “Mind the gap” has become one of the most identifying sounds of the city. The Underground generates more than five million pounds a year on merchandising alone, and carries more passengers during the working week than all other trains in Britain combined. The conquest of streetwear. It is no longer so much the scale, but rather who you sit down to negotiate with. His recent collaborations include Adidas, Arsenal, Prada Linea Rossa, Kurt Geiger and some Nike sneakers with the fabric moquette of the meter that are resold online for around 400 pounds. In 2026 it has launched its third collection with Uniqlowith t-shirts and bags decorated with iconic London transport graphics, celebrating the city’s character and heritage. And before, in 2023 he created a line of streetwear complete —London Underground Studio—in collaboration with the South Korean brand Handsome, from the Hyundai group, with leather bombers and knitted sweaters sold in Seoul. The symbol has transcended its signaling function to become shorthand for the city itself, inspiring fashion and pop music. It is, in the words of its scholars, “cool” in a way that public bodies rarely achieve. The Madrid Metro knows exactly what it wants. And the Metro FC shirts are their first big attempt to achieve this. The trend that opens the door. The Metro movement does not emerge from nowhere: it intelligently takes advantage of the exact moment when the football shirt has become the object of cultural desire of the decade. You just have to walk through any big city to find the image: dozens of young people walk around wearing Real Madrid t-shirts from the 2000s, the 90s Juve or the Japanese national team. They don’t go to the stadium, it’s not even match day. The border between catwalk fashion and sporting passion has been completely blurred. The phenomenon has a name: Soccercore either Blokecore. It takes its name from “bloke” – the common type of British working class of the 80s and 90s – and consists of combining vintage football shirts with baggy jeans or classic sneakers. What began to go viral on TikTok in 2021 has ended up on the Balenciaga catwalks and on the bodies of Bella Hadid, Kim Kardashian or Jennie from Blackpink. Brands like Etro or Stella McCartney have taken the style from the stadiums to the catwalks, and Loewe is in charge of dressing the Spanish team in this World Cup. The business of belonging. It is precisely in that context where the Metro’s commitment takes on all its meaning. The journalist Alejandro Mendo, from his substack Pieces of Fabricidentifies the key movement that explains it: the rise of institutional “white label” clothing. Clubs and institutions have discovered that they can produce their own collections, without large multinationals involved, that cost around 50 or 60 euros and that generate something that no external sponsor can buy: brand loyalty. Metro de Madrid does not sell t-shirts, but rather it sells belonging. The narrative of big brands has also changed in this sense. Mendo points out how brands like Adidas have put aside the muscular epic to sit your stars in chairs during their campaigns. Putting figures like Lamine Yamal or Jude Bellingham in a relaxed and introspective attitude turns them into cultural icons rather than gladiators on the grass. The footballer, like the shirt he wears, has stopped being just an athlete and has become a generational reference. A material anchor in the society of “non-things.” There is something else, however, that explains why these garments work … Read more

The fires have already grown by 218% so far this year and summer has not yet arrived

by

While announcing “the largest deployment of the State”, the Government of Spain has given a disturbing piece of information: the number of fires has skyrocketed by 218% so far this year. And yes, May isn’t over yet. The fine print, however, is interesting. The data, as I say, refers to the number of reported fires, but does not directly correspond to the burned area. In fact, despite to the enormous ‘boom’ of fire outbreaksthe burned land is still below the average of the last decade. In this sense, what is truly interesting is the paradoxThat with reservoirs at historic highs and no signs predicting an upcoming drought, the risk of fire has not stopped. In fact, it has skyrocketed. Clarifying the data on the fire boom. Indeed, between January 1 and May 15, 2026, 127 fires were reported, compared to 40 in the same period in 2025. That is a growth of 218%. And it’s true that “tripling” the fires sounds like a lot: but of those 127 fires, only three were large forest fires and only six required major intervention. The key fact, as we can see, is none of that. During the quarter, 12,946.66 hectares have burned; that is, 2.2 times more than in the same period of 2025 (5,822.12). But it is still 29.6 less than the average for the decade. The key fact is that we have improved a lot in preparing for and putting out fires, but this year the situation is very complicated. The three ways we have of counting fires (Civil Protection, MITECO and EFFIS/Copernicus) say that the year is getting complicated at a forced pace. Above all, because 2025 was a very bad year: three times as many hectares as average were burned. Where is really the problem? In the concentration of damage. According to Greenpeace, less than 1% of fires They already concentrate 86% of the surface burned and the average size of the large fire has gone from 1,500 hectares to more than 6,000. In this context, having more fires means having a greater chance of one of them becoming a superfire. And the countdown has already begun: the fire season is at the door and, despite the grandiose declarations of the administrations, we are not prepared. Image | Marcus Kauffman In Xataka | The satellite that detects fires before firefighters has a problem: it has to avoid space debris and is leaving blind spots on the map

Blue Origin had a plan of 12 launches for this year. A fireball at Cape Canaveral just changed everything

by

Bad news for Blue Origin, Jeff Bezos’ space company. And his New Glenn rocket It exploded this morning into a huge fireball while conducting a ground test at Cape Canaveral, Florida. The accident, which fortunately left no injuries, is a good blow for the company in his race to compete with SpaceXjust when this was going to be his definitive breakout year. That has passed. Around 9:00 p.m. local time (3:00 a.m. on Friday on the Spanish peninsula), the New Glenn exploded during a ‘hotfire’, a test in which the rocket engines are turned on while the vehicle remains anchored to the platform, without taking off. The objective of this test is to check the operation of the engines before a launch. Blue Origin itself He spoke on his X account of an “anomaly” and confirmed that all personnel were located and safe. According to collect The Guardian, the fireball destroyed the platform and the orange glow was seen more than 180 kilometers away, while residents of nearby towns noticed tremors in their homes. A year that was going to be the year of takeoff. The blow is especially hard because of the moment it arrives. Blue Origin had marked 2026 as the year to finally gain pace. Its CEO, Dave Limp, even stated in an interview with Ars Technica that the company could reach double digits in launches this year, until matching its production rate of 12 rockets, and even considering reaching 24 if manufacturing continued to improve. They also mentioned the 12 launches in their request to the FAA to operate from Cape Canaveral. The problem is that it was more of an ambitious goal than a realistic forecast, since the New Glenn has started the year without having flown again since November and experiencing several setbacks. The explosion has now turned that goal into a chimera. Bezos’ reaction. The founder of Blue Origin took the drama out of the matter, counting in Elon Musk also reacted to the event briefly: “Very unfortunate. Rockets are difficult.” Why it is important. The New Glenn It is the key piece with which the company wants to confront the dominance of SpaceX, and it is also called to play a central role in NASA’s Artemis programwhich seeks to return astronauts to the Moon. Just a few days before the explosion, the agency had awarded Blue Origin a contract to participate in the construction of a lunar base. The moment could not have been worse. A streak of setbacks. Blue Origin has accumulated a series of catastrophic misfortunes. On its third flight, in April, the rocket managed to land its reusable booster on a barge at sea, but its upper stage failed and failed to place the satellite it was transporting for AST SpaceMobile into orbit, which ended up falling and disintegrating in the atmosphere. That failure sparked an investigation by the FAA, the US air regulator, which just last week had given the rocket the green light to fly again. Thursday’s test was precisely the preparation of its fourth mission, in which it was going to deploy satellites of the network Leo from Amazona direct competitor to SpaceX’s Starlink. Amazon clarified that none of those satellites were on board at the time of the explosion. Damage assessment. Both the FAA and NASA spoke out quickly. The regulator pointed out that the test was outside the activities it licenses and that it did not affect air traffic. NASA Administrator Jared Isaacman, on the other hand counted that “spaceflight is unforgiving, and developing a new heavy-lift capability is extraordinarily difficult.” The agency promised to support a thorough investigation and, above all, to evaluate how what happened affects its lunar programs. And now what. What we will now see is how Blue Origin rewrites its calendar. NASA was counting on New Glenn to launch the first missions to its lunar base this year, and the agency itself has acknowledged that they still do not know how this accident will affect the mission with Artemis. On the other hand, SpaceX has its own problems with the Starship, also under review by the FAAwhile preparing a historic IPO. The terrain is quite hot. Cover image | NASA Space Flight In Xataka | SpaceX seemed unreachable in its race to the Moon. Blue Origin is proving that anything is possible

Anthropic just surpassed OpenAI as the world’s most valuable AI startup

by

Anthropic is no longer the eternal second fiddle. The company that was always in the shadow of OpenAI has become the main protagonist of this segment in recent months. Its growth is so spectacular that in its latest round of financing it has managed to surpass OpenAI’s valuation. It is an extraordinary milestone, especially for one reason: both hope to go public before the end of the year, and here Anthropic has the upper hand (again). Overtaking on the right. The company founded by the Amodei brothers has raised a colossal financing round of 65 billion dollarsand with it Anthropic’s valuation becomes 965,000 million post money. It is a financial achievement that suddenly destroys OpenAI’s valuation, which is currently stuck at $730 billion. This latest round comes just three months after Anthropic will raise 30,000 million of dollars, cccadadasdsas in an agreement that placed its valuation at 350,000 million dollars. The growth is simply amazing. Anthropic is the coolest company. The valuation reflects a compelling reality: Anthropic is (much) more fashionable than OpenAI. The company has taken great advantage of recent controversies to increase its popularity, and its brand image has been greatly reinforced because it is the company that everyone is talking about. What happened to the Pentagon first and what has happened with the encyclical Magnificent Humanitas of the Pope then they show it. And the one with the best models (seems) to have. OpenAI seemed to be ahead in the AI ​​race with models leading the way. That changed with the arrival of Claude Code and Claude Opus 4.5. Since then, Anthropic’s advances have been striking, and although the differences are small, the popular perception is that Claude Opus is now the model that leads in performance. This has just been confirmed in benchmarks with the recent release of Claude Opus 4.8but above all with Claude Mythos Previewthe model that has been put the world of cybersecurity upside down. They already make money. A few days ago, surprising news leaked: Anthropic could close the second quarter of the year with an operating profit of 559 million dollars. He would make money when the rest of his rivals lose a lot. The projected annual turnover has managed to exceed $47 billion this month, five times more than the amount estimated at the beginning of the year. The reason: the overwhelming success of Anthropic models in companies. That’s where the money isand the company has known how to 1) detect and 2) take advantage of it before anyone else. Memory manufacturers enter the round. The financing round is led by venture capital firms such as Greenoaks, Sequoia, Altimeter and Dragoneer, but this time there are other protagonists. These are the semiconductor firms Samsung, Micron and SK Hynixwho have also participated and who have taken advantage of their current privileged position to also bet on the success of Anthropic. It’s a win-win: they bet on the current winning horse, and Anthropic manages to strengthen relationships with the companies that right now they control one of the big bottlenecks of the AI ​​industry: memory chips. The IPO is imminent. This surprise meteoric intensifies the pressure on OpenAI and further encourages (if that was possible) that other race, which is the IPO of both these two companies and SpaceX. We are in a year that will be remembered for three stratospheric IPOs, but these latest achievements by Anthropic have made the company led by Dario Amodei now the main protagonist in the technology segment. Image | Fortune Brainstorm Tech In Xataka | The surprise of the new Claude Opus 4.8 is not that it is (a little) better. The surprise is the “I only know that I know nothing”

list of new features of the new version of Anthropic’s Artificial Intelligence model

by

Let’s tell you What’s new in Claude 4.8 Opus, the new version of Anthropic’s most advanced and powerful artificial intelligence model. This version has surprised us by arriving just 41 days after Claude Opus 4.7, and it seems that the improvements are minimal, but there is a really important change in its honesty when it comes to telling you if it doesn’t know something. In any case, here you have a complete list with all the new features that come with this new version of Claude 4.8 Opus. We are going to explain each of them briefly so that they are easy to understand. Another thing you should know is that Opus is the most advanced line of Claude models, the one indicated for more complex tasks for programming and the one that uses up your limits the fastest when you use it. There is also the most efficient Sonnet model for day-to-day tasks, which continues in version 4.6 since February 2026, and a Haiku for quick and simple questions that continues in version 4.5 since October 2025. News from Claude 4.8 Opus A more honest AI: The prominence of this new version goes to honesty. He’s significantly more honest about his own work, telling you when he’s unsure about something. It’s also about four times less likely to let bugs in code slip by without flagging them, compared to its predecessor. Performance improvements: The Agentic code score for creating code with agents increases from 64.3% to 69.2%, and multidisciplinary reasoning with tools increases from 54.7% to 57.9%. On other test benchesin the SWE-bench Verified it goes from 87.6% (Opus 4.7) to 88.6%, and in Terminal-Bench 2.1 it rises from 66.1% to 74.6%. GPT-5.5 still falls short in terminal/CLI workflows, although there have been big improvements in Claude, and both models are practically on par in web browsing and graduate-level science topics. Alignment improvements: Alignment assessments show new highs in prosocial traits such as supporting user autonomy and acting in their best interest. Rates of misaligned behavior such as cheating are lower than in Opus 4.7. Fewer hallucinations: As usual, the number of hallucinations is also reduced. Honesty when telling yourself when you don’t know something also helps reduce them. Quick mode: According to AnthropicOpus 4.8’s fast mode is now about 2.5 times faster. The company claims that the improved Quick Mode also costs three times less than before. Effort control– Users can choose between “extra” or “max” levels so that the model spends more tokens and obtains better results. Dynamic Workflows (preview for research): With this new feature, Claude can schedule work and run hundreds of subagents in parallel in a single session, being able to complete codebase-scale migrations of hundreds of thousands of lines. Available for Claude Code on Enterprise, Team and Max plans. No change in base price: The base price of API tokens is unchanged from Opus 4.7. It is 5 dollars per million input tokens, and 25 dollars per million output, with up to 90% savings with prompt caching and 50% with batch processing. In Xataka Basics | How to prevent AI from always being right by default and thus make Claude, Gemini and ChatGPT have fewer hallucinations

No missiles, no rifles, no bombs. Ukrainian drones are carrying a type of cargo unprecedented in war: elderly people

by

During the Soviet blockade of Berlin in 1948, an American pilot began to throw chocolates tied to small cloth parachutes on children watching the planes from Tempelhof airport. That improvised initiative ended up becoming the famous “Operation Little Vittles“, one of the most unexpected images of the Cold War: military aircraft used to carry hope instead of weapons. Decades later, Ukraine is finding equally unusual uses for its war machines. Lifesaving robots. For years, unmanned vehicles were associated with a very specific idea: transporting weapons, ammunition or explosives where the risk for soldiers was too high. The war in Ukraine is expanding that definition with an image that would have seemed unthinkable just a few years ago. In some of the most dangerous sectors of the front, the same ground drones that are part of the war machinery are being used to evacuate elderly people trapped between bombings, mines and artillery fire. In a conflict marked by the automation of combat, one of the most unexpected loads carried by these vehicles are not projectiles or supplies, but old people who no longer have a safe way to leave their homes. Rescue through no man’s land. The last known operation took place near Limánin the Donetsk region. While carrying out a logistics mission, a ground drone unit from the Kraken group was approached by a woman who asked for help to leave the area along with three other people, one of them injured by shrapnel. After coordinating the procedure for days, the operators sent a Zmiy Logistic vehiclea kind of remote-controlled four-wheeled buggy capable of transporting up to about 500 kilos of cargo. The drone traveled about 16 kilometers to the agreed point, rpicked up the four evacuees and began the return journey to a river crossing where Ukrainian soldiers completed the rescue and took the wounded to a hospital. The impossible life in the gray zone. These rescues They show a less visible reality of war. Despite years of fighting, there are still civilians living in the so-called “gray zone”, a strip of land disputed between both armies that can reach between 16 and 20 kilometers wide. There are practically no public services, shops, schools or hospitals left there. Power outages are common and bombings are part of the daily routine. However, many older people continue to resist in those places because they don’t want to leave the houses where they have lived all their lives, because they care for sick relatives or because they hope that the war will end before being forced to leave permanently. Iron soldiers on a new mission. It is not an isolated case. They remembered in Insider that in early April, another 77-year-old Ukrainian woman was evacuated from the same area using a ground drone operated by the 60th Mechanized Brigade. The images They went around the world because the soldiers approached her with a blanket on which a message as simple as it was revealing could be read: “Grandma, get on.” The scene summarizes the extent to which these systems are evolving. Originally designed to transport supplies, plant explosives or even assemble remote weaponry, the so-called “iron soldiers” are beginning to take on rescue tasks that previously would have required exposing soldiers or volunteers to extreme danger. Total automation. Behind these stories there is a much deeper transformation. Ukraine and Russia are accelerating the incorporation of unmanned ground vehicles to carry out missions that They are too risky for people. Some carry ammunition, some carry medical supplies, and some incorporate remote-controlled weapons. The Ukrainian goal is especially ambitious: Minister of Digital Transformation, Mykhailo Fedorov, has announced the purchase of 25,000 ground drones during the first half of 2026 and aspires for all frontline logistics to one day depend on these systems. During the first quarter of the year alone, unmanned vehicles performed more than 21,500 missions. Unexpected consequences. The usual image of military innovation may be associated with increasingly destructive systems, but the Ukrainian experience is showing an unexpected consequence of that technological revolution. The same robots that were born to keep soldiers away from danger are being used to remove vulnerable civilians from some of the most dangerous places in Europe. As militaries race to automate combat, ground-based drones are proving military technology can play a role, too completely different: become the ultimate escape vehicle for those trapped in the ruins of an endless war. Image | ArmyInform In Xataka | Storks have become the best anti-drone weapon of war. And Russia and Ukraine are taking note In Xataka | Ukraine has been terrorizing Russian soldiers with its heavy drones for years. Now they are literally giving it back.

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.