Human

Where is really the limit of the human being?

by

April 26, 2026 will go down in history: Sabastian Sawe won the London marathon and became the first human being to go under two hours in an approved race. What’s more, in that same race Yomif Kejelcha and Jacob Kiplimo They also had better times than the previous world record (although only Sawe and Kejelcha went under two hours). Therefore, the key question is no longer When will we be able to get under two hours in the marathon? and it will start to be “where is the physiological limit of the human body really?” The question is essential, yes; but we cannot approach it naively. There are two factors without which Sawe would not have been able to achieve the record: the sneakers (which, although they are from Adidas, are part of the revolution that promoted Nike a few years ago) and nutrition (worked with Swedish nutrition specialists Maurten for 12 months to design a specific provisioning protocol). If the contemporary quest to break the two-hour mark has shown anything, it is that running goes far beyond a physiological issue. What variables rule in the physiological section? Since 1991, the physiology of the marathon is usually understood following a simple model that said Michael Joyner of the Mayo Clinic.. According to Joyner, sustained marathon pace depends on three variables: The maximum amount of oxygen that the body can absorb per minute. The maximum amount you can sustain for hours without accumulating lactate faster than it is eliminated. The energy cost of maintaining a given speed. Joyner, who lived in a world where the record was 2:06:50, theorized that the limit It should have been around 1:57:58. That’s not enough. Years later, Andy Jones from the University of Exeter added one more factor: physiological resilience. That is, the ability for these three variables to deteriorate as little as possible during two hours of racing. For this reason, shoes and supplies are essential: They are two tools that improve efficiency and resilience. In fact, many experts maintain that the 2016 revolution is a break in the series and the records are “technological, not physiological“. This is important because, using models of the Joyner variables, we can make conjectures of the physiological limit “as long as the same technological conditions are maintained.” And what is the limit right now? Following the real values ​​of the Breaking2 projectsubtracting resilience and considering technological and nutritional improvements, the realistic limit would be between 1:55 and 1:57. Below 1:55 we would need an athlete with physiological capabilities that we have not seen yet. It’s not impossible, but it’s very unlikely. Image | Miguel Amutio In Xataka | More and more people participate in popular marathons. Science knows that being overly optimistic has its risks

How the nutrition revolution has allowed human beings to lose two hours in the marathon

by

“Today I had two slices of bread, honey and tea for breakfast” This was Sabastian Sawe’s brief response to the journalist who had answered what he had had for breakfast that morning. It wasn’t an elevator conversation. It was not pure curiosity among office colleagues before facing a day in front of the computer. Although the answer could very well have been the same. Sawe is not just anyone. Sawe is the fastest man on the planet when it comes to covering 42,195 meters. Sawe is, as of yesterday, the winner of the London Marathon, the recordman of distance and the first person to break down the mythical barrier of two hours in distance. A barrier that was considered unattainable just a handful of years ago. But to round off the accumulation of impossible things that were experienced yesterday in London, Sawe is not the only human to run at a sustained pace of 2’49″/km, repeating the effort up to 42 times and an agonizing and endless 195 metres. Yomif Kejelcha entered just 11 seconds later. Of course, he will have the more than deserved consolation of being the fastest debutant in the history of the distance. Jacob Kiplimo must have been astonished when he realized that the 120 minutes and 28 seconds it took him to cover that same distance had only been enough for him to be third when just two hours earlier they would have made him the new world record holder. All of them are the pure embodiment of a sport that has experienced a revolution in a handful of years. “Yomif didn’t have breakfast on race day,” Alfonso Beltrá, CEO of Holy mother. Beltrá is the founder of a Spanish sports nutrition company, which grew up in cycling but has pivoted to athletics. A few weeks ago, Beltrá himself defended on the Find Your Everest podcast that they were sure that they could be the ones to break the mythical barrier. High competition is the result of research and results that put the athlete in front of a challenge that ends up being decided by details. The differences are minimal and nutrition is what has ended up tipping the balance. A revolution to make the impossible possible Although neither Sawe nor Kejelcha ate a copious breakfast as the most undocumented logic dictates, neither of them stood at the starting line empty. Their stomachs were, but the important thing was that their glycogen stores were bursting with energy. “He doesn’t have breakfast at all on race day, we have a specific drink with 100 grams of carbohydrates that he finishes two hours before the race and another product that we will launch in three months. Then he warms up and six or seven minutes before leaving he drinks a gel with caffeine and 45 grams of carbohydrates,” Beltrán tells us about the Ethiopian athlete who ate 145 grams of carbohydrates before departure. “That’s why we skipped the kilometer five aid station, we couldn’t put in more,” he emphasizes. A very similar amount would have been handled by Sawe’s team. The Kenyan athlete took, in addition to the two slices of bread with honey, a Drink Mix 320 by Maurtena company that has revolutionized sports nutrition in recent years. This is equivalent to 80 grams of carbohydrates and added a Gel 100 that provides another 25 grams of carbohydrates. The total sum between energy drinks, gel and breakfast adds up to about 140-150 grams like Kejelcha. It is a figure similar to taking 200 grams of pasta before running. And then, the performance began. Kejelcha uses Santamadre gels diluted in the appropriate proportion of water to take them with bottles and facilitate ingestion. Sawe, for her part, opted for Maurten’s Drink Mix 160 drink at the first three aid stations (kilometers 5, 10 and 15). In the fourth (km 20) take a little less drink, about 130 ml, but take a gel with caffeine. And from here, at kilometers 25, 30, 35 and 40, 130 ml is prescribed but of the Drink Mix 320, which has a greater carbohydrate load. It is believed, however, that a drum fell. In total, Sawe’s intake remained at about 115 g/h of carbohydrates. Sabastian Sawe race protocol. Click on the image to go to the original post. The figure is very high and unthinkable just a few years ago. And, despite this, it pales in comparison to what was planned for Kejelcha. From Santamadre they explain to us that the final objective was to move around 140 gr/h of carbohydrates in this case. “It’s about keeping your deposits full for as long as possible, being aware that you will always be in deficit, losing more than you earn,” says Beltrá. In this case, they do confirm that the athlete had problems at kilometers 25 and 35, where he lost the bottles. “That coincides with the decline that they experienced at kilometer 41″, laments the CEO of Santamadre who, he emphasizes, for them a debut like this shows that “we don’t know their ceiling. When he arrived he was upset for losing the race, he asked us for forgiveness for losing the two bottles… he wasn’t aware of what he had done, nor did he know it when he hugged me at the finish line.” In this case, Kejelcha used Santamadre gels diluted in water. The limit is set, according to the Spanish company, by the athlete’s own stomach. “The protocol has to be personalized. We have controlled all the variables since we started working together, all meals, rest, body temperature, glycemic peaks… the nutritional strategy was studied to take into account the glycemic index of each moment,” Beltrá explains to us. Kejelcha was, in fact, the only athlete among the first finishers to use a different brand than Maurten. This brand revolutionized athletics with the sponsorship of Eliud Kipchoge and has been the leader in recent years thanks to its famous Hydrogel. With this gel, the brand managed to encapsulate … Read more

Last year, almost no robots finished the Beijing half marathon. This year one has broken the human world record by seven minutes

by

The half marathon world record is held by Jacob Kiplimo with a time of 57:20 achieved just a month ago in Lisbon. This Sunday a humanoid robot called Lightning ran that distance in 50:26achieving for the first time a milestone that had never been achieved. Robots seemed clumsy and unable to outrun humans, but that is no longer true. And it’s just the beginning. Robots are already faster than humans. In the half marathon held on Sunday, April 19, 2026 in Beijing, the absolute dominators were the humanoid robots. Lightning not only broke the human world record by almost seven minutes: he managed to arrive 17 minutes before the first human runner to cross the finish line. The first three classified They were also Lightning models developed by Honor. From disaster to excellence. The first edition of this same event, the Beijing E-Town Half Marathon and Humanoid Robot Half Marathon, It was an absolute disaster for humanoid robots. Only a third of those who ran it managed to finish the race, they were controlled remotely and ran at a pace much lower than that of human runners. This year things were very different: more than 100 robots were presented and most finished the test, but also almost half ran autonomously and several managed to surpass even the best human runners in the world. This is Lightning. The winning robot measures 169 centimeters, weighs 45 kg and was specifically designed to adapt to complex terrain and move at high speed. Its legs measure about 95 cm and its proportions are designed to imitate the stride of elite human runners. It has a liquid cooling system which curiously has been adapted from the one found on Honor smartphones. Du Xiaodi, engineer in charge of this project at Honor, explained that “Running faster may not seem significant at first glance, but it allows technological transfer, for example in structural reliability and cooling, and eventually in industrial applications“. Not everything went well. The race, however, also had moments in which the robots failed. One of them collided with a nearby vehicle although he managed to stabilize himself and continue walking. The H1 model from Unitree, the most famous humanoid robot manufacturer in China, collapsed as it approached the finish line and had to be removed from the road. One of the Lightning models hit a barrier after crossing the finish line, and some other robots they had difficulties with the curves and unevenness of the route. The event also served as a test bed for batteries, joints, motors and algorithms that control these machines. Industrial applications. Xiaodi mentioned it but also Liu Xiangquan, professor of robotics at the University of Science and Information in Beijing. According to him, these long-distance races allow the resistance and behavior of these robots to be evaluated, something essential for their application in industrial environments. Here not only speed is evaluated, but also the aforementioned resistance, stability or the capacity for autonomous navigation in uncontrolled environments. But a key component is missing. Although the demonstration and milestone is fascinating, what this field needs most is other things. For example, advance manual dexterityperceive the real environment in unforeseen situations and be able to perform varied tasks and not focus so much on repetitive movements. Industrial robots are already good at that, but here we are looking for much more versatility because at the moment these robots They are not able to fold clothes or put the plates and cutlery in the dishwasher with sufficient speed and dexterity. China continues to set the robotic pace. The Asian country has completely devoted itself to the world of robotics. Dominate this segment and its companies They manufacture 80% of global production. In recent months we have seen spectacular demonstrations such as the one Unitree carried out with a dozen humanoid robots at a martial arts show. Sunday’s half marathon is one more element of that narrative and that message that China is leaving to the world: robots are our thing. And in a year, what? Breaking the world record is very striking, but this event tells another story: that of how in just one year Chinese manufacturers have managed to improve their models in an amazing way. If everything continues to improve at this rate, it is difficult to predict what the robots that run the next marathon will be capable of, but it seems logical to think that at this point the athletic ability of robots will be absolutely amazing. Image | CGTN In Xataka | In China they are not satisfied with creating advanced robots: a company has developed a head that gestures like a human

teach AI to sound human

by

In recent months, many of us have spoken to an artificial intelligence without thinking too much about it. We have asked him questions, we have asked him for advice or we have simply tested how far his ability to keep a natural conversationl. Tools like voice modes ChatGPT either Gemini They have brought that experience closer to something that, not so long ago, seemed reserved for science fiction, with inevitable echoes of ‘Her’. But there’s one question we rarely ask ourselves while talking to them: how have these machines learned to sound less and less like a system and more like a person. To understand it, it is convenient to separate what we see from what we do not see. On the one hand there are the applications that we use daily, those assistants that respond with an increasingly natural voice. On the other hand, the systems that support them, models trained with large volumes of data that they need to learn not only what to say, but also how to say it. We do not know which specific products end up using this type of recording, but we do know that they are part of the ecosystem with which increasingly fluid and credible voice systems are trained. The human hand behind an artificial voice When we get down to the details, what these workers do is not very similar to the classic idea of ​​“training an AI.” In many cases, it involves having conversations with strangers about seemingly trivial topics, from everyday tastes to open questions that require you to develop an answer. In others, the assignment is more demanding: playing a role, following a script without seeming like it or enter emotional terrain. Bloomberg accountFor example, the case of a worker who recounted painful memories of her life while speaking with a man who introduced himself as a pastor and who, within the exercise, played the role of therapist. All that recorded material serves a very specific purpose: capturing nuances. We are not just talking about words, but about pauses, breaths, changes in tone, hesitations or emotional reactions that make a conversation sound human. There are also labeling tasks, in which workers have to distinguish whether an audio contains a sob, a laugh, or someone talking between laughs. The underlying logic is simple: if a machine wants to stop sounding robotic, it first needs to be exposed to how we really speak. After passing an initial voice test, they can qualify for tasks that start at about $17 per recorded hour. From there, the question is inevitable: how do you access this type of job and how much do you really earn? Platforms like Babel Audio They function as intermediaries that connect these workers with specific projects. After passing an initial voice test, they can opt for tasks that start at around $17 per recorded hour, although the final income depends on the evaluation received and the volume of orders available. Income also varies greatly: a worker cited by the aforementioned media claims to earn about 600 dollars a week. This is what the BabelAudio website looks like As we progress, the work begins to show a less visible side. Beyond the rates and the promise of flexibility, the testimonies point to an environment marked by uncertainty and constant control. Platforms can limit access to tasks, interrupt projects or suspend accounts without detailed explanations, leaving many workers in a fragile position. In addition, each conversation is subject to real-time metrics that assess whether someone speaks too much or too little, expressiveness, language proficiency, depth of exchange and even the length of pauses. When we broaden the focus, the debate stops being solely work-related and also becomes personal. Part of the value of these recordings lies precisely in the fact that they capture how we speak and how we relate, which implies that workers are providing more than just a mechanical task. The terms generally allow those recordings to be used in voice assistants, speech synthesis, and “other audio-related products and services.” When we connect all the pieces, what we see is an industry that works thanks to a complex production chain. The Pulitzer Center describes This ecosystem is like a fragmented work network in which workers are usually subject to confidentiality agreements, operate with very little transparency and, in many cases, do not even know what system they are training or what company their work ends up going to. In this context, the conversations that feed voice systems are only one part of a larger machine, where each task contributes to building increasingly sophisticated technologies. Images | Xataka with Nano Banana 2 | Screenshot In Xataka | Congratulations, you already program without knowing how to program. Now prepare to wait six weeks for Apple to listen to you

There is a much deeper and more important AI race in which China is crushing its competitors: human talent.

by

The AI ​​race It’s about many things. Not only who makes the best AI modelswho has more and better data centers or who has more cheap energy to power this revolution. It’s also about something that right now China dominates with an iron fist: AI experts. China surpasses the US in talent. In The Economist have analyzed the evolution of the publication of studies at NeurIPS, one of the most important conferences in the world on AI. In the 2025 edition they have discovered a singular fact: for the first time in the history of this conference, China has surpassed the United States in studies presented, and that is the definitive sign of how the Asian giant has achieved a victory in a crucial area for the future of this technology. Alarming data. This data is not something isolated, but the result of a trend that began ten years ago. In 2019, 29% of researchers presenting their work at NeurIPS had started their careers in China. In 2025 that figure is 50%. Meanwhile, the proportion of quinees who began their careers in the US has increased from 20% in 2019 to 12% in 2025. The analysis is based on a sample of 600 articles written by almost 4,000 researchers (many studies have several researchers as authors). Chinese universities dominate. This analysis also served to analyze the origin of the researchers who published these studies. Nine of the ten institutions where the most NeurIPS 2025 researchers completed their studies are in China. Tsinghua University is, for example, the protagonist with 4% of all researchers. The prestigious MIT in the USA? Only 1% comes from there. Quantity matters, but also quality. It must be taken into account that this does not necessarily mean that China wins (or loses) in research quality, but it does in quantity. But this parameter is very relevant, because scale matters: when China manages to “produce” a huge number of AI graduates, its chances of those experts being responsible for new advances in this discipline increase. Not only that: it also makes these advances spread faster within the Chinese technological ecosystem. The US depends on Chinese talent. One of the most uncomfortable details of this study is where those who signed studies from US institutions were trained. Of all of them, 35% They graduated from Chinese universitiesthe same proportion as those who did so in US universities. Many leading AI companies in Silicon Valley are drawing on AI experts trained in China, which is increasingly the world’s largest pool of this type of engineers. Come home come back. What is worrying for the US is that the Chinese talent that US companies sign increasingly ends up returning to China. Chinese programs like Thousand Talents Plan They offer up to $100,000 annually plus subsidies for housing and research to attract that talent back. The United States government is also promoting just that, because the funding cutsthe uncertainty with visas and suspicions towards researchers of Chinese origin make working in the US no longer so attractive for these experts. Or what is the same: The US is shooting itself in the foot (again). From the American dream to the Chinese dream. In 2019, about a third of NeurIPS researchers who had graduated in China stayed in the country to work. In 2022 that proportion rose to 58%, and in 2025 the figure already reaches 65%. And as we mentioned, those who had left are returning: in 2019, only 12% of Chinese researchers who had completed postgraduate studies outside of China had returned, but in 2025 that figure has risen to 28%. The case of DeepSeek It is significant: none of its main contributors have a university degree outside of China: the talent who achieved that milestone He didn’t go through Stanford or MIT. The trend doesn’t lie. If we stick to the authors of studies published in NeurIPS as a metric, about 37% of the best researchers in the world now work in Chinese organizations, compared to 32% of those who do so in North American institutions. If this trend continuesin 2028, researchers working in China could outnumber those working in the US by two to one. Silicon Valley may continue to attract a lot of international talent, but the direction of the trend is clear, and that points to a worrying future for the United States. Image | Tommao Wang In Xataka | There is a city in China that goes head to head with Silicon Valley: welcome to Hangzhou, the home of the ‘Six Little Dragons’

Wikipedia has banned using AI to write or rewrite articles in English. Human knowledge begins to raise barriers

by

The English version of Wikipedia has just banned articles made with AI. In the last update of their guidelines are clear: content generated with language models violates content policies. The largest encyclopedia on the internet positions itself as a refuge for content created by humans. AI no thanks. The ‘AI yes or AI no’ debate has been going on for a while generating tension on Wikipedia and they have finally opted to support human content with an overwhelming majority 40 to 2. The new restriction imposed reads as follows: “Text generated by large language models (…) often violates several of Wikipedia’s fundamental content policies.” Those fundamental policies What it refers to are the neutrality of the content, verifiability and that the content cannot be original research, but must be attributed to reliable sources. With this change, editors are prohibited from using LLM “to generate or rewrite article content.” Two exceptions. Wikipedia contemplates two scenarios in which the use of AI is allowed: Basic style suggestions and corrections, as long as the LLM does not introduce its own content. They warn that it must be used with caution since LLMs tend to “go beyond what is asked of them and alter the meaning of the text.” Translation of articles into other languages, as long as it is reviewed by a person competent in the two languages ​​involved. Here it is important to note that Wikipedia has already had dramas in the past because of AI translations. Why is it important. Wikipedia has positioned itself as a repository of genuinely human content in an internet that is flooded with artificial content. At a time when distinguish the authentic from the synthetic is increasingly difficult, the largest encyclopedia in the world chooses to rely on human authorship as a guarantee of reliability. There is certainly something ironic and that is that Wikipedia rejects AI, but AI continues to draw on Wikipedia to provide answerscausing them to lose clicks and saturating your servers. AI generated vs human made. Until recently we thought that the solution was flag artificial content on platforms with the classic ‘AI’ label, but we are already at a point where it is more valuable and useful to highlight the opposite: that it is made by humans. The advancement of image generation tools and the amount of texts made with AI are overwhelming, to the point that an anti-AI current is emerging; Some artists are starting to designing “badly” to differentiate itself from AI homogenizationthey have created extensions to return to the internet before ChatGPTthere is browsers that filter AI results and even ‘Not by AI’ badge has been created. The point is that it is a David against Goliath. The Etsy case. It is perhaps one of the most bloody cases of the flood of low-quality AI content. The platform that It was presented as a refuge for the authentic, today it is an AI market which also tries to pass itself off as artisanal. Ghibli-style portraits for 20 euros, profiles managed entirely by AI that say things like “I can’t wait to draw you”… Etsy allows content made with AI, but says you have to label it as such. Nobody does it. Proof that the label is no longer useful. A key detail. The last paragraph of Wikipedia’s guidelines is especially striking because it talks about possible sanctions for those who violate the rule, the problem is how they plan to detect who uses AI. Wikipedia admits that “some editors may have writing styles similar to those of large language models” and that “more evidence than mere stylistic or linguistic clues is needed to justify the imposition of sanctions.” We have no idea how they are going to do it, what we do know is that AI text detectors fail more than a fairground shotgun. Image | Wikipedia, edited In Xataka | The last barrier against AI is good taste. The problem is that an entire generation is growing up without developing it

the number of human casualties is going to skyrocket

by

In modern conflicts, a single anti-missile interceptor can cost more than a home in a large European city, while the drone you are trying to shoot down can be manufactured for the price of a utility. Still, in just a few days of modern combat, entire armies can consume the equivalent of years of industrial production, revealing along the way the extent to which today’s war is fought so much on the front lines. like in factories. Eating years of arsenals in weeks. The conflict between the United States, Israel and Iran has entered a phase accelerated consumption of unprecedented ammunition, one where equivalent volumes of missiles have been expended in just days to years of productionespecially in systems such as Tomahawk, Patriot or THAAD, whose cost is millions and whose replacement can take years. This dynamic is not only triggering the economic cost and pressure politics in Washington, but also exposes a structural vulnerability: great powers simply they are not prepared to sustain prolonged, high-intensity wars, especially when they must simultaneously engage in global commitments against rivals such as China or Russia. Rheinmetall’s warning. And it is at this point where one of the most authoritative voices to talk about arsenals has appeared. The most forceful warning comes from the largest European arms industry, whose executive director, Armin Papperger, has pointed out in a CNBC interview that air defense arsenals in Europe, the United States and the Middle East are practically empty and that, if the war lasts just another month, they could be almost completely exhausted. The scenario he has drawn is completely unprecedented and the warning is not theoretical, but rather a reflection of a reality already visible: the demand for missiles It is “insane”warehouses are at their limit and the industry cannot increase production at the necessary pace, creating a very specific time horizon in which the conflict must be resolved or radically transformed into something very dangerous. The paradox of cost. At the heart of the problem is an unsustainable equation for weeks: While Iran uses cheap and easily produced drones, the United States and its allies intercept with missiles that cost millions, multiplying the economic and material wear and tear. This asymmetry, which already we had seen it to a lesser extent in Ukraine, turns the war into a kind of wear competition where it is not the one who hits the hardest that wins, but rather the one who can resist the longest by producing and consuming projectiles, and where even a successful defense involves accelerated exhaustion of critical resources. War as a “salvoe competition.” The conflict has thus evolved towards a convoluted logic of constant exchange of massive attacksone where the key is no longer absolute air superiority, but the depth of the arsenals available on each side. In this sense, most analysts match in which the outcome may depend on an unusual situation in modern wars: simply on who exhaust your reserves earlysince not even the most advanced defense systems can guarantee complete protection, and each impact that passes through defenses can have disproportionate strategic and psychological consequences, such as we have seen this week. The scenario without missiles: melee. If that critical point is reached that Rheinmetall anticipatesthe war will not stop instantly, but will very possibly mutate towards most dangerous ways: increased use of conventional artillery, the emergence special operations or even less precise attacks that increase the risk to soldiers and increase the likelihood of errors or collateral damage in urban environments and critical infrastructure. In fact, almost at the same time as the CEO of Rheinmetall discovered the future that awaits the war in the East without missiles, Israel dropped that it was time for “cavalry” on land. Because the loss of precision attack and defense capacity eliminates one of the main containment barriers of modern conflict, making violence more direct, exposed and difficult to control. In short, much more dangerous. Quick victory or wear. Thus, while leaders like Netanyahu they insist in which military objectives are achievable and the conflict could end sooner than expected, the reality on the ground points to the opposite: a war of attrition that has already surpassed the initial forecasts and that is forcing strategic decisions under material pressure. In this context, the true decisive factor stops being the immediate military power and becomes the industrial and logistics capacitywhich turns each week of conflict into a race against time between exhausting the enemy or reaching one’s own limit first. Because the great paradox, remembered by Germanyis that when the missiles fail, not only will the defenses fall, the last invisible wall that contained human casualties will fall, and with it, the war will cease to be precise and will inevitably become more lethal. Image | DoD, Bernd vdB In Xataka | The war in Iran is being given the face of Iraq in 1991. And that is dangerous because we are tied hand and foot In Xataka | Russia is not sending troops or weapons to Iran: it is sending something much more important to take down the US

We believed that human programmers would end up being code reviewers. Anthropic just killed that

by

The rise of the Generative AI The world of software development seemed to follow a clear script: models would write the code and humans would review it. It was the new balance. Well, Anthropic just killed him. The problem of programming with AI. What we know today as vibe codingthis practice of giving instructions in natural language to an AI so that it generates code at full speed, has skyrocketed software production in companies. Anthropic affirms that the amount of code generated by each of its own engineers has grown by 200% in the last year. And now there’s a problem: there’s so much new code that reviewing it has become the bottleneck of the process. Human developers can’t cope. Many pull requests (change proposals that must be reviewed before integrating new code) are skimmed or not read very carefully at all. What Anthropic has done. The company Code Review has been releaseda tool integrated into Claude Code that, instead of waiting for a human to review the code, deploys a team of AI agents to do it automatically every time a pull request is opened. This new system is now available in preview phase for Team and Enterprise plan customers. Cat Wu, Product Manager at Anthropic, explained told TechCrunch that the question they constantly received from their clients’ technical managers was always the same: “Now that Claude Code is generating a ton of pull requests, how do I make sure they are reviewed efficiently?” How it works inside. AI agents work in parallel autonomously the moment a pull request is opened, examining the code from different perspectives. An end agent then aggregates and prioritizes the issues it has found, removing duplicates and sorting them by severity. The result reaches the developer through a featured comment, accompanied by more online comments about specific bugs. The focus, according to Anthropicis in logical errors, not in matters of style, something designed on purpose so that the feedback does not generate too much noise. Issues are labeled by color depending on how important they are: red for critical, yellow for attention, and purple for pre-existing code. Numbers. The company has been using Code Review internally for months before launching it to the market. According to what they saybefore implementing it, only 16% of their pull requests received meaningful review comments. With the tool, that percentage rises to 54%. In large pull requests (more than 1,000 modified lines) 84% returned results, with an average of 7.5 problems detected. And less than 1% of those results are flagged as incorrect by the engineers themselves. In one of the cases documented by the company, they spoke of a single line change that seemed routine. However, Code Review marked it as critical, as it apparently could have broken the entire service’s authentication. The bug was fixed before integration. Furthermore, according to the company, the engineer later acknowledged that he would not have caught it alone. ANDhe new role of the programmer. The narrative that had spread in the last two years was that developers would evolve towards a profile closer to that of a reviewer or supervisor of code generated by AI. Now that transition is also being automated, at least in part. Anthropic does not eliminate the human from the equation (in fact the tool does not approve pull requests), but it does compress the review work that was supposed to be the last bastion. It seems that now the human goes from reviewer to final arbiter. Price. It is not a cheap tool. Each revision has a cost based on token consumption. Anthropic esteem The average price per review is between $15 and $25, depending on the complexity of the code. It is a cost that the company justifies in the context of large technology companies where errors that escape review have a much higher price. Cover image | Compagnons In Xataka | Software companies sank on the stock market for a simple reason: investors are panicking about AI

We thought that human beings began to walk in Africa. This 7.2 million-year-old fossil says otherwise

by

The scientific consensus has been telling us for decades that the cradle of humanity and the origin of our ancestors who began to walk on two legs was in Africa. However, a new paleontological discovery in the Balkans just launched an order to this official story. More specifically, a fossilized femur that suggests our earliest ancestors may have started walking on two legs in Europe. A bone. The centerpiece of this discovery is a femur cataloged as FM3549AZM6 and found at the Azmaka site, in Bulgaria. From this, the research team began to analyze the bone down to the millimeter, highlighting above all the anatomy it had. Researchers here have identified key biomechanical traits that point to partial bipedal locomotion, meaning that our ancestor could walk on two legs. Specifically, they have seen that the neck of the femur is unusually long and it has specific muscle insertion points that strictly arboreal primates do not have. These characteristics suggest that Graecopithecus He spent considerable time walking upright on the ground. A new hypothesis. This finding does not come out of nowhere, since in 2017 this same team of researchers already raised eyebrows in the scientific community by suggesting that the evolutionary divergence between humans and chimpanzees could have occurred in the eastern Mediterranean, and not in Africa. That hypothesis was based on analysis of a jaw found in Greece and a tooth from Bulgaria attributed to Graecopithecus freybergi. Now it comes to light again. At that time, definitive proof of locomotion was missing, but Azmaka’s femur fills that gap that we needed to begin to reach clear conclusions. Why did they stand up? Evolution rarely occurs without a strong environmental push, and the Europe of 7 million years ago looked nothing like it does today. Here investigations at Bulgarian sites, such as the Struma Valley, show that the landscape was dominated by a savanna environment very similar to the African one, caused by a global confrontation and severe droughts in the Mediterranean. This loss of dense forests would have forced the region’s primates to come down from the trees and adapt their movement to travel long distances in open fields in search of food. In this way, it was geography and not the continent that forced bipedalism. The debate. The new Bulgarian femur revives one of the hottest debates in paleontology, since until now, the title of the oldest bipedal hominin was held by Sahelanthropus tchadensisabout 7 million years old and found in Africa. But now, if this team’s dating and analysis are accurate, Graecopithecus would not only equal, but slightly surpass in seniority Sahelanthropusmoving the “kilometer zero” of bipedalism to the Balkans. But at the moment it is too early for the textbooks to change definitively, since, as with previous discoveries, the scientific community will demand more independent analyzes and will seek to debate every notch of the femur. What is undeniable is that the African monopoly on the origin of our lineage now has a serious European competitor. In Xataka | Humans are evolving live on the Tibetan plateau. And understanding what happens there will be essential in space

If we want to increase human fertility, mice have something to tell us: fecal transplants

by

We knew that the bacteria that live in our intestine They are really positive and offer us extra protection against numerous threats from outside or even against Alzheimer’s. Now they have just added a new star function: they can help us improve our reproductive health. And all this with a simple fecal microbiota transplant. New evidence. This same month of March the magazine Nature has published an article that breaks with several scientific paradigms and demonstrates a bidirectional communication between the microbiome and women’s ovaries. The study here wanted to demonstrate that fecal microbiota transplant can completely remodel the behavior of the ovaries, reducing inflammation and even increasing reproductive success. But the most amazing thing about the experiment is not the ‘what’, but the ‘how’, since it has quite surprised the experimenters that the result has gone against what they expected. How is it possible? To understand this finding, we must first know the concept of ‘strobolome‘ which will give a lot to talk about in the coming years. In a simple way, it is a set of intestinal bacteria that are capable of metabolizing and modulating the level of estrogen, which is one of the main female sex hormones and closely related to reproduction. Until now, we knew that the microbiota played a role in almost every part of our body, ranging from digestion to our own mental health. But they wanted to go further, and in the past it was noted that they had already begun to explore how to transfer faeces from young mice to old ones, which would improve their ovarian reserve. But the best thing is that doing the opposite could accelerate the aging of the ovaries. The new study. Knowing this, this research team designed an experiment in which healthy adult mice were taken and given antibiotics to cleanse their intestinal flora. From here, they underwent a fecal microbiota transplant from mice in the ‘estropausal’ phase, which is like human menopause. What the researchers here were expecting is that if they were introduced to the microbiota of aged rats in reproductive decline, their ovaries would suffer damage. But the truth is that they were wrong. Results. The results here showed that adult females who received the “estropausal” microbiota not only did not worsen, but rather improved ovarian function and an increase in their fertility. Because? By thoroughly analyzing the organism of these ringworms, using sequencing of the genetic material of the ovaries, it was seen that the transplant had caused a massive remodeling of the ovarian transcriptome. That is, the way in which the genes in the ovaries were expressed had completely changed towards a “younger” profile. In addition, the analyzes revealed a drastic drop in the expression of genes linked to inflammation. The explanation that scientists are considering points directly to the strobolome, since it is possible that the microbiota of the stropausal mice, in its attempt to survive the natural drop in hormones of its original host, has developed brutal compensatory mechanisms. By transplanting these “surviving, super-optimized bacteria” into a young environment, they boosted the health of the recipient ovary. The future. Although in this case this effect has been seen in mouse animal modelsthe implications of this study may allow us to continue advancing treatments that can improve human reproductive health. The goal here is to be able to isolate what exactly are the specific microbial candidates or metabolites responsible for this improvement and in the future we could be talking about probiotic treatments or microbiome-based therapies to prolong fertility. Images | Kelly Sikkema CDC In Xataka | Drinking coffee in the morning has very positive consequences for someone unexpected in your body: the microbiota

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.