internet

China was isolated from one of the key arteries of the Internet for an hour. It is not clear if it was an essay or a failure

by

On the Internet there are falls that are noticed. And then there is what happened in China. For a whole hour, one of the main road traffic with the outside stopped working without explanation. It was not a total blackout, but a cut that became visible in many parts: applications that were halfway, websites that did not respond, services that failed right in the border crossing. The ruling did not fit with the typical local fall or with a problem of a single operator. It lasted 74 minutes, crossed networks and left behind an awkward silence. What fell was not all the Internet, but a concrete and widely used part: the channel that allows encrypted connections to function safely. According to GFW Reportbetween 00:34 and 01:48 of August 20, Beijing time (UTC+8), a generalized blockade was recorded in the Port 443which is the one that uses the majority of traffic HTTPS. The effect was broad and sustained, but did not affect other channels. The most striking thing is that, until now, no one has explained what caused that cut or why it happened in that strip. The technical analysis points to a very concrete pattern. Every time a connection tried to settle in that port, special packages – the RST+ACK – appeared that act as a closing order. They are the digital equivalent to hang the phone before someone attends. These packages were not sent by mistake: they were being injected into large quantities just at the key moment of the exchange between client and server. According to the specialists, this alteration did not fit with a fall due to congestion and affected both connections that came out of China and to those trying to enter. Despite the serious cut, not everything stopped working. Internal connections within China were still active, and other ports, such as 80 – for non -encrypted traffic – or 22, intended for remote connections, showed no block signs. The problem was focused on port 443. That explains why some services could continue to work while others became inaccessible: it was not a total disconnection, but a very selective interruption of the encryption step that crosses the digital border of the country. There are no public reports indicating that platforms such as WechatBaidu or Weibo suffered a generalized fall in that strip; Its main infrastructure is mostly within the country and does not depend on the international crossing for basic functions. But many Chinese apps incorporate components that depend on Servers outside of China-action, mini-programs, external APIS or cloud functions-and those calls could fail when the encrypted step was blocked. There are no public reports that indicate that platforms such as Wechat, Baidu or Weibo suffered a general fall in that strip The impact was more noticeable in international services. According to The Registerpart of the connections of foreign services – including some functions of Apple or Tesla – could be interrupted for users in China during the blockade window. Other international services that depend on HTTPS, such as certain CDNS or VPN connections that use TCP/443. In summary: the internal network was not broken, the encrypted channel was cut out and that was enough to leave many out. One of the most relevant clues is precisely that technical footprint. Each system that filters or interrupts connections leaves a trail –a fingerprint– Recognizable for the order of the packages, the TTL or the TCP window size. In this case, these values ​​differ from the previous documents, so the researchers propose two hypotheses: a new equipment deployed for tests or a known team that was in an anomalous or poorly configured state. The conclusions remain in condition until there are more data. Traffic returned to normal and there were no visible consequences on a large scale.But the cut was real, it was documented and left an unknown in the air. There were no officials or official explanations. In any case, the great firewall is still there, functioning as always, and with an intervention capacity that this time was recorded. Images | Leon Seibert | Xataka with Gemini 2.5 In Xataka | To have a rather mediocre AI it has been necessary to loot all the internet content. And Reddit has said enough

The future of the Internet is flooding of ia. And there are those who have already seen a business niche: content made by humans

by

Internet has been filled with Searches results generated by artificial intelligence and of ‘Slop‘. The first is already reducing clicks to halfand points to collapse for the media. The second is something that users will have to live if the platforms do not get more serious to improve the filtering. Among all the tangle of new and ancient actors, an old Google opponent has returned to position itself as an alternative to the current in this era. Forks Duckdukgo. What are they doing. Duckckgo is an alternative search engine focused on offering the greatest possible privacy to its users, and for this, use, for example, the Bing searches. After 16 years since its foundation, the company has launched a Filter of images generated with AI that frames in its philosophy of the “private, useful and optional” functions. For the end user, this means being able to see or hide the images generated by the general search of images, something that is already problematic in Google where, for example, we come to see distorted results, even in examples as recognizable as ‘The young woman of the pearl’. The company has already shown that it can be differentiated, with Duck.aian own chatbot that anonymizes conversations and does not use them for training. How do they. To use this filter, the company is not using an engine driven by artificial intelligence or something similar. Those responsible have detailed the method, and is based on using manual open source block lists that the popular content blocker provides Ublock Origin and the Ublacklist’s Hage AI Blocklist, a browser extension created so that users can manually filter concrete websites of their Google results. Effectiveness. One of the big problems of the contents generated with AI is that they not only come from repositories that clearly indicate it, such as Midjourney explore, for example. In this sense, from DuckDuckgo They recognize That the filter “will not capture 100% of the results generated by AI”, but “will considerably reduce the amount of images generated that are shown.” Examples of images generated by AI that do not come out when activating the filter. Some of them is not as obvious as the largest in the first row. In our case, for example, we have sought “dog jumping” without activating the filter so as not to see. And he has found the image of a dog jumping with a suspicious origin: Freepik. It was an image generated by AI that has stopped appearing when activating the filter. Next to him was more of the same source that he has also leaked. The challenge. The filtering works well, but the challenge for companies such as Duckduckgo is capital, because Freepik already serves as “generated with AI”. But what happens when a random image that we find in media like this has no label to indicate it? There they have to play methods that use verification system with detailed analysis of the image, which makes the resources required to do so effectively fire. Beyond images, the difficulty is in the filtering of websites with texts generated by artificial intelligence. Openai, for example, launched A tool to detect plagiarismand then He withdrew it silently. Although according to leaks now they have a tool that operates, They do not throw it. The problem. Detectors do not just work well, and after reaching accuse students of using and erruniversity professors are suffering the phenomenon: To make undetectable texts, students are resorting to rewriting tools that worsen their quality. Google, a strong defender of the generative AI, has opened Its Synthid water brand technology. The inconvenience? We continue needing a universal standard that at the moment does not arrive. And even if it arrives, users will not always want to mark their creation as done with ia. For example, the images generated with Google image indicate its nature with the “AI” icon, but erase or cut it is very simple. The snowball does not stop, The hole. According to the Empty Internet Theory either ‘Dead Internet Theory‘, Much of the content that exists in the network is created by automated bots and algorithms, and with generative intelligence, the thing is going more. In addition, Google’s preview has only been the first step: now it touches Ai mode, which no longer delivers a results, talk. So, although we win comfort, there will be a huge hole for companies such as Duckckgo, looking for that the basic unit of communication on the Internet is still aliveor that knowledge remain something democratized without Loot all Internet content. As an idea, DuckDuckgo can deepen including filters throughout web navigation: it can be distinguished from browsers with AI with a different approach to Your browser. Images | Xataka In Xataka | People did not stop hacking virtual work interviews with AI. Solution: We want to meet you in person

The AI does not stop stealing internet content to improve. And the Internet has swollen the noses

by

That Reddit blocks the action to the Archive Internet is a tragedy. And also a way of stopping a voracious and implacable The price to be paid for having ia It is the looting of all the internet content. In Reddit they know it well, and they have just taken an extreme measure against those indiscriminate robberies: it is not that Block access to their contents to the ‘scrapers’ of the companies of AI directly. Now they also block them if these companies try to access with rodeos. The injured party? Internet Archive. What happened. Reddit, who has always been very proactive when protecting “their” contents (which, by the way, have been generated by users voluntarily and free), has realized something: they were stealing them. But not directly, but through previous versions stored in that gigantic digital hemeroteca library called Internet Archive. Wayback Machine without access. Wayback Machine is the “Machine in Time” of the Internet Archive, and allows access to old versions of any website. But to avoid more content theft, Reddit has banned this platform to index the vast majority of Reddit content. Only the Reddit.com home page can be indexed. Reddit’s argument. Tim Rathschmidt, spokesman for Reddit, explained In The Verge That although Internet Archive is a service aimed at the open web, they had discovered “cases in which artificial intelligence companies violate the platform policies, including ours, and extract data from Wayback Machine.” In addition, he pointed out the following: “Until they are able to defend their site and comply with the policies of the platform (for example, respect the privacy of users, in relation to the elimination of deleted content), we are limiting part of their access to Reddit data to protect Reddit users.” If you want our contents, pay. That message from the spokesman is reasonable, but as little is incomplete. Especially since Reddit has persecuted that kind of looting of AI companies. He has tried to block those who did it with technical means, and the goal both before and now was the same: that companies pay for their contents. It is something that has achieved with the agreements that have reached since this type of processes began. The first thing he did was close his APIa disaster for all the Internet. Then he ended up reaching a Google agreement, which pays 60 million dollars a year In order to have access to those contents. And the same ended up with OpenAiwith which he sealed a pact whose economic details have not been unveiled but that gives access to Reddit’s contents to the models that enhance chatgpt. My content is mine (more or less). Social platforms have been nourishing the content of users for years. Until now the business model focused on advertising, but the arrival of AI has allowed us to have an interesting alternative model: that the companies of AI pay for being able to access those contents. Users barely win, Reddit and social networks do. Contents affirming that they are his – as Reddit, which in June He sued Anthropic– But that actually created the users of these platforms, who without realizing have become slaves of these social networks: they do not stop producing content that others consume, and do so without charging a euro. These platforms are intermediaries that provide the necessary infrastructure for this content to be available for free, but there are hardly any consideration for creators. Only a few can make a living on YouTube, Tiktok or Instagram for example. In Reddit exist Some metallic remuneration for the “taxpayers” who create the most for the platform. Cloudflare and content locks. Content companies are beginning to act in a similar way, and in the last two years we have seen how some editorial groups –Including haste– They reach agreements with AI companies so that they can use their contents. You Shall Not Pass! However, there are companies that go further. We have as clear example to Cloudflare, which has created a system so that companies that use their services can block the “Crawlers of AI” that try to steal their contents. If you are a cloudflare customer, you can activate that block, thus avoiding the problem or at least putting it Much more difficult to the AI companies that try to train their models with your data. Media and platforms such as The Associated Press, Fortune, Time or Stack Overflow are some of the companies that are already using said system. Quid Pro quo. This cat and mouse game is especially striking for the entire content creation segment, because IA companies use All shortcuts that can to capture (and steal) That data, Have or not copyright. What Reddit raises is a model in which creators compensate for the AI to take that data. Or more than creators, platforms that serve as their meeting and showcase. Media groups and audiovisual content producers have an interesting opportunity here Before the potential traffic collapse caused by solutions like Google AI Overviews. In Xataka | The “digital decomposition”: how 38% of the websites that existed in 2013 have disappeared from the Internet

Europe wants users to verify their age to navigate the Internet. Google does not need it: you can estimate it with AI

by

In February Google advertisement They were going to start starting an experimental system based on AI and automatic learning (Machine Learning) for Estimate the age of users of its services. That technology, initially thought to avoid access to adult content on YouTube, will now serve much more, although for now it will be restricted to the United States. Automated restrictions. If Google detects that a user is under 18 in the US, it will apply the same restrictions that it applies when users are proactively identified as minors – for example, when creating Gmail accounts. Thus, it will enable reminders to go to sleep on YouTube, it will limit content recommendations and disable both personalized advertising and access to Adult applications in Play Store. What happens if AI is wrong. If Google’s system incorrectly estimates that a user is under 18, said user will be able to send a photo of his identity card (in the US, that of driving usually), of a credit card or even a selfie. First, that Google ends up having a photo of those documents It is quite disturbing. Second, the selfie option is especially surprising, because a priori might not be enough to verify that someone is not a minor. Remembering How-Ord.net. Microsoft a decade ago launched an AI experiment according to which a user could upload a photo of one or more people, and the service He tried to guess the age of those people. The tool quickly went viral, but it was also controversial because of its possible impact on privacy. A study Of 2022 he revealed that various systems of this type were quite precise when guessing age in young people, but not so much in older people. Companies specialized in this area as Yoti explain that The margin of error is already reducedand its system has an average 1.3 -year error in young people between 13 and 17 years old. How to estimate the google age. Google’s age estimation model will use the existing data of its users, including the websites they visit, what type of videos do they see on YouTube or how long their accounts have had. What about privacy. According to Google, this approach “does not imply the collection of additional data”, but of all that information that you already have thanks to the profiles that users use to access their services. They also ensure that they will not share that estimate with apps or websites. Goal already implemented A similar system on Instagram months ago. The age verification fever. The movements in this regard are increasingly striking. Before the excuse for companies and governments to spy on us and knew more about us was the terrorist threat. Now it is that minors They can access adult content. There is a unique fever for developing age verification systems everywhere, and we are seeing it especially in Europe … and the United Kingdom. The United Kingdom teaches us how the future can be. One of the first countries to move in this area is the United Kingdom, which these days these days has activated its regulation so that minors cannot access adult content. The age verification systems must be present in all types of platforms, something that has been especially controversial there and that has generated a surprising (or perhaps not so much) Collateral effect: British users They have begun to subscribe in mass to VPN services to avoid having to pass those age verification controls. The EU wants its own “pajporte”. The controversy introduction of the Digital portfolio beta In Spain —Popularly known as “Pajortport” – is one of the great steps that EU member countries are facing the future European digital identity. Among the options that will enable this mobile application will be to verify the age of the users, and in fact the Spanish initiative is serving as inspiration For a European system. Said system, yes, is not exempt from controversy. It remains to be seen if the Google system is really effective, and if it is, how its deployment progresses and what impact that has on the intentions of various countries to control the age of users. In Xataka | Allowing an app to “all photos” of our gallery looks like the most harmless action in the world. It is not at all

Azure is being for the AWS was for the Internet

by

Microsoft has made history by joining Nvidia in the exclusive club of the four billion dollars after Spectacular results that demonstrate how their commitment to the cloud and AI has completely transformed the company. Why is it important. Satya’s decision to reposition Microsoft as a company “Cloud First“A decade ago it has turned out to be one of the most successful strategic bets in technological history. Long before The arrival of chatgptMicrosoft decided to focus on building the infrastructure that would feed the AI revolution. In figures. The numbers of the latest quarterly results published are overwhelming: 76.4 billion dollars in quarterly income (+18%). Net profit of 27.2 billion (+24%). Azure exceeding 75,000 million annually with a growth of 34%. These figures fired the company’s action in the After-Hourswhich has catapulted the stock market capitalization above the four billion. The context. During the Ballmer era, Microsoft seemed condemned to be a desk for desktop software. Windows kept dominating, but the future pointed to the mobile and cloud. Nadella’s arrival changed priorities: less Windows, more digital infrastructure. His alliance with Openai in 2019 It ended up being visionary, since he assured access to the most advanced models without astronomical development costs. The alarm signal. Microsoft plans to spend 120,000 million dollars annually in data centers, quadruplying the investment of 2023. The financial director, Amy Hood, has confirmed that “the demand continues to exceed the supply” and that all this expense is “directly linked to contracts already signed.” It is the most expensive arms race in the history of technology. Yes, but. Success generates new problems to solve. The relationship with Openai has been tensioned and Microsoft has diversifying towards models of Grok and calls not to have all the eggs in the same basket. Copilot makes Microsoft continue to dominate in the business market with its 100 million users, but it is still far behind Chatgpt or Gemini, who multiply their adoption. And the winner is … Nadella has shown that deep business transformations are possible. From a Microsoft anchored in software licenses to a digital infrastructure power that competes from you to you with Amazon Web Services. Azure has not only grown: it has become the backbone of the AI era, processing the most demanding workloads of the planet. Script turn. For the first time in its history, Microsoft has publicly revealed the specific income of Azure and not only the accumulated of its division. Thus breaks years of opacity that Even former Steve Ballmer had criticized. Transparency comes just when the division has become its main growth engine and the most valuable asset of the company. In Xataka | Microsoft is in benefits records and also in dropout records. Satya Nadella explained this paradox Outstanding image | Microsoft, Xataka

A student from Girona set up his Cybercrime Office at home. His mistake was to show off his robberies on the Internet

by

A joint operation of Mossos d’Esquadra and National Police He has arrested a hacker in Roses (Girona) Responsible for several cyber attacks operating from a sophisticated technological network … in their own home. Why is it important. The detainee had managed to configure a technological network complex using anonymous applications that allowed him to hide his trail while attacked financial institutions, private companies and public bodies. The researchers discovered that the hacker, a young computer student with very advanced knowledge, stole personal databases of employees and clients, as well as confidential internal documents. Then he sold this information in forums of the Dark Web or published it without asking for anything in return. What has happened. The investigation started in May 2024, when the agents detected in a forum that someone presumed to have data from Spanish banks, a self -school and a public university. The same subject later claimed to have filtered the database of a Barcelona plumbing company. After an arduous discouragement of desanimization, the researchers located Roses’s address from where the attacks were launched. The subsequent investigations identified the inhabitants of the house, including the young computer student. In detail. During the record they seized: A laptop. A dozen mobile phones. Several hard drives. More than thirty SIM cards. Several bank cards in the name of different people. The detainee made available to the Court of Guard of Figueres for a crime of discovery and revelation of secrets. And now what. The investigation remains active to locate new victims and clarify all the scams that this person may have committed using stolen data to the affected entities. Outstanding image | Mossos d’Esquadra In Xataka | We visited the National CNI cryptological center: here is the epicenter of Spanish cybersecurity

Leave an army without internet in the middle of a battle

by

In a war where drones are as crucial as ammunition, cutting Internet access is the equivalent of cutting supply lines. And that was exactly what, according to a new and explosive Reuters reportmade Elon Musk in Ukraine at the end of September 2022, paralyzing a key counteroffensive against Russian troops. It was not a technical failure, nor a Russian cyber attack. According to the agency, it was a direct order of the Musk itself that left the Ukrainian troops blind and showed that the richest man in the world has the power to change the course of a war from his office in California, thousands of kilometers away from the front. The story, which contradicts the public statements of the businessman, is based on the testimony of three people who know the decision of Elon Musk. The first direct proof that the tycoon interfered in the Ukraine War since Walter Isaacson told a similar case in his biography and then rectify. But the incident goes far beyond past controversies and places the tycoon in an unprecedented geopolitical power position for a particular citizen. A deliberate blackout in the front of Jersón. The events occurred during a key counteroffensive of the Ukrainian army to recover the strategic region of Jersón. The Ukrainian troops advanced and depended almost completely on the Starlink terminals for everything: coordinate attacks, point the long -range artillery and pilot the surveillance drones that gave them eyes on the Russian positions. It was then that Elon Musk gave the order to a Spacex engineer, Michael Nicolls, to cut the coverage in the counterattack areas. According to one of the sources, the order was blunt: “We have to do it.” In the Spacex offices, employees complied, seeing how a hundred hexagonal cells that represented Starlink’s coverage on the company’s internal map were turned off. The impact on the front was devastating. The communications were suddenly cut and the surveillance drones were left without a sign, leaving the isolated and vision units on the enemy forces. Artillery units, which are used for the precise geolocation of objectives, began to fail in their shots. The soldiers panic. According to a Ukrainian officer, the attempt to surround a Russian position in the city of Berislav failed after the blackout. Although Ukraine managed to release the area, Musk’s decision temporarily redrawed the front line. Walter Isaacson’s biography fell short. This new report on Jersón is more serious than the most famous controversy to date: the Crimea incident. In September 2023, Walter Isaacson published his biography of Elon Musk telling that the tycoon had ordered to turn off the Starlink connectivity in Crimea to frustrate a Ukrainian attack with submarine drones against the Russian fleet in Sevastopol. That statement caused a media earthquake, but was withdrawn by Isaacson himself in subsequent editions of the biography. According to Musk, what happened was actually that Ukraine had requested emergency activate Starlink’s coverage on Crimea, an area where he was not operational. Musk refused. Your reasoning, expressed in several publications in XIt was clear: doing so would have turned Spacex an accomplice of an act of war, violating the law. So why did you interfere in Jersón? Reuters research points to a reason that Magnate himself has expressed on other occasions: The fear that Ukrainian advances will cause nuclear retaliation by Russia. At that time, Vladimir Putin had threatened to use nuclear weapons if Russia’s “territorial integrity” was compromised. This fear, which according to the report was shared by senior US officials, seems to have been the trigger for Musk to decide that the Ukrainian counteroffensive had gone too far. In his biography, Isaacson collected this kind of existential anguish: “How am I in this war? Starlink was not conceived to be involved in wars. It was for people to see Netflix and relax, not for drone attacks.” The Ukrainian dependence of Starlink. “One of the main factors for which Ukraine was not overwhelmed by Russia is Starlink’s support that I provided, with great risk to spacex of cyber attacks and physical attacks by the Russian military forces,” Musk wrote in the late 2024. “Starlink is the backbone of Ukrainian military communications in the front.” Starlink is a constellation of thousands of satellites, much more difficult to block through interference than Other satellite Internet services. It also offers much larger bandwidth and latency much lower than other operators. This technological superiority has not only been crucial for Ukraine, but has given Elon Musk, the owner of Spacex (which is still a private company) an unprecedented power. Image | The White House In Xataka | China and Europe are investing a fortune in their own Starlink: the US advantage is too big to ignore it

Cut the Internet in half a country

by

The war between Russia and Ukraine is far from being symmetrical: while Russia attacks Shaheds of thermobárica load, Ukraine does so with pellets. However, kyiv is reaping great successes with drones, as in the Spiderweb Operation. They are among the main reasons why the higher commanders of the Zelenski army believe that the country You can resist this year and beyond. And they are causing havoc in infrastructure and Internet cuts to the population. Reality. Associated Press collects the complaints of many Russian citizens who complain about the normal MOBILE NETWORK In his country: “We have the Internet until 12 at night, and recently there has been no connection,” the blogger Pavel Osipyan on Instagram was rapped from the city. In that city the cuts are frequent, but also in others such as Voronezh, near Ukraine or Samara. A difficult day by day for citizens. Internet cuts are leaving many people without being able to pay by card, without being able to use transport applications or even unable to get money from ATMs. The Na Svyazi activist group, whose name translates into Russian for “being connected”, He has reported mobile internet cuts in 73 of 80 regions. And in 41 of them, there are also cuts from the domestic Internet. The official reason. Russian authorities are making internet cuts in many regions because Russian and Ukrainian drones depend on mobile networks to operate and operate key aspects such as navigation, according to, Kateryna StepanenkoRussian analyst at Institute for the Study of War. The mobile connection is also key to sending video in real time and receiving orders. The trend began in May, celebrating the 80th anniversary of the defeat of Germany in World War II. Dmitry Peskov, spokesperson for the Kremim confirmed deliberate restrictions. How long? “It will be done as necessary.” In the Duma, these blockages were talked about as a form of “Preventive protection aimed at suppressing possible threats“ The suspicion: Russia kills two birds with a shot. In Russia they have been common internet cuts and Social Network Blockingand experts like Anastasiya Zhyrmont, of the group Access Nowframe these new Internet blockages as another form of population control. In this sense, constantly chase VPN services. YouTube, for example, is not blocked, but Very slowedalready end of the year was in a 20% of its normal level of activity. From Human Rights Watch They have indicated in the past their restrictive measures as censorship. Disconnection and alternative. Russia has been preparing for a total internet disconnection for years. In 2019 he tested disconnect your network infrastructure of the rest of the Internet, and the experiment was listed as ‘success’. At the end of 2024, the country successfully simulated a Internet disconnection in several regions. Before a possible forced disconnection, such as asked Ukraine in 2022 To the Icann, there are A whole plan ready. Image | Maxim SubotinXataka, Trong Khiem Nguyen (Flickr) In Xataka | The production of Russian drones was so huge that Ukraine has opened them looking for clues. The surprise is China

ASMR’s videos had become one of the most pleasant places on the Internet. Until the AI ​​arrived

by

If in recent days your feeds In social networks they are full of videos in which someone, with a knife, cuts glass planets, stuffed fruits or precious metals in slices of bread, do not worry. It is the latest generation of hyperrealistic videos made with AI, and have found a reef: extremely relaxing, satisfactory images and sounds and thanks to the possibilities of artificial intelligence, also impossible. An unexpected phenomenon. The simplicity of these tiny videos and how easy it is to elaborate them with models such as I see3 have led to have multiplied exponentially in recent weeks. They are clips of just seconds, with a surreal point, in which we see a cut knife, with extraordinary cleaning, all kinds of objects. The small models of planets (usually are of glass or stone) and other glass foods triumph. The grace is that within these objects there may be any substance or surprise: lava, gelatin, flames nuclei … The AI ​​thus exploits its ability to generate hyperrealistic images, but that the viewer knows that they cannot be real. The result intrigues enough to want to know more and stay to the following video, or enter the multiple profiles that make content of this type. Why here and now. Videos take advantage of the ability to see highly realistic textures in the foreground, allowing a prompt to just five lines and a handful of key concepts Very well chosen (“Extreme Macro Perspective, Sticky Separation with Deep Glugs”, “Rich, Slow and Hypnotic” …) these types of videos are generated quickly. That is, they are multiplying at high speed because advanced technical knowledge is not necessary, everything is automatically generated, and there is room for a certain variety: With some imagination, all kinds of objects and sensations can be recreated. Accounts to start. We assure you that as soon as you see a few accounts of this type, you will not have to make more search efforts: the algorithm will take care of the rest and you will have a real avalanche of videos of this type. But if you want some clues, there are some proper names: Impossibleais (well known and followed, glass food), Sicksicee (objects of all kinds, to each more surreal), ASMR_inventory (Video games related objects) … And more: Element.Slice (elements of the periodic table, with realistic effects), Danielleleto (Spread in bread slices impossible objects), click theaii (An original variant: computer key), slice..imposeible (More absurd glass objects) … the list is endless. Where all this comes from. The ASMR (autonomous meridian sensory response) arose Like a phenomenon on the Internet As of 2007, when personal experiences began to share pleasant sensations induced by certain soft sounds and voices. The term was coined in 2010 by Jennifer Allen to describe that sense of tingling and well -being that mainly caused the whispers and soft noises. The first ASMR videos were homemade recordings, made by fans with basic equipment and an experimental approach, and the content was varied: whispers, beaten, crunch and even role -playing games and simulations of invented situations. Its success and massification allowed the subsequent emergence of technologies such as AI, which are radically transforming the way to create and consume this type of content. In fact, these ASMR of knives cutting all kinds of objects and substances are in turn changing, and we are seeing first planes of mouths and hands manipulating all kinds of textures and food, also playing with the surreal (nibbing glass or stones with lava) Exponential growth. And although we all know these videos and their original success, which gave on foot to discussions and parodies of all kinds, they have nothing to do with this new generation of ASMR created with AI, which has automated the contents and multiplied the variants. At the moment the ASMR made with ia is being sneaking into All Trends analysis of the momentand although it is finally a fleeting fashion, that should not make us forget a very important aspect of the content generated with AI: its ability to reinterpret with impossible images and concepts traditional viral materials, which may come to change completely how we consume content. In Xataka | Our concentration has broken so much that we are buying headphones with noise cancellation to be able to sleep

The crochet was a quiet shelter away from stress and internet overinformation. Until the AI ​​arrived

by

There is the way everywhere. The last viral success in networks is made with ia, we hear music made with iathe Wikipedia has been filled with the… despite the initiatives to label itwe constantly encounter content that It gets through Real, And this It is a problem. It is exactly what is happening in the crochet community, a small and harmless Internet corner where the patterns generated with AI are ruining everything. Patterns impossible to weave. Crochet fans or crochet They went to the Internet to find new patterns. Currently, platforms like Etsy are a mines field full of patterns made with ia. Dragons, unicorns, lions … all with complex designs and a suspiciously perfect appearance that, trying to recreate them, are impossible to weave. In other cases, what users receive is absolutely similar to the photos of the product and, when weaving them, they become the meme of “what you ask vs. what comes to you.” And the worst: many are sold. What you ask what comes to you (ETSY) The deception. It has not cost me anything to find several examples of crochet patterns made with ethsy. On the first results page there are a few and do not specify that they have been generated with AI; Rather, quite the opposite. For example, the pattern of This dragon It is sold for 3.40 euros and nowhere does nothing say. Nor do we see it in the seller pagewhere all your designs are clearly made with AI. In fact, in the description ensures that you design each pattern with love and with real photos. In the reviews, this dragon has an assessment of 4.3 out of 5, which is not bad. However, it is enough to dive a bit to realize that most reviews with five stars are from people who have not yet started. If we see the worst assessment, they are people complaining that what they have received is impossible to weave. Many ask for explanations to the creator and the answers continue to insist that the pattern is real and has even been “created by a true craftsman.” A full -fledged scam. Some user reviews that have fallen in this SCAM. A problem that comes from afar. In the crochet community They have been complaining years of the irruption of these false patterns. Two years ago It was already written about this problem And it was common to see images of crochet artificially generated On Instagram and other networks. There were even guides to help people differentiate a real pattern of one made with the. The thing has not improved. There are more and more examples made with AI and the tools have advanced so much that There are people who fall believing they were facing a real pattern. The garbage goes for your hobbies. Recently we talked about Phenomenon of ‘AI Slop’ that flooded the networks such as Instagram or Tiktok with videos of quite disturbing. With the case of crochet, we are seeing that garbage applied to a hobby, and is not the only one. Something similar happens with 3D impression, where more and more Suspiciously perfect models that when they print them, they seem rather little to the photos. Sewing fans too They denounce that ETSY has been filled with patterns AI. If you like the kitchen you are not safe, Facebook is full of recipes made with ia. Labeling the AI ​​content is not enough. Although many fans are able to detect an image generated with a real one, it is enough to read the comments of many of these publications to see the enormous number of people who believe they are real. Platforms like Pinterest, Instagram either Tiktok They are labeling the AI ​​content (although It is not infallible), but the problem of constant spameo or ‘slop’ . The volume of content generated is so large, that in some searches it costs to find images that are not made with AI (try to look for Haircut ideas on Pinterest). John Oliver said it in This great monologueand perhaps it is not disenchanted: “Not all the AI ​​content is spam, but all spam is AI contained.” Cover image | ETSYedited In Xataka | A study has discovered a much more efficient activity for your brain than writing by hand or walking: sewing

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.