artificial

What is Claude Code and what this tool can do to program with artificial intelligence from your computer terminal

by

Let’s explain to you what is Claude Code or Claude CodeAnthropic’s tool to create code with the artificial intelligence directly into your computer terminal. This will mean that you will not need to install anything or be asking questions without stopping. Claude. We are going to start by explaining to you in a simple way what this tool is and the basics of how it works. Then, we will explain to you what things can you do and what this program for developers is for. What is Claude Code Claude Code or Claude Code is a command line application developed by Anthropic, the same creators of Claude’s AI. This is a program that allows you perform programming tasks from the terminal from your computer without having to use another program. The computer’s terminal is that command screen that you have in Windows called PowerShell, and in macOS and GNU/Linux it is simply the terminal. Instead of installing a common program that you have to open, the program is installed directly in the terminal, and you can use it to do so. With this program, you can use Claude to generate code within the terminal. And it not only generates code snippets, but can also act and reason directly on your projects by linking it to Github. Claude Code can read, analyze and edit content in your codebase. But in addition to this, you can also run tests and correct any errors generatedalso managing workflows. The classic way to generate code with Claude is to enter his app or website, explain what you want, and have the AI ​​create the code for you. Then you have to copy the code, paste it into the code editor you have installed and do the tests, so that if something fails you can go back to Claude, explain the problem, have him generate the corrected code again and repeat the process. Meanwhile, with Claud Code the process changes and is radically simplified. You simply open your terminal, run Claud Code in it, write a prompt or command saying what you want and that’s it. Then this AI will access your files, write code, run it, detect errors, fix them, and try again. It does all this autonomously, although you can supervise the process and intervene whenever you want. What Claude Code can do Claude Code has direct access to your file systemand can execute real commands on the computer. With all this, what this tool can do is the following: Read your files to see the code that you already have created in a folder, and thus understand the context of your project. Create new files complete with code, but also with configurations and documentation. Modify existing files editing the code you have in them to make any type of modifications. Work on an interim basisbeing able to read the error messages that appear if something fails in the code, and starting to correct these errors automatically. All this will save you a lot of time in your programming work, since you will not need to manually create folder structures, configure development tools, configure databases, create interfaces, write code, or anything. Claude will do all this automatically with just You explain the type of application you want to create in a prompt. You can also ask you to add features to existing projects with a command in which you mention the project, debug errors, review code, whatever you need. Therefore, we are faced with a tool for developers which will help you save a lot of time. Although as always happens in artificial intelligence, can make mistakes and have hallucinationsalthough within the world of AI programming Claude is one of the best. In Xataka Basics | Claude: 23 functions and some tricks to get the most out of this artificial intelligence

How to create an image of yourself and a Pixar character with your face using artificial intelligence, with Gemini or ChatGPT

by

We are going to explain how to create an image in which you appear holding a 3D character of yours miniature using artificial intelligencelooking like Pixar characters. We are going to use a prompt created for use with Geminialthough it will also work in ChatGPT without problems. It is a fairly simple composition, in which you only need to add a photo of yourself and write the prompt, which is quite long and complex. But the result is quite curious, although you may need several tries to get it completely to your liking. An image of you with a 3D cartoon What you have to do is open a new chat with Gemini, which is the AI ​​with which you will have the best results. Once you have it, upload a photo of yourself in which your face looks goodand then add the following text as a request or prompt: “Use the uploaded photo as the ONLY facial and identity reference. The main subject must look exactly like the person in the uploaded image, preserving identical facial structure, proportions, skin tone, hairstyle, eye shape, nose, lips, jawline and overall identity. Do not embellish, alter or replace facial features. Create a cinematic, ultra-detailed scene of your subject smiling naturally. The subject delicately holds a tiny, cartoon-style miniature version of the same person by the hair between his fingers, like a playful puppet suspended in the air. The miniature character is a Pixar/Disney-style 3D version of the same person, with cute, exaggerated proportions, big, expressive eyes, mouth open with joy, arms raised, and a lively, playful stance. The miniature must clearly resemble the same person and be wearing a matching outfit. The main subject looks at the little character with surprise, delight and affection, creating a whimsical and touching interaction. Lighting is warm professional studio lighting with soft rim light, shallow depth of field, and soft golden bokeh background. The real person’s skin texture is photorealistic, while the miniature character has clean Pixar-style materials, smooth shading, and polished 3D surfaces. Cinematic color grading, high contrast, sharp focus, premium portrait composition, 50mm lens look, f/1.8 aperture, ultra-realism mixed with stylized animation, 4:5 aspect ratio, 8K quality, cinematic finish. Anime, 2D illustration, comic style, flat shading, low poly, plastic skin, wax face, face swap, different identity, facial morphing, beauty filters, excessive smoothing, blur, low resolution, grain, noise, distortion, deformed face, incorrect facial proportions, extra fingers, missing fingers, duplicate hands, floating objects, bad anatomy, inconsistent lighting, harsh shadows, neon colors, cold blue tones, washed out colors, excessive saturation, watermark, text, logo, severed head, face out of frame.” Yes, it is a very long text, but each of the sentences that make it up help with the effect. When you send it, you will receive a composition that shows an image of you holding a Pixar character with your face in your fingers. You will also be able to do it with ChatGPTwhich occasionally releases good results. However, the faces are sometimes somewhat deformed, and for now Gemini seems to do better almost always. In Xataka Basics | How to create a character in ChatGPT and Gemini to use it in all the images you make with artificial intelligence

How to add the Three Wise Men to any photo of your street using artificial intelligence

by

Let’s tell you how to add the Three Wise Men to your photographsso that you can create images full of illusion. The idea is that if you have a photo of your street or a place you usually walk through, you can add these characters to it without altering anything else. We are going to tell you two ways to do this, both with artificial intelligence. First we will go to a website designed exclusively for this, which is the easiest alternative to use. And then we will tell you how to use the most popular artificial intelligence chatbots, such as ChatGPT either Gemini. Use a third party page If you want to do things as easily as possiblethere are pages like fotoalosreyesmagos.comcreated especially to add the Three Wise Men to your photos, and which allows you to see the photos shared by other users from all over Spain. The website offers consistency in designs, although the results are a little less refined. To use it, go to fotoalosreyesmagos.com and click on Upload your photo. Now you will go to a screen where you have to upload the photo you want to use to insert the Three Wise Men. Click on the box or drag the photo to it if you are on the computer. Remember that they must be photos of a street or landscape so that the AI ​​can insert the characters into it. Now you’ll have to choose how to customize your resulting photo. To do this, you just have to decide if you want to include the camels or only to the Kings. Additionally, you have to choose if you want the photo to be public indicating your location or if you want to keep it private and not publish it in the gallery. Now, after deciding whether or not to accept or not give the website a donation of one euro, the photo will be generated. When the photo is generated you can download or share itin addition to publishing it if you want in the public gallery. Add the Three Wise Men with ChatGPT or Gemini The other option is use ChatGPT or Geminiin both cases you will be able to use the same prompt, although today ChatGPT Images offers better results. But you can try both options and stick with the one that suits you best. What you have to do in both options is upload a photograph of your neighborhood, and add the following prompt: I want you to add the three Wise Men in this photo. They should be walking down the street, and you should make them realistic, make them look like real people. Look at the proportions so that they have a realistic size within the photograph, that they have the size of a real person. Don’t touch anything else in the photo, just add to the characters. That’s it, with this the AI ​​will generate a fairly realistic image of these characters. The advantage of this option is that you can add and specify things at the prompt you use, adding objects, specifying sizes, and similar. In Xataka Basics | How to create a character in ChatGPT and Gemini to use it in all the images you make with artificial intelligence

The artificial intelligence race is pushing the US towards an unexpected energy solution: looking to the military sphere

by

The artificial intelligence race is not only being fought in laboratories, chips or data centers, it is increasingly being played in the field of energy. In the United States, the accelerated growth of electrical demand associated with AI has exposed a barely visible fragility: the network is not expanding at the same pace as technological ambitions. This imbalance is forcing us to look beyond conventional solutions and reopen debates that seemed closed, including some that connect directly with the military sphere. What has been put on the table. HGP has submitted an application formal to the United States Department of Energy to redirect two nuclear reactors removed from Navy ships to a civil project linked to data centers in Oak Ridge, Tennessee. The request was channeled through a letter addressed to the Department’s own Office of Energy Dominance Financing, and is part of the so-called Genesis Mission promoted from the White House. According to the documentation, the installation could provide between 450 and 520 megawatts of continuous electricity, aimed at intensive and stable consumption. The main argument in favor of this idea is time. Faced with the construction of new civil reactors, whether large plants or smaller designs, which tend to move on long schedules, or the start-up of large gas plants, also conditioned by permits and infrastructure, the reuse of existing reactors is proposed as a way to gain speed. The logic is simple: start from equipment that is already manufactured and tested, and convert it into a firm supply for the network. It is, at least on paper, a way to add base power while other solutions mature. Behind the scenes of the proposal. The initiative does not come from a newly created startup or from an unknown actor in the energy sector. HGP Intelligent Energy It is a recently created division, but it is presented as part of a developer with previous experience in the US market, supported, according to the company itself, in energy storage projects, electric mobility and development of network-scale assets. At the helm is Gregory Alvaro Forero, president of the division, which appears on your LinkedIn profile as president of HGP Storage since November 2013. That detail helps frame the approach outside of the improvised company pattern. What technology would be reused and at what price. The reactors cited in the proposal come from the US naval nuclear fleet, where aircraft carriers operate with two reactors and submarines typically operate with one. Models A4W, manufactured by Westinghouse, and S8G, developed by General Electric, are mentioned. Adaptation for civil use would have an estimated cost of between one and four million dollars per megawatt, and the project would also require between 1.8 and 2.1 billion dollars in private capital for associated infrastructure. The proposal includes revenue sharing with the Government, a fund for future decommissioning and the intention to request a loan guarantee from the Department of Energy, with a first phase “as soon as 2029”. Just because the idea sounds direct doesn’t mean the path is. Bloomberg notes that Reusing military reactors for civilian use would be unexplored territory, and inevitable questions arise: how is it authorized, who operates, under what standards and with what responsibilities if something fails. Coordination between federal agencies and regulators also comes into play, as well as the logistics of moving and adapting equipment designed for ships, not a grid-connected plant. For now, everything remains at the proposal level. Energy sovereignty as a security argument. HGP tries to support its approach with a framework that goes beyond electricity for data centers. In its materials, the company summarizes the idea with an explicit equation, “Energy Supply Chain Sovereignty = National Defense,” and links supply chain resilience to the country’s ability to secure strategic infrastructure, even noting how geopolitical events or social media posts by managers can affect operations and investments. It is the story with which it seeks political and institutional legitimacy. To reinforce the idea that naval nuclear is not synonymous with improvisation, the context of the World Association of Nuclear Operators enters. According to WANOthe US Navy has accumulated more than 6,200 reactor-years of experience without radiological incidents, with 526 reactor cores, as of 2021. The association attributes that history to the standardization of systems, maintenance and quality of training. It is a relevant fact for the public debate, but it does not close it: a solid record in a military environment does not automatically imply that the jump to civilian use will be immediate or easy. Images | General Dynamics Electric Boat | Igor Omilaev | İsmail Enes Ayhan In Xataka | The race to bring data centers to space promises a lot. Physics says otherwise Images | General Dynamics Electric Boat | Igor Omilaev | İsmail Enes Ayhan In Xataka | The race to bring data centers to space promises a lot. Physics says otherwise

Europe is looking for a place to light its “artificial sun” and Spain only has to defeat Italy and Germany to achieve it

by

For decades, nuclear fusion has been the distant horizon of energy: an almost mythical promise, always thirty years ahead. A future without a map. In full electrification of the economy and with demand pushed by the digital industry and data centers, Europe has begun to set coordinates for that promise: where to build the first commercial centers. For the first time, the “artificial sun” is no longer just a scientific experiment and it becomes a problem of territory, infrastructure and industrial planning. And in this new European energy map, Spain appears among the best positioned countries. A new path. Gauss Fusion, the European company created to power the first generation of commercial fusion plants on the continent, has completed the first comprehensive European study of potential sites for this technology, in collaboration with the Technical University of Munich (TUM). The study culminates in a map that did not exist until now. A map that indicates 150 industrial clusters and up to 900 potential sites spread across nine European countries. Behind each point there is an analysis of geology, seismicity, meteorology, refrigeration, access to the electrical grid and existing infrastructure, aligned with standards of the International Atomic Energy Agency (IAEA). Spain on the horizon. It appears as the third country with the most identified clusters: 17, only behind Germany (53) and Italy (22), and ahead of France, Austria, the Netherlands or Switzerland. This is not a political decision or a formal candidacy, but rather a strictly technical diagnosis: where it would be possible to build a first-generation fusion power plant if it had to be done tomorrow. “That Spain appears as the third country with the most potential clusters is due solely to technical criteria,” emphasizes Milena Roveda, CEO of Gauss Fusion, in an interview with Xataka. “The study follows an objective methodology consistent with international standards. There are no strategic weightings or quotas per country,” he emphasizes. And that nuance is key. The map does not look for winners or distribute investments: it identifies where the minimum physical and industrial conditions already exist to host a fusion power plant. But why Spain? On the one hand, its fusion ecosystem. Spain is one of the European countries with greater historical involvement in ITERhouses the headquarters of Fusion for Energy in Barcelona and has achieved key industrial contracts for national companies. Added to this is the role of CIEMATuniversities with leading groups in plasma physics and materials, and the beginning of the construction of IFMIF-DONES in Granadaa critical infrastructure to validate materials for future reactors. On the other hand, their regulatory experience. “Spain has a nuclear regulatory body with extensive prestige and experience,” highlights Roveda. From an industrial point of view, Roveda insists that Spain should not limit itself to being a host: “It has the potential to be a key piece in the merger value chain. Companies like IDOM already have demonstrated that can design and deliver extremely complex systems. Where could these clusters be? The map does not draw isolated points, but rather broad areas. The study identifies regional clusters capable of containing multiple viable locations. In Spain, they appear spread over a good part of the territory – from Andalusia and Extremadura to Castilla y León, Aragon, Catalonia, Galicia, the Basque Country and the Valencian Community – and are concentrated in industrial areas with high electrical demandgood network connectivity and, in some cases, close to old energy enclaves that could reuse part of their infrastructure. Frédérick Bordry, CTO of Gauss Fusion, explains to Xataka that the objective of the map is not to select a specific place, but “to have a broad database that allows collaboration with authorities, companies and other interested parties.” The final decision, remember, will not come until the end of 2027. What would a commercial fusion center be like? Talking about commercial fusion is no longer talking about experiments like ITER. Gauss Fusion works with the concept of a GIGA plantcapable of producing 1 gigawatt of electricity. This implies very specific industrial requirements. “Assuming an efficiency of 30%, a plant of this type must safely evacuate about 2 GW of heat,” explains Bordry. In practice, this requires access to rivers, reservoirs or the sea, as well as robust electrical infrastructure. Unlike fission, fusion does not produce chain reactions, is self-limiting, does not emit CO₂ and does not generate long-lived radioactive waste. “Due to its safety features, it could and should be integrated near urban and industrial centers,” says Bordry, even supplying waste heat for industrial uses or district heating. This aspect connects with a trend that is already seen in Europe: heat recovery in district heating networks, as happens in Finland with data centerseither the use of large industrial heat pumps. The process now enters a delicate phase. According to Gauss Fusion, the goal is to reduce the European map to between two and five final locations by the end of 2026, and make the final decision in 2027. But the technical criteria will not be the only ones. “Political will, the regulatory framework and social acceptance will be essential,” emphasizes Roveda. In his opinion, Europe needs policies that promote fusion as a new industrial engine, and regulations “adapted to the real risk of these facilities.” Social acceptance will also be key. “Transparency and citizen participation are essential,” he says. “We have to explain well what fusion is and what it is not.” A project that covers a lot. For Bordry, no European country can tackle a project of this magnitude alone. The merger will require a continental industrial alliance, something that Roveda defines as a “fusion Eurofighter”, in which Spain should play a central role, not only as a location, but as a technological and industrial supplier. In a context in which European electricity demand could grow up to 75% by 2050fusion is beginning to be seen not as a distant promise, but as one more piece of the energy puzzle, complementary to renewables, storage and electrification. An open closure, but with a … Read more

what it is and how to use it to create artificial intelligence images from your photos

by

Let’s explain to you what it is and how you can use it ChatGPT Imagesthe new section of the artificial intelligence Designed to help you create and edit photos and images. This is a forceful response to Free Nano Banana of Geminiso amazing that it already represents a new evolutionary leap in images created by AI. Since the launch of Nano Banana on Gemini, Google had managed to compete head to head with ChatGPT in creating images from photographs. Gemini was able to use your face and be recognized, something that OpenAI’s AI could not do… until now. We’re going to start by explaining what this new feature is and what features it has to differentiate it from the rest, because there are some very interesting and innovative things. Then, at the end we will summarize how you can use it to create images with different styles from your photos. What is ChatGPT Images ChatGPT Images a new section dedicated to image creation within ChatGPT. This artificial intelligence chat has updated and improved its image creator from photos so much that it has decided to give it an exclusive section. Just as ChatGPT’s normal chat allows you to create images from scratch or from your photographs, This section is exclusive for creating images from photos. Come on, the idea is that when you want to do this, instead of getting complicated by asking ChatGPT, you can enter the section and speed up the process. This is so because In the Images section you will have several ideas for designs and tools to edit your photos. Thus, it will be as easy as clicking on one of your designs, choosing the photo and that’s it, ChatGPT will do the rest. With this, eliminates the need to know how to write a good promptand the process is simpler and more visual for inexperienced people. When you choose the design and upload the photo, it will automatically be sent to ChatGPT with a pre-generated prompt that you can see. Showing you the prompt changes everything This is important, because being able to see the prompt that ChatGPT uses in its presetyou will also be able to copy and paste it to modify it, or even use it in Gemini or some other competing tool. Thus, ChatGPT Images is also not only a good testing ground, but by offering you several prompts it gives you the basis to later generate a much more personalized image from them. You will also know how the image editing prompt works in a more transparent way, and you will be able to use things from both prompts to create a completely unique one. Until now, when you were faced with creating an image from a photo you had to do it from scratch, composing the prompt on your own or searching the Internet to find them. That’s why showing it to you changes everything, because it makes anyone without knowledge can create images very elegant with AI. To all this we must add an interface that also simplifies everything, and in which ideas are shown to you with an image of the resultso that if you see something you like, you just have to click and choose the photo. How to use ChatGPT Images The first thing you have to do is enter the ChatGPT website or application on your device. Here in the side menu Click on the section Images that will appear just below the search options. This will take you to the main screen of the section Images. In it, at the top you have a search field to write a prompt manually, and below you have pre-generated styles of images and ideas of styles or other things you can do. When you choose one of the designs or ideas, you will go to a screen where you simply have to choose the photo you want to use. You can choose any of the last ones that you have used, or click on Choose a new photo to manually upload another photo. And that’s it. When you do so, a chat with ChatGPT will open that includes the photo and the prompt created to generate the type of image you have chosen. In a few minutes you will have the result. You will be able to copy this prompt to reuse it with other images in the chat itself and even modify it to your liking. In Xataka Basics | How to create a character in ChatGPT and Gemini to use it in all the images you make with artificial intelligence

How to create your own Christmas carol with artificial intelligence

by

Let’s explain to you how to create your own carol with artificial intelligenceso that you can have one completely personalized. Here, it is not so much about creating covers, but rather one with original music to which you can add the lyrics you want. For make a song with artificial intelligence you have many tools to generate music. We will start by telling you several of them, and then we will tell you the step by step that you should follow with Sunofor being the most popular among all. However, the steps will be almost the same with the others. Some tips before starting Before you start, you should consider the type of song you want to create. In this case it is evident that we want to create a carolso it is something we should mention in our prompt so that the AI ​​can relate it and generate a type of music according to this. In this case, the key will be in decide the letter you want to put. The beauty of making your own carol is being able to add to the lyrics references to things that those who are going to hear it will recognize, and here you will have two ways to proceed. On the one hand, you can write the lyrics by hand, try to compose some kind of poem that you think would look good in the song. But you can also use normal AI chatbots like ChatGPT, Copilot, DeepSeek either Gemini inter alia. Even Apps like Suno have their own AI for lyrics. In this case, then what you have to do is tell it the concept or even mention things you want added in the lyrics, and let the artificial intelligence do it all for you. Once you have everything clear, then you will have to proceed with choose an AI tool to create musicsomething for which we are going to help you a little later. You will see that in general they are quite easy to use. Don’t settle for the first resultbecause it may not be good the first time. Repeat the process or you promptmake changes to refine what you want, and little by little you will get something better or closer to what you have in mind. Lastly, you should know that the music you create does not belong to you. You are not a musician creating a song with your skills, but you have generated music by algorithms, and therefore in many cases you cannot register it as yours, nor claim its exclusive ownership. You won’t be able to monetize either. music in most cases, since many platforms will detect it as made by AI. Tools to create music by AI Now, we continue with a small list of AI tools to create music from your text requests. They all work in a similar way: you will have to give it a short description of what you want to achieve, include the lyrics and musical style, and the application will do the rest. Suno: Suno It is the best-known AI music creation service, and one of the first that began to surprise users. It has a simple way to describe what you want and give it creative freedom, and another to specify genres, lyrics, and have greater control. Link: suno.com. Riffusion: Very similar to Suno, although it is a lesser known AI tool. It also has two modes: you can use a prompt in which you explain what you want or by using its advanced panel where you can add genres by hand or the lyrics you want. Link: riffusion.com. udio: This program divides the song creation process into three steps. First you define the style, then the lyrics, and then you specify the vibe, the sensation or feeling it conveys. Link: audio.com. Boomy: Another alternative that allows you to create songs by AI, and even has a function to publish them on streaming services. Link: boomy.com. soundraw: Look for a different mechanic to help you create the song, with a genre selector and the ability to choose the duration and tempos to use. Link: soundraw.io. Create your Christmas carol with Suno Create music with Suno It’s simple. First enter suno.com and log in or create an account. Once you do, on the main screen click on the section Create. You will go to a page where you must choose between simple and custom modeswhich appear with the names Simple and custom. If you choose the option Simpleyou just have to describe the type of song you want using natural language. Mention that you want a Christmas caroland if you don’t write the lyrics, you can at least give it a hint by writing the theme you want it about. You can also choose to make it instrumental only, and in Inspiration you can choose a musical influence. When you have it, press the button Create to start the creation. If you choose the option customwhich will take you to if you decide to add letters in the simple one, you will go to a screen with many more options. Here, you will find the following options: Lyrics: It is the field where you have to write the lyrics of the song for the AI ​​to sing them. you have a button Make Random Lyrics to automatically generate letters randomly. Instrumental: This button is used to skip the lyrics and make an instrumental song. Styles: This is another of the most important fields, in which you have to choose the style of music you want to use for your song. Here you can write “carol”. Advanced options: Includes the option to exclude musical styles. Song Title: The title of the song. Workspace: You can create different workspaces to organize your songs, and choose which one goes to. Create: This is the button you must press next to create your song. Now what you have to do is fill in all the fields as you wantpaying … Read more

How to summarize videos with artificial intelligence to know what they say without having to see them

by

Let’s tell you how to summarize videos using AIso that you can know what is said in them without having to see them. Because sometimes you may be looking for a tutorial, a guide, a recipe or just information, but you don’t feel like watching a 40-minute YouTube video. This is something you can do very easily using artificial intelligence. Of course, you will be able to do it with Geminibut ChatGPT does not allow it to be done. The positive part is that you can do it with the free version of Google AI without problems. Summarize online videos First of all, with Gemini you will be able summarize videos that are on online platformssuch as YouTube or Dailymotion. Of course, you will not be able to do it on others like those on social networks like Instagram. But for YouTube and the like it works. The prompt you can use is the following: “I want you to give me a summary of the content of the video in the link. Make the summary schematically using points or bulletpoints. (Link).” In this prompt that we have used, the request to make bulletpoints is optional. However, if you decide to use it in its entirety, you will have a summary that is not so textual, but rather schematic, point by point. This will make the content easier and faster to understand. Summarize videos by uploading files Gemini also allows you to summarize the videos you attach to the prompt with a file. For this, you must add the video filesomething you can do by choosing to upload the file or linking to it from Google Drive. The prompt you can use is the following: “I want you to give me a summary of the content of the video that I attached. Make the summary schematically using points or bulletpoints.” Come on, what you can use the same promptbut with the difference that instead of adding the link to the video, you have to attach it to the message. Thus, Gemini will analyze the content of the video you have sent and give you a point-by-point summary. Ask questions about the video content You can also ask you questions related to the content of the video. Thus, instead of a summary you can ask a specific point or the precise question you have and want to solve. The prompt you can use is the following: “I want you to look for the information in the video I attached, and tell me (question) (link).” In this way, the question you ask will not be answered based on general information on the Internet, but in what is said in the video about it. It is quite useful to extract more precise information. In Xataka Basics | The best prompts to save hours of work and do your tasks with ChatGPT, Gemini, Copilot or other artificial intelligence

When they sold us generative “Artificial Intelligence” we did not know that it was going to be artificial and generative but not “intelligent”

by

A few months ago, a group of Spanish researchers thought of putting an AI chatbot to the test with a curious test. They uploaded an image of an analog clock to the chatbot and asked the AI ​​a simple “What time is it on that clock?” The AI ​​failed disturbingly. Machine, can you tell me the time? Researchers from the Polytechnic University of Madrid, the University of Valladolid and the Politecnico de Milano signed a month ago a study in which they wanted to evaluate how intelligent the artificial intelligence of those models was. To do this, they built a large set of synthetic images of analog clocks—available in Hugging Face— in which 43,000 different hours were shown. Before fine-tuning their behavior, the AI ​​models consistently failed when trying to tell the time. After the adjustment the behavior was much better, but still imperfect. That should not happen with such a “simple” issue for humans. disastrous result. From there they asked four generative AI models what time those images of those analog clocks showed. None of them managed to tell the time accurately. That group of models was made up of GPT-4o, Gemma3-12B, LlaMa3.2-11B and QwenVL-2.5-7B, and all of them had serious problems “reading” the time and differentiating, for example, the hands or the angle and direction of those hands in relation to the numbers marked on the watch. Fine tuning to improve. After these first tests, the group of researchers managed to significantly improve the behavior of these models after performing fine tuning: they trained them with 5,000 additional images from that data set and then re-evaluated the behavior of the models. However, the models again failed consistently when tested with a different set of images of analog clocks. The conclusion was clear. They don’t know how to generalize. What they discovered with this test was confirmation of what we have been observing from the beginning with AI models: they are good at recognizing data that they are familiar with (memorized), but they often fail in scenarios that they have never faced and that are not part of their training sets. Or what is the same: they were incapable of generalizing. Dalí enters the scene. To try to find out the causes of these failures, the researchers created new sets of images in which, for example, they used the Dalí’s famous distorted clocksor those that included arrows at the end of the hands. Humans are able to tell time on analog clocks even if they are distorted, but for AI models that was a huge problem. If they do this with watches, imagine with medical analysis. The danger of these conclusions is that they reignite the debate about whether generative AI models are indeed artificial and generative, but not very intelligent. If they have these difficulties in identifying the hands or their orientations, things are dangerous if what the models have to analyze are medical images or, for example, real-time images of an autonomous car driving through a city. AIs are stupid. Although it is true that generative AI models are fantastic as aids in various scenarios such as programming, the reality is that what they do is “regurgitate” responses that are already part of their training data. As Thomas Wolf, Chief Science Officer of Hugging Face, explained, a generative AI “will never ask questions that no one had thought of or that no one had dared to ask.” Although thanks to their enormous memory and training they can recover a multitude of data and present it in useful ways, finding solutions to problems for which they have not been trained is very complicated. For experts like Yann LeCun, the reality is clear: generative AI it’s very stupid and, furthermore, a dead end. Source: clocks.brianmoore.com AI doesn’t draw watches very well either. Added to the experiment of these researchers is another small test that once again calls into question the capacity of generative AI. It involves asking different models to create the code that allows an analog clock to be displayed with the current time. A designer named Brian Moore wanted to share the result of several AI models and the truth is that the result obtained in most of them is terrible, although others like Kimi K2 achieve a good result. We have tested with the recent Grok 4.1 and GPT-5.1. After a little insistence, Grok 4.1 has drawn the perfect clock and it works. With GPT-5.1 there has been no way, at least in our tests. A worrying reality. This inability to solve tasks that seem simple certainly means that these models are not in a good place. It is true that a good prompt can help resolve some of these limitations, but what is becoming increasingly evident is that AI models continue to make mistakes despite the passage of time. The theoretical revolution of this technology precisely needs to eradicate them, and it does not seem that we are on the way to achieving it. The models improve, yes, but not enough for us to trust them 100%. Image | Yaniv Knobel In Xataka | As if there weren’t enough AI companies, Jeff Bezos has just returned from the shadows to build another one, according to the NYT

What are the news about Google’s new artificial intelligence model?

by

Let’s tell you what they are the main news of Gemini 3the new version of the model artificial intelligence announced by Google. We already have the first data on its main characteristics. As always, the flashes will go to Gemini 3 Pro, which will be the most advanced version. Here, one thing you want to know is that You will notice few of these new developments when you are using Gemini in a conventional way. Most of these changes are aimed at advanced users. Gemini 3 news A step forward in all areas: Google has presented the results of its model in various types of tests, comparing it with the previous version and its direct competition. He is ahead of everyone in everything, from mathematics to understanding what is happening on the screen or creating code. Reason “at the doctoral level”: That is what the test results also indicate, although where it advances the most is in the mathematical results, with a score of 23.4% for the MathArena Apex test compared to 1.0% for GPT 5.1 or 1.6% for Claude Sonnet 4.5. Integrates with Google Search: Gemini 3 is linked to Google’s AI Mode, integrating into the search engine. Generate visual elements: Gemini 3, has the ability to create interactive visual elements, such as calculators, simulations or widgets in real time. Something especially useful when integrated into the search engine. Sometimes it may not respond to you with text, but with an interactive webapp. More direct answers: Google has fine-tuned the way its model responds, offering more concise responses that offer more valuable information and less flattery, clichés and clichés. Improvements in “Deep Thinking”: Another of the most notable improvements is in deep thinking, in addition to advances in code execution, abstract reasoning and visual understanding. Larger context window: This model has a context window of up to one million tokens, being able to analyze large code repositories or very long texts on which you can later work. Better contextual reasoning: Reasoning is improved, especially with long contexts, to avoid hallucinations. Parallel reasoning improvements: Abilities to reason with visual and textual data at the same time are improved, improving accuracy when interpreting tables, diagrams and interfaces. Improvements in its multimodal mode: The analysis of all types of information is improved. For example, you can decipher and even translate handwritten recipes in different languages, and use them to create a cookbook that you can share. You can also analyze sports matches, scrutinize research data and generate code from it. Programming improvements: As we said at the beginning, one of the biggest improvements of this model is in its ability to program. Improved your agent mode: Your ability to use tools and operate a computer through the terminal using agent mode has also been improved. Agents with Gemini can now autonomously plan and improve more complex software tasks. Gemini 3 will begin to be available in the coming daysalthough as we told you at the beginning, it is possible that many of the differences will not be noticed unless you are going to try to take advantage of them in an advanced way. In Xataka Basics | The best prompts to save hours of work and do your tasks with ChatGPT, Gemini, Copilot or other artificial intelligence

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.