The new great models of generative the AI ​​do not stop delaying. It is a dangerous indication that we have touched the roof

We expected to have GPT-5 available at the beginning of the year, but OpenAi gave us GPT-4.5. The generative AI model, which theoretically represented a remarkable leap with respect to its predecessors, ended up disappointing and the company announced that it will eliminate it from its API in July. It was too expensive and simply did not compensate. That was already a bad sign of AI advance, but there is more. And GPT-5, what? It was expected that GPT-5 will arrive in the middle of the year. Sam Altman has been creating Hypebut in December we knew that the arrival of that model was being problematic. The jump in benefits I wasn’t being expected And the cost of developing it is huge. What did they do in Openai? Delay it and launch in its place GPT-4.5 That, as we have seen, it was one of the great disappointments in the history of Openai. Bad signal one signal. Behemoth is delayed. As indicated In The Wall Street JournalGoal will delay the launch and deployment of its most ambitious model to date, call 4 Behemoth. This “monster” with 288,000 million active parameters (two billion in total) is the third member of the newly presented family calls 4. However, according to WSJ “Company engineers are having difficulties in significantly improving” their capabilities. It should have arrived in April, but now it is estimated that it will arrive in autumn, or even later. Frustration. Sources close to the company indicate that managers are frustrated with the performance of the team that is developing calls 4 Behemoth. “Significant management changes” are already contemplated that would raise internal movements (and who knows if layoffs) as a result of these bad results. And not that the available flame models 4 are having a good receptionlet it be said. Bad signal number two. Unbalanced. In WSJ also highlight how the first flame version was created by its fundamental research team, formed by academics and researchers. Since then 11 of the 14 researchers have left the company. Anthropic does not advance either. We also expected a “round” leap in Claude, the Chabot of the generative of Anthropic, but in February the company presented Claude 3.7. It is true that this model Yes offered striking benefitsbut at the moment its Opus version, the most ambitious, still does not appear, and nothing is known about Claude 4.0. Bad signal number three. Jumps not, at most jumps. What we are seeing in recent months are not significant leaps in the capacity of the models, but striking improvements only in some sections or effective characteristics. It happened with Gemini 2.5 Pro, especially powerful in programming and that has allowed Google to win integersbut also with Openai and the famous images that imitate Studio Ghibli Oa Grok 3, who has become more famous for his lack of censorship that for its accuracy or quality (which is not bad). Deceleration. All this triggers the debate on a potential “deceleration” of AI: the climb no longer seems to work so well, and that of using more GPUS and more data to train models is not offering the expected return. Jaime Sevilla, CEO of Cophai, I did believe That the rhythm of improvement was being expected, but these delays of course make the future progress of generative the AI ​​again. The agents and the AI ​​that “reason” are hope. The models with “reasoning” capacity Yes, they have allowed us to propose striking improvements in some areas, and companies have launched to present this type of variants and deep research modes For specialized uses. The other great hope of 2025 are the agents of AI capable of completing task sequences autonomously to solve a problem, even connecting to other services or data sources. At the moment we already have outstanding examples in the schedule of programmingbut practical applications for end users are limited. Image | Goal In Xataka | There are too many AI models. That raises a true death sentence for Anthropic and Claude

Sending the ID to a stranger is dangerous. This new app has a free solution: put water brand

The DNI is a document as important as dangerous is to share it with third parties without taking concerns. Surely we have all ever faced someone, whether a person or establishment, who asks us for a copy of the ID to, for example, complete a reservation or a purchase. It is an extended practicebut not sure. The consequences can be disastrous, so it is recommended edit the document Before sending it. There are several optionsbut This tool It promises more security and put it more difficult for the IAS for the modest price of zero euros. The most interesting thing is that, although it works through the browser, there is no cloud in between. Everything works at home. His name: Saferlayer. “Hey, the copy of the ID …”. A scammer can use our ID to commit crimes (such as impersonation of identity), access to an betting house or ask for a loan in our name. Nothing good, of course. Therefore, it is important that if we are forced to share a photocopy of the DNI we do it prior edition. This type of DNI -related scams They are relatively known on the purchase and rental platforms. The lack of supply, the enormous demand and despair to find a floor can make, with the rush, let’s forget to take the necessary precautions and we will find two negative things: that the announced floor did not exist and that a total unknown has Our ID. The general recommendation is never to send a copy of the DNI or passport without editing What is the problem? Not everyone knows, wants or can use dedicated tools like Photoshop. Pixelar an image or a piece of an image It is very simple (and not as sure as it seems). Also add text or a watermark, but you have to know. Not only that, but these water marks can be edited and removed afterwards and leave the image as if nothing had happened. The solution: a dedicated tool. Carlos Sánchez, Head of Product in Real Madrid, and his “Compi de Side Projects“Guido Fioravantti, got to work after Carlos had to send a copy of his email documentation. In telephone conversation with Xataka, Carlos Sánchez explains that” you take any traditional water brand and pass it through some tools for some tools Specialized artificial intelligence, you’ll see that they take it off in seconds. What we wanted was to avoid that. “ After several tests and iterations, they found a “water mark that covers the entire document, which deceives the AI; makes it difficult to differentiate the water mark from the background image, so when you try to remove it, load it is loaded part of the document and is proof that it has been manipulated in some way. “ Copy of the original ID (above) and edited with Saferlayer (below) | Image: Saferlayer The water brand. It consists of a wave pattern with a degraded one that goes from fuchsia to green. “This has been a very proof and error process,” explains Carlos. “In that process we detected that this undemed undulating pattern, that flag effect waving, made it difficult for the AI. These color changes throughout the pattern also complicate it. In the end, we had to find the balance between whatever Very difficult to remove for AI, but also that it was readable, “he says. For Carlos, it was important that it was “evident to the person generated by the document and for whom he receives that there is a watermark that can be read, especially so that the person who receives it does not have incentives to use it for purposes Fraudulent “. This is the document edited with Saferlayer when we pass it through one of the many existing tools to eliminate the water mark | Image: Saferlayer, Xataka Everything at home. One of Saferlayer’s keys is that It works at homenothing leaves the browser. The image we are going to edit does not climb any server. This can be easily verified with the inspector of the browser itself or, simply, adding the Webapp to the home screen and putting the mobile in plane mode. We will verify that it continues to work without problem. When you upload an image, we simply have to specify the use that the copy has, for example, “Valid copy only to confirm the ABCDE hotel reservation.” Once this is done, the tool is responsible for adding the watermark. It is practically instantaneous. The tool works with all types of documents in image format (JPEG and PNG), but not with PDFs. Water brand detail | Image: Xataka A success. The tool was launched on Sunday afternoon and spread mainly Through x. It soon went viral and that can be seen in the figures. As Sánchez confirms us, “right now we are doing a water mark every nine seconds.” To the elderly, the idea of ​​the creators, their vision, is to develop a tool that allows the “total traceability of the document. That allows the individual to protect their document add a water mark, but we also have the idea of ​​covering the other part , to the company that receives the document and wants to make sure whether it is legitimate or not. ” In Carlos’s words, “we would like to evolve this to verify that the documents that have Saferlayer water mark have been manipulated in some way, or if it is the water mark as it has been generated by the tool.” “Right now we are doing a water mark every nine seconds” In that sense, Carlos tells us that a hotel in the center of Madrid has contacted them because he was “very interested in exploring tools that facilitate these processes (in reference to the new lodging record) since it generates a brutal administrative burden and makes Check-in the process more uncomfortable for customers. “According to Carlos, the hotel wanted to talk to them to see if Saferlayer could” potentially help them. “ … Read more

Dangerous forest fires registered in San Diego, California

During Monday night and early Tuesday morning, multiple wildfires broke out in northern San Diego Countywhich occurs amid a new wave of Santa Ana winds that has generated meteorological alerts about the risk of fires. A new fire known as the Friars Fire began spreading rapidly Tuesday afternoon near the Fashion Valley Mall, quickly reaching two acres, reported Fox Weather. Authorities issued evacuation orders for neighborhoods adjacent to the affected hillside.. However, thanks to the quick intervention of fire crews, the fire was contained and cleaning work is currently being carried out in the area. On the other hand, the Lilac Fire originated in Bonsall and led local authorities to issue evacuation orders for those south of Pala Mesa and west of Interstate 15. So far, at least 86 people have been evacuated due to these incidents, and two buildings have suffered significant damage.. According to the North County Fire Protection District, which is collaborating with CAL FIRE, the Lilac Fire began shortly before 1:30 a.m. local time and has consumed approximately 85 acres so far. Despite the imminent danger to life in the area, firefighters reported that the fire is 50% contained. They close schools and establish shelters The San Diego County Fire Department announced that the spread of the fire has been stopped. In response to the situation, several schools in the area have closed their doors. Temporary shelters have been set up for both people and animals affected by the evacuations; A specialized team was even deployed by the San Diego Humane Society to rescue six horses. Just a few miles north, another fire known as the Pala fire broke out within an hour of the Lilac Fire starting. This fire affected approximately 17 acres before being completely contained by emergency crews. Additionally, a third fire called Riverview broke out in Fallbrook; although it threatened some nearby structuresits spread was controlled after burning around an acre with no injuries reported so far. Keep reading: – FireAid, support concert for the fires in California, announces venues and poster– Editor of ‘The Brutalist’, possible Oscar candidate, confesses that he used Artificial Intelligence– Fires in California: Why has the hiring of private firefighters caused controversy?

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.