Open

Vibe coding wants to help Open Source. But developers don’t want AI botches

by

If you like Open Source software, vibe coding now gives you a fantastic opportunity: to take that code and modify it to your liking with the help of vibe coding and the AI ​​agents that program. Let them tell it to me. You may have good ideas and the AI ​​will solve them with new code generated with these tools, but there is a problem: the quality of that code may not be adequate. what has happened. Steve Ruiz (@steveruizok) is the creator and responsible for TLDrawa nice Open Source application that allows you to turn your browser into a canvas so you can easily draw whatever you want on it. On January 15, Steve posted a message on X in which he announced something very striking: he would stop accepting code contributions (pull requests, PRs) in the TLDraw GitHub repository. We don’t need low quality code. “Due to the influx of pull requests of low quality, we will soon close those requests to external contributors,” said the person responsible for a project in an additional post on the official blog of the project. The message was clear: although people’s intentions are surely good when trying to contribute their ideas to an existing project, this developer soon realized that the code contributed by new programmers, fans of vibe coding, was of low quality. Solution? Ban those contributions made with AI. AI-generated code can serve. In that article I indicated that this was not a measure against vibe coding, but against code (any code, human or AI) of poor quality. Ruiz explained how: “We already accept code written with AI. I write code with AI tools. I hope my team uses those AI tools too. If you know the project’s code base and know what you’re doing, writing great code has never been easier thanks to these tools.” AI Slop, but from code. Although we often talk about “slop created by AI“(AI Slop) in reference to low-quality text, images, music and videos, the term can also be applied to code. Ruiz explained how in September he began to detect many requests for code contributions that seemed correct but that after a deeper analysis, although they worked, could potentially introduce future problems and complexity to the project. I correct here, I correct there. In addition, many of the contributors had profiles in which they could be seen jumping from Open Source project to Open Source project and then disappearing. They simply contributed without following the policies or requirements of the project and moved on to another. This is a plague. In the debates that this decision generated in Hacker News and x Ruiz found himself with a surprise: people not only did not protest, but they valued the measure positively. He commented how “this seems to be the standard experience for all public repository maintainers right now.” He cited the example of Excalidrawanother similar project that “received more than twice as many PRs in the fourth quarter of 2025 than in the third” in your repository. More and more vetoes of low-quality AI code. Other projects are going through that same phase. ghosttya terminal emulator for macOS and Linux, recently published their “AI policy” in the site’s public GitHub repository with important notices. For example, that “PRs created by AI must have been fully verified with human use”, and further that “all use of AI in any way, shape or form must be disclosed.” That’s cheating. Curl, a very popular utility for command line users, had announced the opening of a bounty program to detect bugs and vulnerabilities in its code. What have many people done? Use AI to find them and take the money. Those responsible for the program have announced that They will close it this month in the face of the avalanche of low-quality vulnerability reports clearly generated by AI. Linus already said it. Linus Torvalds, creator of the Linux kernel, admitted to using vibe coding tools for some small personal project. While recognizing that these tools can be great, he warned of the danger of all that AI-generated code: “AI will be a tool, and it will make people more productive. I think vibe coding is great for getting people to start programming. I think (the code it generates) is going to be horrible to maintain… so I don’t think programmers will go away. You’ll still want to have people who know how to maintain the output.” AI code works, but it is not usually “quality”. The developer community has been warning and experiencing this for some time. Although AI tools can help program and solve many routine tasks, the generated code must be reviewed by a human programmer to avoid future problems. It is reasonable to think that this code will be increasingly better and of higher quality, but today in many cases the situation is clear: it may work, yes, but that is not enough for many projects in production, especially when they are used by thousands (let alone millions) of people. In Xataka | Bill Gates and Linus Torvalds had been rivals for 30 years. The funny thing is that they just met and took a selfie

Qwen and open models

by

Alibaba’s Qwen family of open AI models is quietly taking the world by storm. Until the current month of January these models have overcome and to 700 million downloads on the Hugging Face platform. The milestone is significant and confirms the supremacy of Chinese companies in this type of models. what has happened. Those responsible for the development of this family of models explain how Hugging Face data They don’t lie: these projects have become the most popular open models worldwide, at least if we look at the number of downloads. Unstoppable. In October 2025 the Qwen model family managed to surpass the previous leader in this segment, the Llama de Meta family of models. Two months later, the Qwen models had been downloaded so much that the total number exceeded the combined figure of the eight AI models next in popularity. That group is made up of the models from Meta, DeepSeek, OpenAI, Mistral, NVIDIA, Zhipu.ai, Moonshot and Minimax. Alibaba is a steamroller. Since Qwen launched in 2023, the advancement of the models in this family has been unstoppable. Although initially accessing them was more uncomfortableAlibaba has taken advantage of its infrastructure and its size to popularize them little by little, but above all its engineers have done something else: not stop launching models. The pace has been frenetic, but the models are also notable and comparable to proprietary models from major US technology companies. Qwen wants to be the Android of AI. The family’s model catalog is enormous. Hugging Face currently includes 300 different models that cover both slightly older versions and various variants of each new major version. For example, we have models specialized in visual data recognition (Qwen VL), in programming (Qwen Coder) or models in image generation (Qwen Image Edit) or “generic” models absolutely gigantic or others like Omni that already compete with Grok 4 or GPT 5 Pro. The intention is obvious: to be the “standard” AI models on the market. Even if it is due to saturation. Surprise: the most popular is not the most powerful. a study Published in October, it did provide surprising information on the growth in popularity of this family of models. One would expect that the most downloaded would be some of the versions of Qwen 3, the most modern and capable version. Actually the most downloaded is Qwen 2.5-1.5B-Instruct, a “light” model that can even run on modest mobile phones and laptops. Tiny but bully. Currently the Hugging Face list indicates that the most popular in downloads is Qwen2.5-3B-Instruct, more modern and somewhat less lightweight, but still “small” by today’s standards. It seems clear that there is notable interest in being able to run this model on mobile phones, tablets and computers with little video memory. Thousands of derived projects. The possibility of obtaining and using these open models in a simple way has made many developers and companies take advantage of them to customize them and adapt them to their own needs. That has made according to Xinhua That family has been used in more than 180,000 derived versions. Flame fades. Meanwhile, Silicon Valley confirms that its vision is different. Goal, what led this area initially thanks to Llama, has given a rudder stroke. The company is still expected to launch new versions of this model, but in the meantime most technology giants focus on their closed and proprietary models. Of course: they keep some models open with a promising future, as we have seen with Gemma (Google) Phi (Microsoft) or gpt-oss (OpenAI). Without forgetting that Mistral is a great European benchmark and also offers open variants. In Xataka | China and the United States have started an antagonistic race in AI through a simple question: whether to be open source or not

At the moment it has an open field and four streets of a PAU

by

The official website says that “you no longer have to imagine it. Now you can live it.” But the truth is that, as the matter stands, it is difficult to imagine. Let’s hope we can live it. Because when there are nine months left, Madringthe capital’s Formula 1 street circuit looks like anything but a circuit. At these moments, it becomes difficult to think that through those streets and those half-assembly curves you can search Fernando Alonso his 33. If everything goes as planned, the September 11 The cars have to start rolling south of the city. Among the streets of Valdebebas, one of the Madrid PAUsand the bowels of IFEMA, which competes with the Fira de Barcelona for being the largest fairground and convention center in Spain. The layout is a good reflection of everything that Formula 1 is rewarding: urban layouts that allow attracting large investments from cities to export their image to the world even if, as in the case of Madrid, all the attractions around the circuit are residential buildings, a fairground and a half-built City of Justice for more than 20 years. Now, with 243 days left until Madrid returns to the Formula 1 calendar 45 years later, a doubt is beginning to float: whether Madrid will return to the Formula 1 calendar. Some works in diapers The FIA ​​said that everything is going well, that there is no problem. That’s what they collected in I amMotora portal specialized in motorsport competitions, a little less than a month ago. “There are no delays or concerns within the FIA,” they then stated from the media, which also claimed that a commission from the International Automobile Federation had been supervising the status of the works in the streets of Valdebebas. From IFEMA they also defend that they are advancing at the expected pace. “The works are going within the established deadline. The paving has begun and part of it is already finished, although it is being done little by little due to the rain,” they point out to Autobild. And they make it clear that the circuit still remains to be completed: “the last layer of asphalt is expected to be completed during the summer.” That IFEMA has come out to speak is no coincidence. The information points to a closing of ranks after the viability of the project was questioned. At least for this year. The Italian media Rmc Motori claimed last November that Liberty Media, owner of the rights to the sport, was considering removing 2026 calendar to Madrid in favor of Imola, given the progress of the works. The legendary Italian circuitwill not be part (at least for the moment) next year of the F1 calendar. What is certain is that the works seem to be in an embryonic phase. Those who are walking around the circuit these days are finding the streets of Valdebebas without any type of modification. Ready for you the cars pass at full speed but not single-seaters. Click on the image to go to the original tweet With few exceptionsit is difficult to intuit the circuit along the 22 curves that make it up. The Monumental, a banked curve with a 24% inclination that has become one of the great attractions of the circuit, it’s a muddy mess right now. Yes, progress has been made on the route but there is no sign of progress in the surrounding services and the first asphalt is conspicuous by its absence. The times are also much tighter than we might think. In August the circuit must be ready so that Eurocup-3, a single-seater category inferior to Formula 1, can compete in one of its grand prizes. If it arrives, the intention is to make it the first big test before Fernando Alonso, Carlos Sainz or Max Verstappen set foot on the soil of Valdebebas. Click on the image to go to the original tweet The circuit, in addition, has to fight with the opposition of the neighbors. Pave the way for cars to pass is causing profound changes in its streetsconstant works and the anticipation that the noise suffered during that weekend will be much higher than the averages that have paralyzed the concerts at the Santiago Bernabéu. Besides, environmental associations They defend that the project threatens the conservation of wetlands and “non-transplantable” trees in the area. Nor is Madrid the first city where the viability of a Formula 1 Grand Prix is ​​doubted a few months before its celebration. In South Korea, Yeongam circuit was not reviewed by the FIA up to 10 days before the traffic light went out on the finish line in a clear example of “out of sight, out of mind” heart. That same weekend work was being done on the track and in some areas the asphalt was not well established. In Las Vegas, Formula 1 has been fighting with a recurring problem for three years now: the sewers become loose with the passage of cars. And in Hanoi, 600 million euros were spent on a circuit so that five years later a total of zero cars raced before its abandonment. Photos | Ifema In Xataka | Madrid says that F1 will not be paid for with public money. Valencia promised the same and it cost them 300 million euros

The great plan of Chinese brands to open hundreds of dealerships in Spain: a movement against the current

by

Search, compare and if you find something better, buy it It was the 80s and Colón’s detergents had snuck their famous slogan into every house in Spain. 40 years later, a BYD worker explained to me how they sought to break down prejudices in Spain: “We are letting people take the car home. We don’t want to do the typical 20-minute test with the salesperson inside the car. We tell the customer not to be afraid, to take the car and bring it to us the next day” BYD barely had a handful of points of sale in Spain those days in 2023. Shortly before we had attended the official presentation of the brand. The Chinese company arrived with three electric cars (two of them with a clear premium focus) and I saw it clearly: the brand had to attract the customer to the dealership. Let him sit in the car, touch it and feel it. It was the only way to dynamit prejudices. We are just over a month away from the end of 2025. At the end of October, BYD has sold 22,357 exclusively plug-in units in Spain according to data from Anfac. They easily double Fiat. They surpass Mazda and Volvo. They left Tesla behind a long time ago. They have Opel or Cupra on the near horizon. They begin to approach Ford. At the same time, BYD will close 100 dealerships this year (96 are already active throughout Spain) and they plan to open another thirty next year. At the same time, Chery has placed 31,493 cars in our country at the end of November between Ebro, Omoda and Jaecoo. And we are facing the first full year in which they have sold cars in our country. The sum of all of them also easily exceeds one hundred points of sale. MG adds 38,989 units between January and October 2025. With 11 points of sale available throughout Spain. The irruption is such that 10% of the cars purchased in Spain they are Chinese. It was a figure that was difficult to imagine just a few years ago. A figure that has been achieved by taking the customer to the dealership. And dynamiting their prejudices. The importance of being on the street There are many factors that explain the brutal growth of Chinese brands in our country. We can talk about its low prices, about offering a gateway to a technology (electric or plug-in hybrid) that has made the product more expensive or the extensive equipment offered in each car. But in addition to the price, which overrides all the previous arguments, we find an expansive effort by all these brands to be on the street, at the customer’s feet, with the dealers. “We had been waiting for its arrival for a long time. Already 15 years ago, in the brand’s first foray, we had one of its models, now I don’t remember exactly the name. In terms of volume, manufacturing capacity and development, it is a really excellent product. We think it is above the rest of the Chinese brands that are arriving in our country.” The speaker is José María Blitz, Project Director at the dealership that BYD has on Concha Espina Street in Madrid (next to the Santiago Bernabéu) and which belongs to Astara Retaila distributor with a presence in 19 countries and that sells you a Bentley or a Rolls-Royce as well as a BYD. This time it is the Chinese brand he is referring to. He tells us that the public has welcomed the company with open arms and that since they opened their first dealership, this one next to the Madrid stadium, interest has only increased. “The client It has already surpassed that of ‘unknown brand’. There could have been one at the beginning, three or four years ago, but I think it is practically expired. What’s more, the customer’s perception of the brand is excellent,” he explains. Added to that, “the European product was very expensive and the equipment was very fair. Chinese brands offer a quality product at a competitive price with a much higher level of equipment. We can easily be 20% or 25% below competing brands with higher equipment levels,” says Blitz. BYD is just one of the Chinese brands that sell in our country. Together with MG and the Chery Group (Omoda, Jaecoo and Ebro) they form a kind of quintet representative of the 28 Chinese brands that already sell in our country, according to Faconauto. The dealer association of our country says it has about 600 points of sale right now, between dealers and official services. Of course, they point out to us that “it is convenient to contextualize this figure: the majority of points of sale in Spain continue to correspond to consolidated manufacturers – European, Japanese, Korean or American – whose presence is structural and has developed over decades. Right now, Faconauto has 28 Chinese companies selling cars through 600 dealerships spread throughout the country. What happened to find us with this explosion? Blitz is clear, the product, he assures, is part of the success. But also who these brands have partnered with. “Their strategy has been to close agreements with large groups of dealers, people who are really professional,” they say from this dealership owned by Astara. The same is the opinion of Faconauto, who point out that “they have decided to enter our market with a ‘traditional’ model, taking advantage of the establishment of business dealers. And the key is the word ‘businessmen’, who choose where to invest their money. It is evident that many dealer groups have seen a good opportunity in investing in the distribution of Chinese brands.” This commitment generates trust in the customer, which has allowed them to grow “to the level of any other European brand,” for Blitz. The key: a disruptive product and good after-sales service. “They are very agile and they want their employees to be agile too. There is a sense of … Read more

Europe has left a crack open to using combustion engines in 2035. It is a goal pass to China

by

The European Commission has spoken. Now it is up to the rest of the European organizations to buy the proposal. Everything indicates that this will be the case and that we will have a relaxation in emissions standards in 2035. One that points to very expensive combustion engines and highly electrified options. Options in which China leads. The approved. First, we must start with what has been approved. It is the proposal of the European Commission regarding the emissions targets that manufacturers must meet in 2035. This points to a slight reduction. With the 100% reduction in carbon emissions that was approved, the combustion engine was almost doomed. Why does an electric car have less autonomy than advertised? And it is that only those moved by efuel they could work if they were carbon neutral. With the changes, the average emissions of the manufacturers’ fleet must move in 11 gr/km of CO2. These are figures almost impossible to achieve for any car that is not purely electric. Therefore, most options involve selling the vast majority of electric vehicles and a touch of combustion. Expensive and exceptional. Combustion cars “will become the Swiss luxury watches of the automotive industry.” The words are by Matthias Schmidta market analyst who points out that the rule is nothing more than a “Porsche amendment.” This explains the exceptional nature of the combustion cars that will be sold on the street. And the use of “green steel” and synthetic fuel, produced in Europe, will be key to receiving emissions bonuses that increase the average CO2 allowed to each manufacturer. Requirements that, presumably, will make the cost of the car even more expensivewhich will have to be passed on to the end customer. That leads to two paths. One, as we say, is to offer a few very expensive combustion cars as a status symbol. The second aims to sell exclusively electric cars. Or, if necessary, a type of plug-in hybrid called extended range electric. A type in which, again, China has the lead. The extended range. The extended range electric car is a type of car designed by and to be used as an electric car. The objective is for it to be supported by a combustion engine but only to be used as an emergency measure. Mazda sold us the MX-30 R-EV using this name but the cars of 2035 will have to go one step further. And it is that the SUV electric Mazda plug-in hybrid It already approves 21 gr/km of CO2, a figure that will skyrocket when the new approval criteria come in. The alternative for those looking for a car with a combustion engine for peace of mind or because their needs demand it will have to go for a type of extended range electric vehicle forgotten in Europe. This extended range is what was already proposed with the BMW i3 REX. The BMW electric car, ahead of its time, did have a combustion engine but it barely had 38 HP and was supported by a 9-liter tank. Because the fundamental idea is that the engine would act as an electrical generator in emergency conditions, when the battery had run out and there was no outlet nearby. China, always China. This type of car is one of the few with combustion engines that aspire to be relatively affordable. Right now, in the Spanish market, the best example is the Leapmotor C10 REEV. This car, as in the case of Mazda, has a 50-liter tank for an 88 HP engine, but its usage pattern has allowed it to approve 0.4 l/100 km of consumption and 10 g/km of CO2, a real rarity in the market. Given this expected increase in the approved emissions figures, this type of car will have no choice but to expand the battery (in the Leapmotor it is only 28.4 kWh) and reduce the gasoline tank. While maintaining its operation as a pure electric vehicle and, if necessary, as a series hybrid. This technology is used by many cars in China. In this list you have the most purchasedamong which are cars of all price ranges. We find cars like himLi Auto L6 EREV with 212 kilometers of electric autonomy but that extends over a thousand thanks to its combustion engines or the Aito M9powered by Huawei. BYD with its YangWang U8 It shows that there is a market for all types of options. The series hybrid. If the Leapmotor manages to reduce its consumption and emissions to such low figures with a 50-liter tank, it is largely because of how it uses its technology. China has specialized in serial hybrids, a small rarity in Europe. Toyota, for example, combines the technology with the parallel hybrid, where the combustion engine can drive the wheels at the same time as the electric motor but separately. In a series hybridthe gasoline engine works as an electrical generator that provides electricity to the battery. The electric motors draw power from this. And the hybrids that are coming to us from China, both plug-in and the Omoda 9 SHSas non-pluggable, as the Omoda 5 SHS-Happly this system to try to improve their efficiency. What they achieve is that the combustion engine operates at a speed range that is considered optimal, where they deliver the greatest power with the lowest possible consumption. When more power is needed, the car can deliver it and increase the engine revolutions but they try by all means to prevent this from happening. The driver, for his part, has the feel of an electric car, with less noise and vibrations, which is a plus in comfort. One more time. As we say, these cars will have to increase their electric range and reduce their gasoline tanks to operate very punctually with this system and reduce emissions, but again China is one step ahead of Europe in this technology. Leaving the door open for this configuration to be an interesting alternative to have a minimum safety net with … Read more

The elite of the open models spoke in Chinese. Mistral has just placed Europe at a level that not even the US managed to reach

by

Over the last year, the elite of open models for assisted programming, at least in benchmarks as SWE-Bench Verifiedhas spoken with a Chinese accent. Names like DeepSeek, Kimi either qwen They had settled into the top positions in testing and were setting the pace in complex software engineering tasks, while Europe was still searching for its position. The arrival of Devstral 2 alters that distribution. It does not displace those who were already at the top, but it places Mistral at the same level of demand and turns a European company into a real contender in a field that until now seemed reserved for others. League change: the technical leap that had been brewing for some time. During recent months, the open models developed in Europe and the United States had shown constant evolution, although still without the performance necessary to compete in the most demanding tests. The progress was evident, but there was a lack of a project capable of consolidating it at a higher level and demonstrating that this path could give results comparable to those of the sector. Devstral 2 in data: performance, size and licenses. The new Mistral model reaches 123B parameters in a dense architecture and offers an expanded context of 256K tokens, accompanied by a modified MIT license that facilitates its adoption in open environments. Its compact version, Devstral Small 2, reduces the model to 24B licensed parameters Apache 2.0. In the SWE-Bench Verified figures published by the companyDevstral 2 obtains 72.2%, a mark that places it in the most competitive section of the open models evaluated and that confirms its presence among the most advanced alternatives in the segment. It is reflected by a panorama concentrated in the upper part of the benchmark. Among the open models, DeepSeek V3.2 leads the group with 73.1%, followed by Kimi K2 Thinking with 71.3% and for proposals such as Qwen 3 Coder Plus and Minimax M2, which are around 69 points. At lower levels GLM 4.6, GPT-OSS-120B, CWM and DeepSWE appear, with more moderate results. In the closed commercial environment (proprietary models), the graph incorporates higher scores: Gemini 3 Pro reaches 76.2%, GPT 5.1 Codex Max rises to 77.9% and Claude Sonnet 4.5 scores 77.2%, all of them above the best brands registered for open models. What SWE-Bench Verified Really Measures and Why It Matters. SWE-Bench Verified is a test designed to evaluate whether a model can solve real programming tasks, not synthetic exercises. Each case presents a bug in an open source repository and requires a patch to pass the previously failed tests. The evaluation seeks to measure whether the system understands the structure of the project, identifies the cause of the problem and proposes a coherent solution. It is a useful and demanding metric, although limited to Python repositories and a specific set of situations that do not cover the full breadth of software work. From co-pilots to agents who act on the project. The arrival of Devstral 2 coincides with a broader change in the way of working with programming tools. It is no longer just about receiving suggestions in the editor, but about having agents capable of exploring an entire repository, interpreting its structure and proposing changes consistent with its real state. In this context, Vibe CLI appears, a tool that allows Devstral to analyze files, modify parts of the code and execute actions directly from the terminal, bringing these capabilities closer to the daily workflow of developers. Cost and deployment: what each type of user can do with Devstral. The model will be available for free for an initial period and will then cost $0.40 per million tokens for input and $2.00 per million for output, while the Small 2 version will be priced lower. Its deployment also makes a difference: Devstral 2 requires at least four H100-class GPUs, aimed at data centers, while Devstral Small 2 is intended to run on a single GPU and, according to Mistral documentation, the Devstral Small family can also run in CPU-only configurations, without a dedicated GPU. This variety allows both companies and individual developers to find a suitable entry point. The appearance of Devstral 2 introduces an unexpected element in a space where Chinese companies set the pace and where not even the United States, despite its leadership in artificial intelligence, had an open model in this high performance range in SWE-Bench Verified. Mistral does not displace those who were already at the top, but it does broaden the conversation and shows that Europe can compete in a field where it did not appear until now. It is a movement that does not alter the general hierarchy, although it does open a new margin for the evolution of assisted programming tools. Images | Xataka with Gemini 3 In Xataka | OpenAI and Google deny that they are going to put ads in ChatGPT and Gemini. The reality is that accounts do not come only with subscriptions

There are so many English people living in Alicante that the largest British pub chain has decided something: open there

by

The millions of British tourists who land in the province of Alicante each year will now have a piece of their country just before they leave. As if Benidorm, Torrevieja or the entire Costa Blanca had not been enough, next January the first Wetherspoon in all of continental Europe will open at the Alicante–Elche Miguel Hernández airport. A “100 Montaditos British style”, but installed in the boarding area and designed, paradoxically, for those who are already queuing to return to the United Kingdom. The very British landing. According to The Guardianthe chain has confirmed that its premiere outside the United Kingdom and Ireland will be in Alicante, where it will open a newly built pub called Castell de Santa Bàrbera (when in Valencian it would be Castell de Santa Bàrbara), in “homage” to the fortress that crowns the city. This is a striking move for the company founded by Tim Martin more than four decades ago and which had never operated on continental European territory. For its part, as The Independent has detailedthe store will open every day from 6:00 a.m. to 9:00 p.m. and will be located in the departures area, aimed mainly at British people returning from vacation. The space will be about 93 square meters and will have an outdoor terrace. In addition, the menu will replicate 90% of the typical Wetherspoon pub menu: full English breakfasts, fish and chips, burgers and pizzas. Even so, it will also incorporate some typical Spanish dish such as garlic prawns or Spanish tortilla, an adaptation that the company has already confirmed. The choice is not accidental. British tourism in the province of Alicante is one of the most important in the region; Benidorm is well known for this. According to data collected by La Vanguardiaalmost 90% of English people choose the province as their favorite destination. Although a decade ago the owner publicly celebrated Brexitthe chain has recently experienced slowing growth in the UK: like-for-like sales of 3.7% in the first 14 weeks of the financial year, lower than in previous years. According to The Telegraphthe company is suffering from the increase in labor, energy and tax costs, which has led its president to explore new markets, and hence its strategy focused on airports: places where traffic is guaranteed and the clientele is usually predisposed to consume, even at times when most bars would not open. A British icon, almost invisible for Alicante. Despite the commotion that the news has generated in the province, the truth is that this first Wetherspoon on the European continent will be out of reach of the general public. It will be necessary to pass security control to access, which makes it a rarity: a British icon installed in Alicante, but almost invisible to the people of Alicante. Although Alicante will be the first, it will not be the only one. Tim Martin has reiterated in different British media that his intention is to open “several pubs abroad in the coming months and years, including some in airports”. The new location at Alicante airport will, therefore, be a test by fire. One last drink before heading home. Alicante can now boast of having the first Wetherspoon on the continent, although only travelers who fly will be able to enjoy it. For British tourists, it will be the last sip of home before returning; For the province, further proof of the weight that this market has in its economy. Time will tell if this little pub next to the departure gates is the start of a new European conquest or simply a last pint in the sun before heading home. Image | FreePik Xataka | Years ago Alicante thought it was a good idea to build an artificial island with a luxurious restaurant. It didn’t turn out as I expected

We believed that no open model could outperform GPT-5. A Chinese startup proves us wrong

by

A Chinese startup called Moonshot just launched Kimi K2 Thinkinga gigantic open model with a trillion parameters that has done something that seemed almost impossible: surpass the best proprietary models from companies like OpenAI, Google or Anthropic. If we thought that “Open Source” models could never compete with GPT-5, Gemini 2.5 Pro or Claude, we were wrong. what has happened. This “AI laboratory” had already announced Kimi K2 in July with that gigantic size of one trillion parameters, but now they have released the “Thinking” version with that same size (32 billion active parameters, Mixture of Experts architecture). According to those responsible, the model is capable of maintaining stable use of agentic tools over between 200 and 300 sequential calls. Or what is the same: it can chain long sequences of actions autonomously and apparently without error. The best of all is not that: it is that it surpasses GPT-5 or Claude Sonnet 4.5 in various tests and costs much less than those models. The benchmarks. Those responsible for Moonshot explained how Kimi K2 Thinking achieves the highest scores in Humanity’s Last Exam (general knowledge, 44.9%) and BrowserComp (agent browsers, 60.2%). He is almost at Claude’s level in the SWE software development test, and is also almost the best in another of those benchmarks, LiveCodeBench v6. It is true that in some tests still slightly behind of its “western” rivals, but the achievement is spectacular. More benchmarks. Those responsible for Artificial Analysis have shown their first conclusions after evaluating it with various tests. Thus, they highlight its behavior in agentic tasks that simulate that the model is acting as a customer service agent. In this test it obtained 93% of the maximum, surpassing all its competitors by far (GPT-5 Codex High obtained 87%, for example). They will do more tests, but for now the prospects are fantastic. And on top of that, cheap. On CNBC indicate that training the model cost $4.6 million, a ridiculous figure considering that training proprietary models like GPT-5 It cost about 500 million dollars according to estimates. Using the Kimi K2 Thinking API is also very affordable: $0.6 per million tokens in and $2.5 per million tokens out. GPT-5 Chat costs $1.25/10 respectively, while Claude Sonnet 4.5 costs $3/15 respectively. The details. The model makes use of an INT4 quantization to improve its efficiency without compromising the precision and quality of its responses. Its context window—the “size” of the data we can enter when making prompts—is 256k, a relatively modest figure for large models but still notable. And as a good open model, we can download it to use locally… if we have a real monster at our disposal. The model weighs 594 GB, and for example joining two Mac Studio M3 Ultra It is possible to make it work locally relatively smoothly at about 15 t/s. Alibaba is behindyes. Although the model is developed by an independent startup called Moonshot, this firm has been financially supported by Alibaba, which is becoming an absolute powerhouse in this field. Already not only conforms with developing its own models, which are outstanding (Qwen is the clear example), but is also financing the development of other models such as Kimi K2/Thinking. China and its love for open AI models. During the last few months we have seen how China dominated in the field of open AI models —not “Open Source”—. The Asian giant has adopted an overwhelming philosophy with increasingly better models but which until now seemed to be several steps behind the large proprietary models of OpenAI, Anthropic or Google. This is no longer the case. The race is lively. This achievement represents a new vote of confidence for the open models coming from Chinese companies. It is true that they are huge and that makes it very difficult to use them in practice by end users, but they present an interesting alternative for companies. Image | idnaklss with Midjourney In Xataka | There are many “internal” races within the greater AI race. And Alibaba is winning Open Source

does everything the other way around in Spain, is losing a fortune… and plans to open more stores

by

Costco has announced that it will continue opening warehouses in Spain despite accumulating 150 million euros in losses since its arrival in 2014, according to reports Digital Economy. The Spanish subsidiary recorded another 7.5 million losses in 2024, although its sales shot up to 607 million. The company already operates five centers (Seville, Getafe, Las Rozas, Sestao and Zaragoza) and is looking for new land. Its latest establishment in Zaragoza started with 15,000 members on the day of its inauguration. Why is it important. Costco represents the complete opposite of the model that dominates Spain: Mercadona triumphs with medium storesa reduced assortment and no membership. He doesn’t even have his own card. Costco is committed to large stores of more than 15,000 square meters, buying in bulk and charging a membership fee. It is the clash between two philosophies: the Spanish one that lives in your neighborhood and offers small, domestic formats, versus the American one of “pay 36 euros a year and get 24 rolls of toilet paper.” The strategy. Costco is playing the game Amazon played for two decades: lose money in a controlled way while it grows and build market share. Its 750,000 members (15% more than in 2023) and constant sales growth suggest that the model is finding its place. The company earned 11.5 million just from membership fees in 2024. Each new center comes with a gas station and Kirkland’s own brand products. Its average salary of 24,044 euros is above the sector. The contrast. Where Mercadona has immediate success, Costco has sustained losses. Where Mercadona optimizes margins (3.88% net profitin an upward trend), Costco optimizes volume and loyalty. Where Mercadona dominates with a 28% national share, Costco is building small niches. Yes, but. The bet has obvious risks. Costco needs critical mass for its model to work, and Spain is not the United States. Spanish purchasing habits favor proximity over volume, and competition in large stores – dominated by the French – is fierce. In fact, the hypermarket is going down in favor of the supermarket. We no longer make shopping a three-hour ritual on Saturday, but instead take advantage of empty spaces to make small purchases any day. And now what. Costco maintains that 2025 will bring more investment and land prospecting. The key will be if it manages to replicate in Spain what it achieved in other markets: convert initial losses into long-term leadership. It took Amazon twenty years to become profitable. Costco has been in our country for ten years and continues to invest. In Xataka | Spain has become a country addicted to something that a few years ago enjoyed little prestige: white label. Featured image | Marcus Reubenstein

Alibaba has one of the best open source AI models. Your next step: use it in robotics

by

Alibaba has taken another step in its commitment to artificial intelligence by creating an internal team dedicated to roboticswhich will operate from qwenits AI modeling division. The Chinese giant, owner of one of the best open source AI models, now wants the Qwen team to know how to apply its knowledge in robotics, a sector that is beginning to awaken interest, not only in industry, but also with the arrival of projects in the domestic sphere. Who leads the project. Justin Lin, technology manager at Qwen and expert in multimodal models (capable of processing text, sound and images), was the one who has confirmed the creation of this “small team for robotics and embodied AI” through their social networks. Lin has worked on the development of the Qwen models, which are currently among the most popular in open source globally. The vision behind the movement. According to Linmultimodal AI models are evolving into “fundamental agents” capable of performing complex long-term reasoning tasks thanks to reinforcement learning. “They should definitely make the leap from the virtual world to the physical world,” he said. explained the manager, making clear the intention to apply these technologies in tangible devices. Alibaba’s big bet. This announcement is part of Alibaba’s broader strategy in the sector. Last month, the company led a financing round of 140 million dollars at X Square Robot, a Chinese robotics startup. In addition, its CEO Eddie Wu esteem that global investment in AI will reach $4 trillion in the next five years, a figure that reflects the sector’s expectations. Global competition. Alibaba is not alone in this race. Nvidia and SoftBank are also making significant moves in smart robotics. SoftBank just announced the acquisition of ABB’s industrial robots business for $5.4 billion, while Nvidia CEO Jensen Huang has qualified the combination of AI and robotics as a “multi-billion dollar” long-term growth opportunity. China is also the world’s leading power in the robotics sector. And only in 2024, Chinese factories installed nearly 300,000 industrial robotsa figure higher than the rest of the world combined. The Qwen factor. The choice to place this team within Qwen makes all the sense in the world. Seven models of the Qwen series are currently listed in the top 10 Hugging Facewith the multimodal model Qwen3-Omni occupying first place. This strength in AI provides the company with a solid foundation to develop advanced robotic applications based on the journey they already have with Qwen. Cover image | zhang hui and Possessed Photography In Xataka | AI companies have just encountered an unexpected challenge: insurers have started to turn their backs on them

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.