We have been relying on the Nutri-Score in stores for years. Science believes that its real impact is zero

He Nutriscore what we can see in some foods born with an ambitious promise: simplify the nutritional complexity of products into a code of easy to understand colors to know if a food is healthy or not. However, what on paper seemed like the definitive solution against obesity and poor diet is facing a much grayer scientific reality. His dark side. Although the idea seemed quite good, the reality is that new scientific reviews are setting off alarm bells. The conclusion being drawn is quite clear: the real impact on the shopping basket is minimal and the algorithm categorizes foods that are essential as something very bad. A good gap. One of the strongest arguments in favor of Nutri-Score comes from studies conducted in controlled environments, i.e. a laboratory. But what happens when we go down to the real, everyday world? This is what they wanted to analyze in a recent narrative reviewwhich evaluates consumer behavior in physical supermarkets and throws cold water on the system. And with this food color coding, the data shows that the improvement in the nutritional score of the purchase is only 2.5%. That is to say, it has hardly been noticed that a person begins to eat much more appropriate foods with this color code. Something that quite disagrees with the laboratory results that predicted that the effect was going to be much better. The real victim. The fact that some people’s shopping baskets have improved a little is the motivation that some producers of these foods have to change their ingredients to achieve a better Nutri-Score. as seen on Eroski. But this does not mean that citizens have changed the way they shop. The great blind spot. The fiercest criticism from the scientific field, highlighted by organizations such as the Puleva Nutrition Instituteis the omission of micronutrients. The current algorithm focuses almost exclusively on macronutrients, which are fat, sugars and proteins, but forgets other points that are fundamental. One of these points are vitamins and minerals, which are logically essential for the body, especially because some of them must be taken as they are not produced by the body. But polyphenols or bioactive compounds also stand out, which are essential antioxidants that can prevent chronic diseases. Unfair penalty. The system that is implemented right now also penalizes foods for their total fat content without differentiating whether they are healthy, something that has led to putting a bad score for olive oil. A paradoxical situation. The study from the University of Granada wanted to see the same thing about soluble cocoa to highlight these large discrepancies that force us to question Nutri-score. The result of the research team indicates that while pure cocoas with a higher bioactive profile can receive low grades such as C or D. But, on the other hand, others ultra-processed products with additives They achieve better scores, even A, simply by adjusting their sugar or fiber levels, without necessarily being healthier. Trying to correct it. The scientific community is no stranger to this problem and logically when something goes wrong you want to fix it to make it fit reality and that it truly fulfills the objective with which it was created. In fact, recent updates have already tried to correct the algorithm to better treat vegetable oils and nuts and penalize ultra-processed foods more strongly. However, the validations insist that, although there is an association between scores and macronutrientsthere remain huge gaps with comprehensive dietary guidelines. And we must keep in mind that the Nutri-Score measures “isolated nutrients” but not the overall quality of the food. ¿Where are we going? Science seems to indicate that the Nutri-Score is a useful but overly simplistic tool. By trying to condense health into a letter, nuances are lost that really make a difference in longevity and disease prevention. Although the algorithm is being refined to better align with European recommendations, the risk of the consumer blindly trusting an “A” for a processed product versus a “C” for a natural food remains present. Images | Franki Chamaki In Xataka | Ozempic’s “great rebound”, in figures: science reveals that the weight returns four times faster than with a diet

Relying on US AI is a strategic danger

When DeepSeek R1 It was presented a year ago now, caused a real earthquake in the technological world. What was surprising was not its capabilities, but that China had managed to reach that level despite the blockades and setbacks of the United States. DeepSeek was proof that AI can be done without the United States and now it is Europe that needs to replicate this success. Tensions and dependence. Relations between the United States and Europe they are going through their worst moment. Trump’s obsession with take control of Greenland and the response of several European countries that They have sent their troops to the region have caused an unprecedented clash. Amid the threats of invasion, the deployment of troops and tariffs, there is also the issue of technological war, a war in which Europe is in a position of strong disengagement from the US. The US executed and Europe regulated. Far behind. If China lags behind the US in AI, Europe is light years away. While American companies were developing the models and infrastructure to train their AI models, in Europe regulation was reinforced with the AI Act. The European Union itself understood that this approach was leaving them behind in the AI ​​race and recently They greatly simplified the rules. It was late, the technological gap was already enormous. Dependence. The United States not only controls the language models, it also controls the chips to train them, the data centers and, above all, the investment to get all this going. Miguel De Bruycker, head of the Brussels Cybersecurity Center, is very forceful: “Europe has lost the internet (…) If I want my information to be 100% in the EU… keep dreaming,” he told the Financial Times. In the current context, this dependency puts Europe in a very vulnerable position and becomes a major strategic risk. The US could use its dominance as a pressure point in negotiations or, in the worst case, restrict access to its services. A sovereign AI. They count in Wired that the concern to create a European AI is growing and there are already several projects underway to achieve it. The best known model is the French one Mistralbut there are others like Apertus in Switzerland or ALIA in Spain. In Germany they are developing SOOFIa project that aims to launch an open source language model with 100 billion parameters designed specifically to reduce European dependence on the US. Chinese inspiration. The US seemed unattainable, but DeepSeek showed that it was possible to achieve competitive results without having the best GPUs or the largest data centers. The fact of bet on open source It also gives an advantage since it allows creating a larger user base in less time, in addition to more actors can participate in the developments. There is also talk that Europe could encourage its companies to use its own AI, a strategy similar to that followed by China with the use of national chips. Image | Karola G, Pexels. Engin Akyurt, Unsplash In Xataka | The ASML-Mistral alliance reveals the European plan B: if we cannot manufacture chips, at least we will control how they are manufactured

Chinese startups have been relying on NVIDIA chips to train their models for years. That is already changing

The name of the Chinese startup Zhipu AI (Z.ai) may not sound familiar to you, but perhaps GLM, its AI model, does a little more than its latest version, GLM-4.7already competes with Claude Sonnet 4.5 or GPT-5.1. The real surprise of this “Chinese AI tiger” is the launch of GLM-Image…and not so much for what he does, but for how he has managed to do it. what has happened. GLM-Image is a multimodal generative AI model that focuses on image generation. The idea, of course, is to compete with options like Nano Bananafrom Google. That’s interesting, but even more striking is the fact that the model has not been trained with conventional chips. Trained with Chinese chips. According to those responsible for Z.ai, this model is the first developed in China that has been fully trained with “local” chips. Specifically, it has been trained with Huawei’s Ascend chips thanks to the use of servers Huawei Ascend Atlas 800T A2 and a framework called MindSpore. Thus, traditional NVIDIA AI chips, which are usually the usual choice for AI model developers in Chinese startups, have not been used. Turning point? This milestone demonstrates the real feasibility of training high-performance generative AI models on a platform developed entirely in China. We are not dealing with something minor: it is validation that it is possible to continue innovating in this area despite the restrictions imposed by the US. In fact, Zhipu AI — included last year on the US blacklist — has intensified its collaboration with other local manufacturers, such as the promising firm Cambricon that has risen from the ashes thanks to tariffs. Threat to NVIDIA. The news comes at a unique time, because NVIDIA has not stopped pressuring the US government to once again allow it to sell its advanced AI chips to Chinese companies. He has obtained that permission—which It won’t be free—, but now the one that might not be interested is China, which he hasn’t said anything at all. That chips from companies like Huawei are a valid alternative for training quality AI models can change many things in this area. Zhipu goes like a shot. The Chinese startup has also just gone public, and since it has done so its shares they have shot up more than 80%. Investors see the company no longer as a rival to Google or OpenAI, but as a banner. One that shows that it is possible to compete without depending on the US and its companies. Huawei, great beneficiary. If the trend continues, Huawei can become the Chinese NVIDIA, and the company prepares an increase in production of its AI chips. It is not the only one: Cambricon plans triple your production by 2026, which seems to make it clear that the Chinese industrial machinery is moving quickly to neutralize the impact of US vetoes. Challenges…Despite everything, Zhipu already has warned that the price war in the AI ​​sector will become international. If Chinese companies end up controlling the entire chain (or rather, their chain), they could offer AI services at much lower costs than their Western competitors, who must pay NVIDIA’s margins and Big Tech’s cloud infrastructure. …and unknowns. This technological achievement raises other questions. One of the most important is how powerful and capable Huawei chips are compared to NVIDIA’s in these processes: is training much slower? Is it more expensive in time and resources? The efficiency of the MindSpore framework compared to Pytorch or TensorFlow is another of the key components of these developments. In Xataka | Faced with the US strategy, China has a plan to revive its technology industry: that AI belongs to everyone

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.