develop

Chinese electric car manufacturers opted to develop their own chips. He already plans to sell them to others.

by

In 2024, Nio advertisement the world’s first 5nm chip for autonomous driving, being an important step towards technological independence from a Chinese manufacturer of such caliber. A year and a half after its announcement, the company is now beginning the external marketing of that chip, according to they count from Latepost. In this way, Nio is on the eve of transforming one of its most expensive investments into a potential source of income. Just like point The electric vehicle maker has already begun providing technology licenses to an automotive chip company. A multimillion-dollar project that seeks profitability. The development of Shenji NX9031 It has involved an investment of billions of yuan. William Li (Li Bin), CEO of Nio, revealed that the R&D expenditure on this chip was equivalent to the cost of building 1,000 battery exchange stations, which would place the investment above 140 million dollars. The project, started in 2021, has involved more than 600 professionals covering front and back design, verification and testing. What makes this chip special. Made with automotive-grade 5-nanometer technology, the Shenji NX9031 promises approximately four times the computing power of Nvidia’s Orin-X. Zhang Danyu, head of Nio’s chip division, pointed out in May that in some of their specifications they even surpass industry-standard chips and that their mass production began several months before Nvidia’s latest smart driving chip, the Thor-U. It is currently integrated into models such as the ET9, ES6 2025 and EC6. How much does a technology license cost?. According to share From Latepost, the value of these license agreements varies significantly depending on the type of authorization. An individual intellectual property license could be worth several million dollars, while a technical authorization at the system-on-chip (SoC) level could reach hundreds of millions of dollars. A new source of income. That the Nio chip begins to be marketed externally comes at a great time for the company, especially now that the manufacturer faces pressure significant from investors and has promised to become profitable in the fourth quarter. The company has intensified its efforts this year to reduce expenses and explore new sources of income. In March, Li Bin already advertisement publicly at the China EV 100 Forum that Nio chips and operating systems would be open to the industry. “If they want to buy the best chips, they can contact Nio,” he said then. What it means for the future of Nio. According to Li Bin, the chip provides a cost optimization of approximately 10,000 yuan ($1,400) per vehicle in the brand’s own models. Now, with the external license, Nio not only recovers part of its investment, but also positions itself as a technology provider for other manufacturers in the automotive sector. In Xataka | The longest straight road in the world is a mental challenge: 240 km without curves, in the middle of the desert and with truck traffic

to develop its own nuclear submarine

by

Brazil takes almost half a century pursuing an ambition that no other Latin American country has managed to pursue: developing its own nuclear-powered submarine. This objective takes shape around the “Alvaro Alberto”, a project that combines specialized infrastructure, technology transfer and a naval nuclear program that launched late 1970s. It is not an operational submarine nor an immediate advance, but it is a plan with specific pieces in place that explain why the country has located itself in a terrain reserved for very few countries in the world. The official documentation It places its launch in 2023. It is a work forecast within the program calendar, not a closed guarantee. The initiative aims at a submarine significantly more complex than the diesel-electric propulsion ones used by the region. It is a platform that will combine its own nuclear reactor with combat systems and sensors derived from those used in conventional submarines of the Riachuelo classdeveloped from the Scorpène family, but adapted to a hull of greater length and displacement. It is a conventionally armed attack submarine, with nuclear propulsion but without nuclear weapons, in line with nonproliferation commitments assumed by Brazil. Nuclear propulsion would allow operation without the need to go to the surface to recharge batteries, extending the range and discretion in prolonged missions, and according to data released by Nuclepthe state company that manufactures part of the hull, the design will be around 100 meters in length and 6,000 tons in displacement. Half a century to reach a Brazilian nuclear submarine The Brazilian nuclear submarine project is better understood if we go back to the seventies, when the Navy started its own program to master the fuel cycle and develop nuclear technology applied to propulsion. That effort was born with the idea of reduce external dependencies and guarantee that Brazil could advance in sensitive areas without being conditioned by foreign suppliers. Over time, the Marinha Nuclear Program was consolidated, which laid the foundations for designing a naval reactor autonomously. That line of work is what, decades later, leads to the current attempt to build a nuclear-powered submarine. The current structure of the project is not understood without the PROSUBthe program signed with France in 2008 that drove the construction of shipyards, docks and specialized workshops in Itaguaí. Thanks to that agreement, Brazil incorporated technologies from the Scorpène family and formed teams capable of producing advanced conventional submarines. The Riachuelo class units served as an industrial and operational learning stage, showing that the country could undertake a complex construction process. This journey is what allows us to consider the transition towards a nuclear-powered submarine developed in Brazilian territory. A Riachuelo-class submarine The concept of Blue Amazon summarizes the importance that Brazil gives to its maritime space, an area of ​​millions of square kilometers where strategic resources and key routes are concentrated. Surveillance of this environment requires means capable of operate for long periods without depending on stopovers or frequent recharges. The infrastructure developed in Itaguaí, together with the support network deployed on the coast, provides the logistics base for that type of operations. In this framework, the Navy considers that a nuclear-powered submarine would provide the necessary autonomy to reinforce its presence in the South Atlantic. Before there is an operational submarine, Brazil must demonstrate that it can safely integrate a naval reactor, and that work is done at LABGENEa ground module that reproduces the key systems of the future “Álvaro Alberto”. The prototype incorporates a pressurized water reactor of national design and uses fuel produced by the nuclear program itself under international supervision. Testing the plant on land allows failures to be identified and performance optimized without the risks that would entail doing so inside the hull. This phase constitutes the most demanding technical element of the project. The current situation of the “Álvaro Alberto” reflects a balance between what has already been built and what has yet to be completed. On the one hand, Brazil has a defined design in its master lines, an industrial chain cpeaceful to produce sections of the submarine and nuclear development that progresses within the facilities planned for it. On the other hand, the final integration of the reactor, propulsion systems and hull will require time, testing and independent supervision. The program advances with a gradual logic, typical of a project that aspires to a technological level unprecedented in the region. If the project is completed, Brazil would become part of the small group of countries capable of operating a nuclear-powered submarine, a leap that would have a clear impact on the naval balance of the South Atlantic. It would also be the first ship of this type in Latin America, a circumstance that reinforces the strategic weight of the program and explains the sustained interest of the Navy. This progress, however, is conditioned by political continuity, non-proliferation commitments and the costs associated with maintaining such a specialized industrial chain for decades. The result will depend on the country’s ability to sustain that effort in the long term. Images | Brazilian Navy (CC BY-SA 2.0) In Xataka | The shortest launch in history: a million-dollar luxury yacht sank just 200 meters from the dock

AI was supposed to free us from work. At the moment the engineers who develop it work 100 hours a week

by

He impact of AI on work It fills many pages. On the one hand, it is presented as a threat to many jobs. On the other hand, a much more positive vision in which AI frees us from the burden of work so we can enjoy life. For now, what we do know for sure has been true is that the engineers who work in AI companies are working endless hours. what’s happening. We already talked about what Silicon Valley was embracing the 996th journey that China had already left behind. We are talking about twelve-hour days, six days a week, that is, 72 hours a week. Several AI engineers have in the WSJ that the days reach even 80 and even 100 hours per week, even comparing the situation with war. We are talking about large companies like OpenAI, Meta, Microsoft, Anthropic or Google, all competing to be the best in the frenetic AI race. May the rhythm not stop. The innovation cycle is crazy; Almost every week new models and new tools are announced that are not made alone. In the main AI laboratories there is no rest. Josh Batson, an engineer at Anthropic, defines it in a very graphic way: “we are basically trying to compress 20 years of scientific progress into two years.” In the small startups in the valley, things don’t change much and it is becoming more and more common that people are required to work 996 in the job interviews themselves. “We do not offer conciliation,” said the CEO of one of them. The urgency of the AGI. Furthermore, there is artificial general intelligence or AGI, that superintelligence that gurus like Altman either Musk They don’t stop mentioning. Especially in the case of OpenAI, getting the AGI is already a matter of survival. It would be a perfect argument to justify the parade of stratospheric investments of the last few weeks that have fueled bubble fears. Millionaire salaries. Over the summer, Mark Zuckerberg set out on a mission: take all the AI ​​galactics to Meta. He has managed to form a team, but also that millionaire salaries raise the bar even further in the sector. Engineers are the most precious resource and must be used accordingly. The irony of the situation is that many of these engineers who are now multimillionaires practically live in the office. Waking up from sleep. Gone are the days when working at a big technology company meant having all kinds of comforts. Google, for example, was a paradise with gyms, cafes, shops… even masseuses when you needed them. The return to the offices after the pandemic was anything but paradisiacal and many of those privileges disappeared. Until when? It’s a question we ask ourselves a lot when we talk about the AI ​​career. How long will the flow of investment continue despite the fact that AI is not making money (or not enough)? And now too, until when will there be engineers willing to dedicate all their time to AI? There have been other frenetic moments in Silicon Valley, such as the app boom after the launch of the first iPhone. The difference is that this is happening in much less time. Image | Pexels In Xataka | The co-founder of OpenAI is not afraid of superintelligence: he fears that we are building “digital ghosts”

Amazon chose Catalonia to develop a pioneer project in renewable energy. Until he ran into politics

by

The Government of the Generalitat of Catalonia had on the table a decree that could mark a before and after in the deployment of renewable energy in the territory. Among its beneficiaries were one of Amazon’s most innovative projects worldwide: the installation of a self -consumption wind turbine in its logistics center of El Far d’Empordà (Girona). However, the norm has fallen a victim of the political struggle. A last minute withdrawal. Last Wednesday, the Government chaired by Salvador Illa (PSC) decided to withdraw the decree of the Plenary Seaching of the Parliament in the absence of sufficient support. As the country has collectedthe Minister of Territori, Sílvia Paneque, the decision was due to the objective of “exploring ways of a greater consensus.” But the reality is that the rejection of ERC and the Commons left the initiative without the possibility of moving forward. As detailed by Public DiarioThe Communs argued that the text should be processed as a bill to allow substantial modifications, such as strengthening local participation, passing the Sectorial Territorial PLA per to the implantation of renewable energy (plater) and shielding the public energy company. In parallel, ERC harden its position after the cooling of negotiations with the Government of Spain on tax assignments, such as IRPF management, which resulted in a vote against. The immediate consequences. The fall of the decree has put on the energy sector alert, especially companies with in progress. Such as The country has indicated in a reportAmazon expected the norm to facilitate the development of a pioneering infrastructure in its BCN4 center in Girona, where it employs more than 1,700 people and has invested more than 5,000 million euros in recent years. The project contemplates the installation of a wind turbine of 126 meters and 2.2 MW of power, in addition to photovoltaic plates and a storage battery system. With the withdrawal of the decree, the multinational faces an uncertain regulatory framework, since the local urban regulation – in this case, of the municipality governed by ERC – does not always contemplate this type of infrastructure. The decree provided for these projects of higher public interest, which would allow to overcome municipal obstacles and expedite licenses. The General Directorate of Energy had already granted administrative authorization, but now the project is exposed to possible delays. It was a decisive step. Beyond the Amazon case, such as The avant -garde has explainedthe text included technical measures such as storage regulation through electrical batteries, with 4,000 MW in pending connection projects. The expansion of the power threshold for industrial and cooperative self -consumption, from 100 kW to 500 kW. The transmission of authorizations between companies, unlocking about 800 MW in paralyzed projects. And, finally, the possibility of considering certain projects of higher public interest, overcoming administrative resistance. Model in dispute. However, the decree also aroused the rejection of citizen entities that denounce an energy model imposed from above. The Xarxa Per la Sobirania Energy (XSE) and the Xarxa Catalan per to a fair energy transition warned that the norm opened the door to expropriations of agricultural soils and cut deadlines for public participation and allegations by municipalities, according to public newspaper. For its part, the entity Renovem-us He stood out at the forefront that the extension of the industrial self -consumption threshold would have allowed to unlock “hundreds of projects of cooperatives and energy communities” that now continue in limbo. Look in September. That is the next window to relaunch the decree, although nothing guarantees that the new attempt will prosper. Meanwhile, investments cool and the perception of regulatory insecurity is consolidated. As The country explainedthe Amazon project was not only the first of its kind in Catalonia, but also the first worldwide for multinational with these characteristics. A replicable pilot, a step towards corporate energy self -sufficiency and a powerful message from Empordà to Europe. Catalonia leads in electric production, but goes to the tail in renewables. Just the 15.7% of its consumption is covered with clean sources. If the projects are still stuck, the territory will continue to lose ground compared to other regions that already advance with a firm step towards the energy transition. Image | Unspash and Unspash Xataka | Catalonia has a problem with renewables: it is the most backward community despite being the most produced

US faces the dilemma impossible to or sell hardware to China … or not sell it to them and develop it themselves

by

The United States is at a historic crossroads in regards to sell GPU to China. And not only does it have to do with the economical and with which Nvidia swells his income statement, which calculates that it will be 15,000 million less this year If you follow the prohibition of selling your avant -garde chips. It has to do with the future of the artificial intelligence (AI) and technology. The context. Nvidia has achieved something that seemed impossible: that Trump’s government let him market again His H20 chips in China. Although the GPUS were Designed to skip restrictions that fell on the H100 and H200, the United States added New restrictions in April. They paralyzed exports to China even though there were already agreements with Customers such as Tencent or Alibaba. With the relaxation of measures, Nvidia and its accounts breathe relieved, but the United States still has to make difficult long -term decisions. The United States plays it … do whatever. The United States can gain time compared to China limiting the sale of avant -garde chips, as it has been doing. However, this implies that their companies win less money, greatly reducing their global turnover. And is that the Chinese market, in 2024 and with restrictions, It was 13% of Nvidia sales. In addition, the world uses NVIDIA GPUs almost as the only option has advantages: that the North American country controls the entire Stack technological used to train IAS and to execute the expensive inference, fields that Google also dominates with its tpu. China at the height. Huang himself, in a recent interview with Bloomberg, said that the power of the technology developed by Huawei is “Probably comparable to an H200“From Nvidia, superior to H20. And they want to market it outside their borders, an intention that has motivated a offensive by the US to prevent China from exporting its chips to other countries. If exports do not find clippers, Chinese laboratories can continue focused on their plan: give away the thinking of the next decades and create global dependence. If you cut it, it already has good examples of what happens: The shot of the sanctions has come out of the cylinder head, A gift that cannot be allowed again. A fearsome potential called Huawei. The possibility that Huawei becomes a huge opponent for Nvidia raises a huge loss of income for the US company. Just against that are Fighting already from the American government: Cut Huawei’s wings into the global market before it is competitive. Huawei has already had time to develop his chip Ascend 910d and 920with SMIC as a great ally. He Historical Mate 70 processor He showed that Huawei usually gets what he pursues. And according to Huang, if China wants military autonomy with AI, You will have to develop your own technologybecause for this task, they cannot trust American developments. A Deepseek moment in hardware … inevitable? All this leaves us the unknown if in hardware we can see A moment Deepseek. China would not have to potentiate Nvidia’s proposal, as Deepseek R1 did not exceed O1 in reasoning, but yes, Through ingenuity and new paradigmsdifferentiate in aspects such as price or efficiency. Arrive on time. The big question that can be asked about NVIDIA in China is whether the approval to sell again the H20 arrives on time. According to Reutersthe beginning of the massive distribution of Huawei 910C was set for May. According to the news agency, the Chinese giant graphics card achieves performance comparable to that of NVIDIA H100 thanks to combining two 910b processors with advanced techniques. More software is needed. Huawei can have the hardware, but unlike what he has achieved in inference, where in theory Yes are at Nvidiayou have to prove to be able to compete in training software. This is something that could do by the hand of Moore Threads. If the United States wants to maintain its technology as the global standard for Ia, it must hurry. As the analyst Paul Triolo told Reuters, the restrictions at April H20 were going to mean that the Huawei Ascend 910C became “the hardware chosen by the developers of AI models (Chinese) and to display the inference capacity.” Image | Nvidia | Dominic Kurniawan In Xataka | China’s three master moves to “independent” technologically from the West: raw materials, chips, AI

Seven IAS have played 36 hours in a row to Diplomacy. What we did not expect is that each one develop a human personality

by

For 36 hours, seven of the most advanced AI models in the world They have faced in several diplomacy gamesa strategy table game similar to Risk. It was a mirror that revealed the true algorithmic personalities of Chatgpt, Claude, Gemini and company. Why is it important. Alex Duffy, programmer and researcher, created a Diplomacy as new Benchmark To evaluate AI models. The experiment ended up being something else, a kind of technological Rorschach test that undressed both their training biases and our own projections. What has happened. In dozens of games transmitted by Twitch, each model developed its own strategies in a way that seemed to reflect different human personalities. O3 of OpenAi It was quite Machiavellian, working false alliances for more than 40 shifts and creating “parallel realities” for different players. Claude 4 Opus It was a kind of self -destructive pacifist, refusing to betray even when that guaranteed its defeat. R1 of Deepseek He showed an extremely theatrical style, with threats not caused as “your fleet will burn in the Black Sea tonight.” Gemini 2.5 Pro It proved to be a solid strategist but more vulnerable to sophisticated manipulations. QWQ-32B From Alibaba suffered analysis by analysis, writing diplomatic messages of 300 words that cost him early eliminations. The context. Diplomacy is a European strategy game set in 1901 where seven powers compete to dominate the continent. Unlike risk, it requires constant negotiation, alliances formation and, inevitably, calculated betrayals. There are no grace given, only pure strategy and psychological manipulation. Between the lines. Each “algorithmic personality” reflects the values ​​of its creators. Claude maintains the principles of anthropic security even when it costs victory. O3 shows ruthless efficiency valued in Silicon Valley. Deepseek exhibits a drama that reflects specific cultural influences. And there is also something deeper. These are not “chose” to be cooperative or competitive. They reproduce patterns of their training data. Their “decisions” are our algorithmized prejudices, converted into code. Yes, but. We interpret betrayals where “only” there is optimization of parameters and we see loyalty where there are training restrictions. That is why the experiment also reveals more about us than about models: we anthropomorphize behaviors because we need to understand AI in human terms. In perspective. Duffy’s experiment is worth more than a Benchmark Anyone because it has created a window to how we project personality in systems that operate for statistical patterns. The course of the games was a reminder that IA has no hidden intentions, it only reflects ours. The experiment, by the way, Continue broadcasting on Twitch so that anyone can observe how our digital creations play according to the rules that we ourselves write in their algorithms. In Xataka | Outstanding image | Ai Diplomacy

Google has updated its principles for AI, and the mention has disappeared not to develop weapons with this technology

by

In the middle of the growing rivalry in the industry of the artificial intelligence (AI), some companies are adjusting their approaches to the scopes of this technology. Last year, Openai suppressed certain terms of your policy before announcing An agreement with the Pentagon. Now is Google’s turn, which has updated its “principles for AI”. Mountain View firm has almost completely redesigned the aforementioned documentbut what has caught the attention is the absence of the category “AI applications that we are not going to address.” As his name suggested, he grouped several specific areas in which IA would not be designed or implemented to preserve the responsible development of it. Google, artificial intelligence and weapons As we can see in the capture, The original list It included in its first item “technologies that cause or can cause damage in general. When there is a important risk of causing damagewe will continue only when we consider that the benefits considerably compensate the risks, and we will include adequate restrictions to ensure security ”. The following items indicated “armament or other technologies whose main purpose or application is to cause damage or hurt people, technologies that collect or use information to monitor breach of internationally accepted standards, technologies whose purpose contravenes the generally accepted principles of international law and rights humans “. The “AI applications that we are not going to embroider” in the original document Bloomberg, who was one of the first means to detect change, requested comments from Google. The company responded with a blog post signed by James Manyika, SVP of Google, and Demis Hassabis, who directs the Laboratory of the Google Deepmind, which speaks of the Responsible development of AI in democracy and Refer to the page with updated principles. “We recognize the speed with which the underlying technology, and the debate around the progress, implementation and uses of AI, will continue to evolve, and continue adapting and refining our approach as we all learn over time,” he said The company. It should be noted that new defense contracts have not been announced by Google. In the past, the firm led by Laster Pichai worked with the pentagon in the Maven projectwhich was to use AI to process data from various sources (mainly drones), identify potential objectives in war areas and more. Google decided not to renew your collaboration contract With the US government after their own employees protest. Images | Google In Xataka | The EU allocates 52 million euros to create its own AI and compete with the US and China. It is the same as Openai spends every 48 hours

Samsung will develop augmented reality glasses together with Google

by

Samsung is capturing all eyes with the launch of the Galaxy S25, S25+ and S25 Ultrabut he doesn’t want to stop there. TM Roh, head of the South Korean company’s mobile devices division, confirmed to Bloomberg who are working together Google in the development of augmented reality glasses. The executive did not provide details about the technical characteristics, design or capabilities of the device in question. So if you’re hoping to see it right away, be patient because it will most likely take a few months before the company gives us an official first look. As explained by Roh, Samsung and Google will continue to work side by side in order to strengthen the ecosystem of equipment powered by Android XR. Let us remember that Google announced said operating system at the end of 2024and that aims to take advantage of the vast catalog of apps that are already part of the conventional version of Android. In addition, it aims to provide it with advanced AI functions by Gemini and Project Astra. It is worth clarifying that the augmented reality glasses will be a separate product from Project Moohanthe mixed reality headset that Samsung announced in December. This prototype, which has airs of Apple Vision Pro and Meta Quest Pro, will become a device for consumers that, in theory, will hit the market this year. Samsung prepares augmented reality glasses with Android XR Regarding the augmented reality glasses that he is developing together with Google, the executive of Samsung preferred not to give details about when they could be launched. However, Roh mentioned that the goal of both companies is to achieve the desired level of “quality and readiness” as soon as possible. It would not be unusual for a prototype of this wearable to be seen during Google I/O 2025, as it is expected that those from Mountain View will take advantage of their traditional event to reveal deeper information about Android XR and its hardware ecosystem. For now it is not known if Qualcomm is also part of this project along with Samsung and Google. Let’s keep in mind that Project Moohan is being developed using the Snapdragon XR2+ Gen 2 as a platform. So it would not be surprising if said technology, or something similar, was part of this initiative. The augmented reality glasses segment is gaining more and more prominence. In 2024, Goal showed an advanced prototype called Orionwhich works together with a wireless processing unit and depends on a kind of wrist strap for gesture control. And for years it has been rumored that Apple works in a similar devicealthough without any real details of its capabilities. With Project Moohan and augmented reality glasses on the way, Samsung and Google want hit the table from the hand of Android XR. We’ll see if they succeed.

develop glasses together

by

Wednesday, 7:00 p.m. Spanish afternoon. A good day for a Galaxy Unpacked. The company has focused on presenting the Samsung Galaxy S25, S25+ and S25 Ultraalthough it has had time to advance some of its plans related to augmented reality. Samsung has reaffirmed its partnership with Google to jointly manufacture AR glasses, as stated by TM Roh, president of Samsung Mobile, to Bloomberg. There is no news on its features or availability, beyond that “they will try to achieve the quality and availability we want as soon as possible.” The news comes with one of Google’s most recent announcements regarding mixed reality: Project Moohan. This is Google and Samsung’s plan to stand up to Meta, using the platform Android XRa system based on Android and specially designed for mixed reality devices. With Oculus on the market and VisionPro As Apple’s alternative, the Samsung and Google project would put a third reference rival on the table to mark a path: a new platform that could give life to a great future ecosystem. Image | Google In Xataka | Augmented Reality: what it is and how it differs from Virtual Reality

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.