Houston, we have a problem with Outlook. Microsoft spends millions on AI, but Artemis II does not escape the failures of its email

On April 2, we experienced a historic event for humanity: the mission Artemis II It successfully took off towards the moon after more than 50 years without orbiting near the Earth’s satellite. Although the takeoff was a success, the path to get here was not without problems: it already had to delay the first date launch and also the second. Even on the official day there were problems. In the previous hours it was necessary check an anomaly in a temperature sensor of a battery abort system and also appeared another incident in the flight termination system (the safety mechanism that makes it possible to destroy the rocket if it deviates from its trajectory and becomes a threat). When the Orion spacecraft was flying almost 150,000 kilometers from Earth according to FortuneCommander Reid Wiseman encountered a mundane problem faced by any mortal with a computer and Microsoft email: an Outlook crash. The incident. The launch of Artemis II could be followed live and in that live, approximately 13 hours and 15 minutes after the broadcast began there is a fragment where the problem appears: “I see that I have two Microsoft Outlook accounts, and neither one works. If you could connect remotely and check Optimus and those two Outlook accounts, that would be great.” At first, Wiseman had issues related to the Optimus software, but then he pointed out a more trivial concern: There were two instances of Outlook running on his personal computing device. As a curiosity, the live stream to follow the takeoff still available on YouTube. Why it is important. The Artemis II mission is historic and the stream has left for posterity its first hours of flight and this anecdote that constitutes what is probably the first Microsoft technical support ticket generated from space. Beyond the joke, the episode shows that today’s space exploration and its cutting-edge technology coexist with commercial productivity software and its common failures. When an agency standardizes its entire infrastructure on a single technological ecosystem, the problems of that ecosystem also become problems of the mission. Tap to go to the post There is a support ticket from the Moon. As with any standard corporate ticket, the user first reported the incident, the technical team took over remotely, and finally closed the case. Houston accepted the request for remote access to the commander’s device, identified in records as PCD 1, and about an hour later, Outlook was back up and running. After 14 hours and 20 minutes of broadcast, someone from mission control communication said: “We managed to open Outlook. It will appear as “offline”, as expected”, as pick up Tom’s Hardware. Why they use Outlook in space. That there is Microsoft software on board is not something casual or improvised: Microsoft is a strategic partner of NASA that provides everything from productivity software to cloud data infrastructure and artificial intelligence (NASA Earth Copilot), hardware and mixed reality and Minburn Technology Group is your partner for software support and maintenance. In fact and according to NASAthe personal devices of the astronauts on the Orion spacecraft are Microsoft Surface Pro and the software they run is Commercial Off-The-Shelf, That is, standard commercial software for everyday tasks like talking to your family or managing your photos and videos. Another thing is the spacecraft and main flight systems: these are powered by specialized radiation-resistant hardware and specialized software with rigorous maintenance. The bathroom was also broken. The Outlook failure was not the only technical problem in the first hours of the flight, as can be seen in the broadcast. About two hours after launch, a malfunction light came on in the ship’s waste management system: the urine extractor fan had jammed. This component is responsible for sucking urine into a collector, avoiding the uncomfortable and unhygienic effects of microgravity. NASA confirmed shortly after the toilet problem had been solved. In Xataka | NASA had been refusing to allow its astronauts to carry iPhones for decades. For Artemis II you have made a historic decision In Xataka | The Artemis II astronauts will carry out experiments in what will be their own study models Cover | POT and Ed Hardie

Netflix spends 17 billion on producing content and YouTube does it for free. And that’s why YouTube is winning the game

Alphabet first revealed in its Q4 2025 earnings report that YouTube generated more than 60 billion dollars over the past year, adding advertising revenue and subscriptions. The figure is 33% higher than the 45,000 million that Netflix reached in the same period and places the video platform above all the entertainment giants except Disney, which had a turnover of 95.7 billion. The data confirms what many in the industry already sensed: YouTube is not simply another competitor in the online video market, but the main beneficiary of the transformation in audiovisual consumption habits. Paradigm shift. YouTube’s victory reflects a profound transformation in how we consume video. While subscription platforms opted for the Netflix model (closed catalogs of professional productions), YouTube added in July 2025 13.4% of total television viewing time in the United States. It expanded its lead over Disney (9.4%) to establish the largest difference recorded since these measurements began. Youtube on your TV. Time spent watching YouTube on television has grown 53% since February 2023. The traditional streaming market, meanwhile, is going through what is known as “subscription fatigue“: the average number of subscriptions per consumer in the European market has stagnated at 2.35 in both 2023 and 2024, after growing systematically for years. This saturation has caused structural changes: the number of original series released in the United States fell 11% in 2025third consecutive year of declines from the 2022 peak. The difference in the plan. Breaking down where the money comes from can point to the reasons for this triumph. Of the 60,000 million in YouTube revenue, we have: Advertising revenue in the last quarter of the year was 11.38 billion dollars, with a growth of 8.7% year-on-year 325 million paid subscriptions on all your consumer services, such as YouTube Music or YouTube Premium For its part, Netflix: It reported revenue of $12.05 billion in the fourth quarter of 2025, with a growth of 17.6% For the year as a whole, the platform reached $45.2 billion, with more than 325 million paid memberships The most notable difference lies in the business model. While YouTube maintains a hybrid model where advertising remains dominant, Netflix revealed its advertising figures for the first time: in 2025, its third year selling ads, advertising revenue exceeded $1.5 billion, multiplying by more than 2.5 compared to 2024. The company projects double that ad revenue in 2026. Why YouTube wins. YouTube’s competitive advantage lies in features that traditional platforms cannot replicate. On the one hand, the radical democratization of content creation: Netflix invests 17 billion dollars annually to produce, while on YouTube the creators assume the production costs. The base of 69 million creators generates a volume of content that is impossible to match: every minute 500 hours of video are uploaded to the platform The second differentiating factor is the algorithmic recommendation system. YouTube’s recommendation system uses large-scale language models that can handle massive amounts of data. This allows YouTube to do something that closed catalog platforms cannot: recommend videos based not only on general categories, but by fine-tuning suggestions based on specific interests. In 2025, YouTube’s recommendation system is the most sophisticated and user-focused. The third advantage is the absence of entry barriers for the public. While Netflix requires a mandatory subscription, YouTube offers free ad-supported access, with premium subscription as an option. This hybrid model maximizes potential reach: YouTube’s monthly active user base reached approximately 2.7 billion people in early 2025. This means that more than 25% of the world’s population uses YouTube in any given month. What it points to. YouTube’s triumph over Netflix in annual revenue represents more than a change in leadership: it signals a structural transformation in how audiovisual content is produced, distributed and monetized. The centralized studio model, a direct heir to the Hollywood system, is giving way to a decentralized ecosystem where millions of creators generate content for hyper-segmented audiences. And the implications for the industry are very profound. Header | Photo of NordWood Themes in Unsplash In Xataka | A YouTube video that lasts 140 years has gone viral. Nobody is clear why

Alibaba has just demonstrated that Openai spends 78 million to do the same as them for $ 500,000

There is a new star technique to train AI models super efficiently. It is at least what Alibaba seems to have demonstrated, that Friday presented His family of QWEN3-next models and did so presuming from spectacular efficiency that even Leave behind the one he achieved Deepseek R1. What happened. Alibaba Cloud, the Alibaba group’s cloud infrastructure division, presented a new generation of LLMS on Friday that described as “the future of efficient LLMs.” According to those responsible, these new models are 13 times smaller than the largest model that that company has launched, and that was presented just a week earlier. You can try QWen3-Next On the Alibaba website (Remember to choose it from the drop -down menu, in the upper left). QWen3-Next. This is what the models of this family are called, among which it stands out especially QWen3-Next-80b-A3Bwhich according to developers is up to 10 times faster than the QWEN3-32B model that was launched in April. The really remarkable thing is that it also manages to be much faster with a 90% reduction in training costs. $ 500,000 is nothing. According to AI Index Report From Stanford University, to train GPT-4 OpenAI invested $ 78 million in computation. Google was further spent on Gemini Ultra, and according to that study the figure amounted to 191 million dollars. It is estimated that QWEN3-Next has only cost $ 500,000 in that training phase. Better than its competitors. According to the benchmarks made By the artificial firm Analysis, QWen3-Next-80B-A3B has managed to overcome both the latest version of Deepseek R1 and Kimi-K2. Alibaba’s new reasoning model is not the best in global terms-GPT-5, Grok 4, Gemini 2.5 Pro Claude 4.1 Opus overcome it-but still achieves outstanding performance taking into account its training cost. How have you done it? Mixture of experts. These models make use of the Mixture of Expert architecture (MOE). With it, the model is “divided” into a kind of neuronal subnets that are the “experts” specialized in data subsets. Alibaba in this case increased the number of “experts”: while Depseek-V3 and Kimi-K2 make use of 256 and 384 experts, QWen3-Next-80b-A3B makes use of 512 experts, but only activates 10 at the same time. Hybrid attention. The key to that efficiency is in the so -called hybrid attention. Current models usually see their efficiency reduced if the input length is very long and have to “pay more attention” and that implies more computing. In Qwen3-Next-80b-A3B, a technique called “Gated Deltanet” is used that They developed and shared MIT and NVIDIA in March. GATED DELTANET. This technique improves the way in which the models pay attention when making certain adjustments to the input data. The technique determines what information retain and which can be discarded. That allows creating a precise and super -efficient cost mechanism. In fact, QWEN3-Next-80B-A3B is comparable to the most powerful Alibaba model, Qwern3-235B-A22B-Thinking-2507. Efficient and small models. The growing costs of training new models of AI begin to be worrisome, and that has made more and more efforts to create “small” language models that are cheaper to train, are more specialized and especially efficient. Last month Tencent presented models below 7,000 million parameters, and another startup called Z.AI published its GLM-4.5 Air model with only 12,000 million active parameters. Meanwhile, large models such as GPT-5 or Claude use many more parameters, which makes the necessary computation to use them much greater. In Xataka | If the question is which of the great technology is winning the AI ​​career, the answer is: None

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.