Mobile Tech & Gadgets Archives - Page 1148 of 1159

give up fast money to win the war

January 24, 2025 by usatoday24

Openai has chosen a conservative path to Operator: limit it to users of Chatgpt Prothose who pay 200 dollars a month. Only they can access this tool that automates tasks on the web. It is what OpenAi describes as a “prior research view”, and that lets us glimpse more an experiment than a finished product. Why is it important. The IA giant continues to prioritize product refinement over immediate income. 200 dollars a month per user capable of using operator are many dollars … but that is an important limit to its adoption in this phase. And it is deliberate. Between the lines. Openai has signed alliances with companies such as Dordash, Instacart, Priceline, Openable or Uber. That could suggest trade agreements to monetize operator on the side of the companies, but the OpenAi Coo, Brad Lightcap, has been blunt in the Wall Street Journal: There is no financial relationship in these collaborations. They are only optimizing the user experience to become more compatible with Operator. Therefore, the income that Operator now will come exclusively from Pro subscriptions. But that can change in the future: Openai has announced expansion plans to Users Plus, Team and Enterprise. He has also promised to release the API of Your Cua model (Computer-Useing agent) For developers. At the moment, there are no dates for either of those two promises. In perspective. Openai operates in a market with a giant potential. Analysts expect that The generative AI will move more than one billion dollars a year in less than a decade. However, Openai is betting on perfecting operator with a controlled implementation rather than maximizing short -term income. The bet. Initially limiting operator to its most committed and enthusiastic users (those who pay $ 200 per month), OpenAi seeks feedback quality. The one that is most capable of giving this type of user. In addition, errors are less expensive in this phase, and more valuable learning. This strategy suggests something: that the monetization plans with greater reach will come when the product is more mature. Deepen. Openai’s caution is striking because it contrasts with its aggressiveness on other fronts. For example, a few days ago we knew that it was also part of the Stargate projectvalued in half a billion dollars. Operator seems like an experimental strategic project, but not so much an immediate source of income. In Xataka | OpenAi prepares an AI with a doctoral level. He is so promising that he will first show it to the US government Outstanding image | OpenAI, Xataka with Mockuuuups Studio

If the question is who wants to build the world’s largest data center, the answer is surprising: India

January 24, 2025 by usatoday24

The data centers is the new gold fever of technology. That Mukesh Ambani, the founding billionaire of the company Reliance Group. His intention is to go beyond what the great American technological ones promise and build in India the largest data center worldwide. 3 GW. That is the energy capacity that Ambani projects for a new data center that wants to build in Jamnagar. This is stated by sources close to the project In Bloomberg. Akash Abani, CEO and son of the founder, said in a recent statement that they were building a data center in less than two years “as we have always done in – in 24 months.” An irreverent figure. At present, the largest mute data centers are below 1 GW of energy capacity, as indicated in Bloomberg with data from the DC Byte consultant. In visual capitalist They showed a year ago the largest data centers in the world. To be able to compare with the Indian project, a fact: all AWS data centers in the Northern Virginia region, in the US, about 300, add a total of 2.5 GW, when one in India could have more energy capacity that all those United centers. Amazing. The most capacity data center in the world is currently from Microsoft in Boydton, Virginia, with almost 600 MW of energy capacity. The more energy, the better. Data centers are usually measured by the energy they can provide to their servers, their cooling systems and the rest of the equipment of these facilities. The larger the figure, the more computing capacity, and here the AI models have created a career to build more and larger data centers. Ambani for all. The tycoon is known for his aggressive tactics to have his rivals with price wars that few can sustain. It happened in the mobile segment and now wants to take the strategy to AI. Its objective is to reduce inference costs to a minimum to become an absolute reference in this market. This was stated In a statement highlighting that “our goal is to create the cheapest artificial intelligence in the world, right here, in India.” Astronomical cost. However, it is not clear how Reliance will finance this potential operation. According to Bloomberg estimates, a data center could have a cost of between 20 and 30,000 million dollars. Reliance Industry, the main division of the group, has about 26,000 million dollars. Jamnagar. The Indian city has a population of 650,000 people. There is the refining and petrochemical complex of Reliance, the largest in the world, but also Reliance wants to build there a renewable energy complex huge These energy sources would be especially used in that data center. Stargate in the spotlight. These days there is much of what to talk about in data centers, especially after the announcement of Openai and Softbank of PROJECT STARGATE to create a series of data centers in the US with a gigantic investment of 500,000 million dollars. There are no specific details of the energy capacity that is discussed in that case, but with that investment that data would be equally colossal. Image | Microsoft In Xataka | Stargate impossible accounts: the challenge of finding 500,000 million dollars for the largest IA project

Nepal has decided to get serious about the touristification of the Himalayas

January 24, 2025 by usatoday24

Everest is an extreme place, dangerousdifficult to access, roof of the world and one of its most unknown points. But of course It is not an inhospitable place.. Its magnetism is so powerful that every year hundreds and hundreds of mountaineers, plus their Sherpas, put their backpacks on their backs to climb its 8,849 meters. It is estimated that in 2018, before the pandemic, they reached the summit more than 800 climbers. And the forecasts for this year’s spring season are generous, with about 700 summits from Nepal and Tibet. With these data, the Nepalese authorities have made a decision: increase the rates They charge almost 40%. A demanding goal… and expensive. Summiting Everest is not easy. Not cheap either. Especially from September onwards, when Nepal expects skyrocket rates which charges climbers in exchange for the permits necessary to climb the highest mountain on the planet. Prices had not been touched for almost a decade, so in Kathmandu they have decided to go all out, increasing their rates about 36%. In hard cash that translates into thousands of extra dollars for those hoping to summit Everest. How much will they cost now? Starting in September, those who want a Nepalese permit to climb the mountain will have to pay $15,0004,000 more than now. This is during the most popular climbing season, April to May, and on the standard route of the southeast ridge, the South Col. From September to November and from December to February, less crowded months, rates will also increase by 36% to reach, respectively, 7,500 and 3,750 dollars. An expected rise. The update of rates (the first “in a long time”, defend from the Nepal Tourism Department), will make the climb to Everest even more expensive; but it will probably have caught few climbers by surprise. “We expected this increase in rates,” recognizes Guardian Lukas Furtenbach, from the Austrian expedition company Furtenbach Adventures. In his opinion, this is an “understandable” decision by the Nepalese Government and he trusts that the extra dollars that enter its coffers starting in September will be used on the mountain itself, protecting its ecosystem and reinforcing security. Key to the region. If we talk about mountains, Nepal is not just any country: it is home to eight of the 14 famous ‘peaks 8,000’ on the planet, making it a coveted destination for mountaineers. And that in turn generates an interesting source of income for its population. The BBC estimates that mountaineering and hiking more than 4% of the country’s economy and throughout recent years there have been several experts who have tried to quantify exactly how much mountain tourism contributes to the region. “Mountain tourism is an important source of income for the inhabitants of the Himalayas, since agriculture is limited in these places due to harsh climatic conditions and poor crops,” points out Sayas D. Joshi at Nepal Economic Forum. According to the data it managed in 2022, the activity leaves the villagers between 150 and 250 million Nepalese rupees each year in salaries and services, in addition to the activity it generates for agencies and operators. A valuable source of income. “According to data from the Nepal Mountaineering Association, seasonal employment in all categories of supporting labor ranges between 40,000 and 50,000 people. Climbing is a crucial business that brings much-needed income to the rural economy,” duck. “The 2021 spring season saw the highest number of Everest expedition teams since COVID-19, with more than 400 climbing permits issued. Mountaineering licenses have generated about 470 million Nepalese rupees in revenue for the Government”. For reference, that sum is equivalent to $3.4 million. The other bill. The influx of mountaineers aiming for the summit does not only translate into rupees, employment and fees. For Everest it means something equally or even more palpable: garbage and excrement, tons of waste which have even led the authorities to demand that mountaineers collect their own feces in biodegradable bags. The saturation of the mountain, with groups in which hundreds of people gather, also generates debate for years. Hundreds of permits. Reuters points out that each year some 300 permits to climb Everest, but the truth is that the total influx of mountaineers comfortably exceeds that figure. In 2019 it is estimated that they reached the summit 807 mountaineers and there are estimates that suggest that this spring season could be close to a similar number of summits, including those on both the Tibet and Nepal sides. Proof that the limits and sustainability of the mountain generate debate in Nepal is that not even a year ago, in April 2024, its Supreme Court demanded that the Government set limits on the issuance of permits to ascend Everest. The BBC remembers However, the order did not detail a maximum number. When money doesn’t matter. Nepalese fees are only a part (and not a very high one) of what it costs a mountaineer to climb Everest if all his expenses are taken into account. There are estimates that speak of $45,000others place the normal fork between 40,000 and 60,000they bring it closer to 70,000or even extend the total to the 200,000dollars. The truth is that money is not a problem for everyone. Financial Times spoke recently of a special seven-day package to ascend Everest. Its cost: 150,000. Why so few days and such a large sum? Partly due to a special pre-acclimatization treatment. Images | Mário Simoes (Flickr) and Munu Nepal (Flickr) In Xataka | Everest is growing faster than we thought because one river stole water from another 90,000 years ago

what it is, how it works and is used, how to hire it and how much it costs

January 24, 2025 by usatoday24

Let’s tell you what is the OpenAI Operatora new launch from the company responsible for ChatGPT with which they are betting on AI agents. But being a quite different product and with other purposes, it is perhaps a little more complicated to understand and more confusing. Therefore, we are going to try to explain everything to you in a simple way. We will start by explaining what AI agents are and what Operator is. Then we will tell you how it works, and we will finish by talking to you about how to access it and the price it costs. What are AI agents This is what Operator looks like in action Operator is an Artificial Intelligence agent created by OpenAI. AI agents are automations that perform actions for you on the computersuch as in the browser. The idea is to streamline processes in which you have to perform several different tasks, and allow the agent to do it for you. It’s going to be something like a virtual secretary. For example, think that you want to go on a trip to a city, and that you have two or three destinations in mind. Currently, what you do is look for hotel options in the cities, look at the weather information, and even see what events that interest you are in these cities to make the decision, and then you book and pay. The agents’ idea is that do all these processes for you. For example, you just write to the agent on your computer (or tell him with your voice) that you want to leave on certain days and that you are thinking about these two or three destinations. Then, the computer is in charge of checking the weather, seeing interesting events, offering you hotel options and collecting and giving you all the options so that you only have to choose one and confirm the payment. However, this is the ultimate goal of the operators. Currently, the options that are being launched are much simpler, although they continue with the objective of being able to do as many things for you. In a few years, this could be a reality. What is Operator? Operator is an AI agent created by OpenAI. It is currently in the development phase, although there are already some users who can start using it. This is an agent that can navigate the web to perform tasks for us. Operator will have the ability to use its own browser to browse the web, and interact with a page writing and clicking and scrolling to find information and perform tasks for us. In this way, what OpenAI wants is for you to be able to tell Operator something like “Reserve me a table for two at Beretta for 7 p.m.”, and for this agent to take care of doing it for you Opening a reservation website, search for the venue, select the date and time, and make the reservation for yourself. If a problem arises such as there is no free place where you want to go, or if some sensitive information about you is required, then it will ask you to intervene. Come on, the idea is not to do things crazy, but to If he can’t do something, he’ll ask you. so you can decide what to do. Following the reservation example, this allows you to choose another date or time, or for you to write your payment method so that the operator does not use your card on its own. But it can also ask you for confirmation of actions or helpthe first in payments and when making final decisions, and the second in other obstacles such as a CAPTCHA or having to write username and password. You will also be able to create custom tasks, configure specific instructions and preferences for automate tasks you do frequently. For example, if every month you make a specific purchase in an online supermarket or do some periodic online action, you can configure this so that Operator does it on its own from time to time. At the moment Operator has only been announced and released to a small group of people. It is in the investigation phasewhich means that they will use the data and feedback from the first users to start fine-tuning it and improving its operation. This also means that at the moment it is not very reliable, and that it can make mistakes. How Operator works Operator uses an own artificial intelligence model. Is called Computer-Using Agent (CUA)and is based on GPT-40. This model is able to “see” what is happening in the browser by taking screenshots, and then allows you to use typical controls such as the cursor or mouse to navigate and take actions on the web page you go to. This model is capable of analyzing and understanding each pixel of the screenshots taken on a website, and thus knowing identify objects on screentext and images. Once it analyzes the content of the website, it does a kind of internal “reasoning” and follows a line of “thought” to carry out the actions. Of course, it is also adapting taking into account past actions. To carry out the tasks, Operator uses an internal browserso you don’t have to install anything. This browser works remotely from OpenAI’s servers, and both your data, what you ask of it, and the captures of the websites end up stored on the company’s servers. This means that, as happens in ChatGPT and almost all AI models that do not work locally on a computer, the company owns the data. OpenAI will then use this data to detect fraudulent activity and to train and improve the model. In any case, you can deactivate this collection from the settings, or decide how long the data is stored. How much does Operator cost? Operator will be included in ChatGPT Pro subscriptionwhich is the one that has a price of 200 dollars a month. Therefore, you cannot hire … Read more

Perplexity launches an assistant for Android. An all in one that marks a before and after

January 24, 2025 by usatoday24

The AI Assistant Perplexity comes to Android with an enormous promise: to become the first truly useful assistant on your mobile. Or at least that it is capable of going far beyond what we knew. Why is it important. This is not just another AI app. Perplexity integrates real-time web search, multitasking control and advanced automation in an interface that seeks to understand the context of your needs. The context. Current mobile assistants are limited and frustrating when compared to the capabilities of an AI chatbot. Perplexity breaks these barriers by being able to maintain coherent conversations while doing tasks in different apps. That is, it combines the capabilities of an AI chatbot with the classic assistant. It integrates with apps like Uber, Spotify and YouTube, as well as messaging and watch apps. More will come. It is multimodal, so it supports voice, text and camera… …so it can even “see” what you have on the screen. It is available in fifteen languages. Maintains context between different actions. Access information in real time thanks to reading websites. Between the lines. The most important thing about Perplexity Assistant for Android is its ability to maintain context. For example, if you ask it about Italian restaurants near you, it will be able to automatically reserve the place that best suits your previous preferences and your usual dining time. Context. More case studies: You can identify a product, even if it is newly launched, using the camera. The Verge mentions his success in spotting a promotional toy released just two days ago. You can compose and send emails maintaining our usual writing style. You can manage our calendar by creating contextualized events. For example, “remind me to take my heart pill two hours before the Valencia game.” Something more concrete: you can ask him ‘play the song that plays at the end of the movie’Interstellar‘”. The big question. Is Perplexity going to be able to unseat the Google Assistant, or at least be more capable of competing with ChatGPT by having your own scope of action? Although it is not easy at all, it is in a good position: it is already valued at 9 billion dollars, so it has the financial and technological muscle to try. Go deeper. The power of this assistant lies, above all, in its proprietary search engine. It is what allows you to give updated and accurate answers. And what makes it overcome the limitations of chatbots that are based on pre-trained systems. The assistant for Android can be activated from the Perplexity appreplacing the native Google one. As for whether it will come to iOS, it seems complicated: the company remains open to this possibility as long as Apple “grants them the appropriate permissions.” Featured image | Perplexity In Xataka | If you have Movistar, you have Perplexity Pro for free: so you can use it and get the most out of it

a bug that prevents the car from moving

January 24, 2025 by usatoday24

He Renault 5 It is a car that is being liked. We verified this by driving it during our first impressions where we say that it is, without a doubt, one of the cars with which I have felt the most observed. And it is easy to verify with sales. The deployment in the French car market has been spectacular. During the month of December he fought face to face with the almighty Tesla Model Ywho managed to take away the honor of best-selling electric car in France for just over 100 units (4,807 units of the Tesla Model Y compared to 4,681 units of the Renault 5). The emergence has been such that with just over two months on the market it has reached close to 10,000 units and has managed to become the seventh best-selling electric car in France, according to data collected by Clean Technica. But some Renault 5 buyers have reported a worrying bug. One that leaves the car unusable because it does not allow them to move from position N. A problem that Renault is already solving The story comes to us through our French colleagues from Jeuxvideowho echoed the case via an official Renault forum in France. In it, the owner of an electric Renault 5 says that, just a week after purchasing the car, his small electric vehicle does not change gears. It is “frozen in N (Neutral) mode,” he explains in the post that you can read at this link. In it it states that the car cannot jump to D/B mode (to resume driving) or R (to reverse). Furthermore, he assures that other drivers have had the same problem, as explained to him by the person who towed the car to the workshop. Of a case collected three weeks ago, “he told me that another owner (of whom I attach a photo) had exactly the same problem“he emphasizes in the publication. In that same publication, two other drivers claim to have had the same problem. One of them assures that it was solved after half an hour and that the vehicle has to be updated. The problem is known to Renault. In fact, the company itself has offered an official response which explains that they have provided a patch to fix this bug that leaves the car unusable. If this happens, they recommend performing a complete reset of the vehicle, which can take between 20 and 30 minutes. Simply using the car again should have solved the problem. In Xataka We have contacted Renault, who have guaranteed us that they have no reports that this could have happened in any of the Renault 5s that already operate in our country. Despite everything, if we find ourselves in this situation and are not at home, Renault also explains that we can wait inside the vehicle for the reset to be completed but that we must comply with the following instructions: Make sure the vehicle is not charging. Do not use interior lights (dome lights, etc.). Do not use screen mirroring, radio or applications. Lock the vehicle using the key and/or the button on the instrument panel. Wait for the SOS LED to turn off before starting the car again Photo | renault In Xataka | Renault surprises with a bombshell: we will have an electric Renault 5 Turbo, rear-wheel drive and 500 HP

A PS5 for 400 euros or a macbook at an outlet price, these are some of the best chollos of the week

January 24, 2025 by usatoday24

Many stores have already started the second sales, so getting chollos is now much easier. If you are thinking about renewing some of your technological devices; As every Friday, the time has come to collect some of the best offers we have found this week … it’s time for Hunt bargains! Apple laptop computer MacBook Air 2023 Chip M2 by 999 euros: With 16 GB of RAM and chip m2. Tablet Lenovo Tab Plus by 209 euros: With 2K resolution and 26 W speakers. Bluetooth headphones Xiaomi Redmi Buds 4 Lite by 14.99 euros: With autonomy of 20 hours and noise cancellation. Smart TV TCL P75 Series 50p755 by 299 euros: 50 inches and with Google TV. Console PS5 Slim by 404.10 euros: reconditioned and with a dualsense command. APPLE MACBOOK AIR 2023 CHIP M2 portable computer If with the new year you have set out to be more productive at work or studiesthis MacBook Air It will be a good partner in your day to day. It is difficult to find discounts on Apple products but, now, this laptop from Cupertino is available by 999 euros In Amazon. It is the penultimate generation of the Air Macbook launched by Apple and comes with the Chip m2. It has one 16 GB RAM and 256 GB of SSD storage. Its screen is 13.6 inches and the reduced model is the star white. Apple Macbook Air (2023) 13.6 “Retina, Apple chip, 8 -core CPU, 8 -core GPU, 16 GB RAM, 256GB of SSD, White Estrella * Some price may have changed from the last review Tablet Lenovo Tab plus If the Magi did not bring you the tablet you had asked, now you can buy a good, beautiful and cheap. In pccomponentes, you can now get the Lenovo Tab Plus reduced, by 209 euros. This Tablet of the Lenovo firm presents an 11.5 fleas screen and 2K resolution. Although there is something for what it stands out is to have Eight JBL speakers (offered by 26 w). The processor that mounts is Mediatek Helio G99 and comes with 128 GB of storage, through MicroSD cardup to 1 TB. * Some price may have changed from the last review Bluetooth headphones Xiaomi Redmi Buds 4 Lite The wireless headphones sector is very diversified and there are many options that we find today in the market. If you want good, beautiful and cheap, these Xiaomi Redmi Buds 4 Lite They are one of the options to consider. Now, they are available for 14.99 euros In the official store. These Wireless headphones They connect through Bluetooth 5.3 and are compatible with Google Fast Pair. As for their battery, they offer a Autonomy of up to 20 hours With the load case. To make clear calls, they integrate noise cancellation with AI. Xiaomi Redmi Buds 4 Lite – Wireless headphones * Some price may have changed from the last review SMART TV TCL P75 Series 50p755 If you are looking to renew the old TV of your home, in PcComponentes, we have found this TCL P75 50P755 series of 50 inches at an irresistible price. Before it cost 449 euros but, now, it is available for 299 euros. Mount a LED panel with backlight Direct LED and resolution 4k ultra HD and its speakers offer a power of 20 W. comes with Google TV And it is compatible with Apple AirPlay 2. As far as connectivity is concerned, integrates WiFi 4, Bluetooth 5 and has Ethernet port, HDMI input 2.1, a USB 3.2 port and another USB 2.0. TCL P75 Series 50p755 TV 127 cm (50 “) 4K Ultra HD Smart TV Wi -Fi Titane * Some price may have changed from the last review PS5 Slim console Although Sony recently launched the PlayStation 5 Prothe truth is that it is expensive. If you are thinking of buying a console, but your budget is smaller, this PS5 Slim Digital is now at an unbeatable price in the Mediamarkt outlet on eBay. You can take it for 404.10 euros. This PS5 Slim It comes without the record reader, which makes its size much more compact, than the normal version. Of course, you should know that it is a reconditioned product And that comes, in addition, with a dualsense command. * Some price may have changed from the last review Some of the links of this article are affiliated and can report a benefit to Xataka. In case of non -availability, offers may vary. Images | Xiaomi, Sony, Tcl, Lenovo and Apple In Xataka | The best Xiaomi mobile price: purchase and comparative guide In Xataka | Fast loaders for your mobile phone or tablet: better models to buy for your power and safety

lock all chickens

January 24, 2025 by usatoday24

We are surrounded. Or, rather, the aviar flu has us surrounded. With positive in France and Portugal, the Ministry of Agriculture, Fisheries and Food knows that it is a matter of time that the positives explode in Spain. The mission is now to delay everything that can. What does the ministry say? They don’t walk around: “In recent months, outbreaks of high pathogenicity avian influence in poultry, captive and wild (…) in the central and northern area of Europe.” And the situation has gone worse. “In particular, in recent weeks there has been an increase in the number of cases detected (…) indicating a tendency to worsen up of the situation with the consequent increase in risk for Spain,” they continue. All that, “make it advisable to increase the level of risk nationwide and apply risk mitigation measures.” What implies all this. In terms of measures, the ministry has launched a whole series of measures: things such as prohibiting the “breeding of ducks and geese with other species of poultry”, the “breeding of poultry out of the outdoor” or even, even, even “Give water to corral birds from water tanks to which wild birds can access.” However, the standard opens the door to the use of birds or other devices that prevent the entry of wild birds. An inconvenience … necessary. Will we not be exaggerating? As Eloy Ureña, responsible for the Poultry Sector of the COAG, said In eldiario.es“The ministry’s decision can be an inconvenience for those who have layers, but it is worse to have a positive, because they die and because they suppose a relevant loss for exploitation.” And, of course, we are not talking about a futures. And it is not because, right now, the world lives the greatest outbreak of this flu ever registered. That gives us examples for almost everything. In the US, without going any further, the uncontrollable virus sprouts They have caused a shortage of eggs As you don’t remember. In California, without going any further, the dozen has reached 8.97 dollars: the average price of a dozen in the country in January 2002 was 1.93. In Portugal, the sacrifices have already begun and the risk that the wick pledge in Spain is larger every day that passes. Images | Rowan s | Chris Boland In Xataka |

Operator works differently (and better) to other agents who see our screen. Your secret: Cu

January 24, 2025 by usatoday24

We already have OpenAi’s agent. It is called Operatorand it is a system capable of seeing our screen and performing actions autonomously in the browser from our requests. It is something we had already seen with ‘Computer Use’ from Anthropic either Deepmind Marinerbut here the company led by Sam Altman has its own special ingredient. Computer-Useing Agent (Cua). Operator uses a model called Computer-Useing Agent (CUA) that is based on GPT-4O. CUA interprets screenshots and interacts with websites through the typical browser controls, such as a cursor or a mouse. How Cua works. As they explain In Openai’s documentationthis system processes those “raw pixels” of the captures that you make and use a mouse and a virtual keyboard to complete its actions. Once you have the screenshot, “reason” and follow a “thought” line in which the past actions take into account. Promising performance. There are several benchmarks since they allow to evaluate the ability of these agricultural models. According to them tests performed internally In Openai, CUA achieves a 38.1% performance in OSWORLD (Use of a computer in general) against platforms such as Anthropic, which achieves 22%. Humans, yes, achieve 72.4% on average, which makes it clear that these systems still have a lot of improvement margin. In the use of the browser, the Benchmarks Webarena and Webvoyager also allow Operator to score very high: 58.1% and 87% respectively, compared to 36.2% and 56% of their competitors. What about those catches that I collect operator. Operator continuously performs screenshots to “see” the browser interface with which he interacts. That browser does not run on our PC, but in a remote browser on OpenAI servers. User data, including these catches, are used according to OpenAI’s privacy policy. This is: they can be used to detect fraudulent activities and to improve the service. That implies that our data can be used to train and improve the model, although We can deactivate that option In operator settings. The user, yes, has the capacity for how long this data is stored in Operator. By default these data are saved until the user decides to delete them. An agent who asks for help (and confirmation) when he needs them. As we have seen in other agents such as ‘Computer Use’ of Anthropic, Operator is an agent who does not act crazy. If you meet an obstacle – like a captcha code or the request to introduce user and password on a website – you will ask that the user take control, and will also ask for the final user confirmation if for example we have to validate a reservation or purchase of a product that has sought Operator. The operator user can also take control at all times. This is how it works. Source: OpenAi Do not release the steering wheel. This reminds us of assisted driving systems such as Tesla FSD. It is true that it is able to take us from one place to another once we introduce the destination address, but it is important to continue paying attention and have our hands in the steering wheel in case they occur unforeseen. With Operator and the rest of this type agents something similar happens. There are things that cannot be done. At the moment Operator cannot complete specialized tasks such as managing complex calendar systems or interacting with very personalized or non -standard websites. You will also refuse to do some tasks with high risk of provoking damages. For example, send emails, make electronic transactions or delete calendar events. Its benefits and capabilities will increase, without a doubt, but they will gradually do so and always guaranteeing that the possibility of error is the least possible. Image | OpenAI In Xataka | The generative AI seems stagnant. Big tech believe they have an ace in the sleeve: “agents” who do things for us

In the midst of a campaign against TikTok, ByteDance has an ace up its sleeve: becoming an AI powerhouse

January 24, 2025 by usatoday24

At ByteDance there is frenetic activity. At the moment, your social network TikTok is on a kind of legal limbo in the USbut ByteDance has many more open fronts. One is especially striking: AI. Especially because she had barely shown any signs of being involved in this segment. A most “interstellar” AI agent. First of all, the launch of the UI-TARS project is surprising. The name is probably a nod to the movie ‘Interstellar’, which featured an AI robot called TARS. The proposal is very reminiscent of Anthropic’s ‘Computer Use’ system, but unlike it, UI-TARS is Open Source, something surprising and which suggests that other startups and independent developers can use it as a basis for their projects. The code is available on GitHub. “Talking” to your computer. As in the case of Anthropic, it is possible to ask UI-TARS Desktop that executes tasks on your computer interface, which is capable of analyzing and recognizing to execute them. The developers warn that at the moment performance and behavior may be erratic. Your own programming assistant. Bytedance recently launched a developer assistant. Is called Bring and offers the same advantages as already popular alternatives as Cursor. With this platform we have a “virtual software engineer” who answers questions about the code, completes and generates it, and is even capable of developing projects from scratch based on user prompts. And a model that reasons…. The thing doesn’t end there, of course. ByteDance launched this week Doubao-1.5-proan update to its AI model that is now capable of “reasoning” and thus competing with OpenAI’s o1. In fact, those responsible for it claim that it surpasses its competitor in the AIME benchmark, in which the response of these models to complex instructions is measured. And even though it is affordable, cheap and gives benefits. As they point out on PandaDailythe model is especially efficient and for example the price per million input tokens is only 11 cents. Not only that: even though it costs relatively little, it promises a gross margin of 50%, which in a reasoning model – which usually requires significantly more resources – is striking. {“videoId”:”x9c3qcu”,”autoplay”:false,”title”:”The BRUTAL Chinese SURVEILLANCE system”, “tag”:”video surveillance”, “duration”:”73″} Huge investment in chips. In 2025 ByteDance plans to spend $12.3 billion on infrastructure for the development of its AI models. This is indicated by sources close to the company, indicate in Financial Times. 5.5 billion will be dedicated to acquiring chips in China, while 6.8 billion dollars will be invested abroad to develop their models taking advantage of advanced NVIDIA chips. Be careful with this other Bytedance. These are just the latest developments in the field of AI for ByteDance. This division continues to make losses – like most AI companies – because among other things it is reinvesting over and over again in R&D, and at this rate we may soon have to talk about another AI giant. In Xataka | China strikes back in the AI war: its new open model matches Claude and GPT-4, and it is giving it away (function() { window._JS_MODULES = window._JS_MODULES || {}; var headElement = document.getElementsByTagName(‘head’)(0); if (_JS_MODULES.instagram) { var instagramScript = document.createElement(‘script’) ; instagramScript.src=”https://platform.instagram.com/en_US/embeds.js”; instagramScript.async = true; instagramScript.defer = true; headElement.appendChild(instagramScript); – The news In the midst of a campaign against TikTok, ByteDance has an ace up its sleeve: becoming an AI powerhouse was originally published in Xataka by Javier Pastor .

Log In

Sign In

Forgot password?

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections