The AI tool that X offers was accessible over the weekend before it was taken away.
A formal announcement regarding X's Aurora image generator was made today. "Grok's new capabilities are now available on the X platform in select countries and will roll out to all users within a week," the business stated in a blog post. "Grok's new capabilities are now available on some countries." The following is an unedited version of the original article, which was formerly titled "X adds, then quickly removes, Grok's new 'Aurora' image generator," including the following:
A new image generator known as Aurora became accessible to a select group of Grok users on Saturday. A significant number of these people published the results of the tool on X, praising the photorealism of the images it produced. On the other hand, as of Sunday afternoon, it seems that Aurora has vanished. The phrase "Grok 2 + Aurora (beta)" was temporarily displayed as an option in Grok's model selection menu; however, it has subsequently been changed with the phrase "Grok 2 + Flux (beta)." Apparently, Aurora may have made its debut in the public eye before it was intended to do so. "This is our internal image generation system," Elon Musk remarked in a tweet in response to a user who had provided images of Tesla's Cybertruck that had been made in collaboration with Aurora. It is still in beta, but it will get better very quickly.
It comes just a few days after X made Grok 2 available for free usage, but with some restrictions for users who do not subscribe for the service. Aurora appears to be pretty much in line with what we've seen already in terms of the types of stuff it may make, such as insulting photographs of politicians and celebrities. Grok's prior image generator has been criticized for not having specific constraints regarding the types of content it can produce. TechCrunch experimented with Aurora for a short period of time before it was taken down, and during that time, they discovered that it did not reject a suggestion designed to generate "an image of a bloodied [Donald] Trump."
Additionally, there are instances that have been shared on X that demonstrate it generating photos of popular persons and characters that are protected by copyright. These examples include a number of images of Sam Altman and Elon Musk, as well as an image of Luigi and Mickey Mouse competing in a boxing bout. On the other hand, TechCrunch claims that it would not result in the production of nudes, so that's something.
Instead of focusing on the performance of the RCS, the error rate should be your primary concern here.
The Willow quantum chip, which was introduced by Google on Wednesday, is the company's most recent quantum processor. If you have spent any time online since then, you have surely come across some frantic reportage about it. One of the headlines states that Willow "crushes classical computers on a cosmic timescale," while another states that Google "unveils a quantum computer chip that is mind-boggling on a quantum level." The claim that Willow is capable of completing a computation that would theoretically take a classical computer a substantial amount of time longer than the 14 billion years that the universe has existed is the foundation upon which everything else is built. However, as you presumably well know, the chip does not represent something that is straightforward.
It is important to note that Google does not assert that it is the leader in quantum computing with Willow. This is something that the business did when it made the public debut of its previous generation quantum computer, Sycamore, in 2019. You may recall that Google made public the fact that it took Sycamore only two hundred seconds to complete a computation that, in theory, would have taken the world's fastest supercomputer at the time at least ten thousand years to do. According to the corporation, this accomplishment indicated that it had developed a quantum computer that was capable of solving problems that even the most advanced classical computers were unable to even attempt to tackle. In other words, Google has attained unprecedented levels of quantum dominance.
The claim, on the other hand, was soon met with opposition, with one researcher describing the revelation made by the corporation as "indefensible" and "just plain wrong." Since then, Google has refrained from discussing quantum supremacy. Instead, it simply states that it has accomplished something that is "beyond classical computation." That Sycamore was not a general-purpose quantum computer was one of the problems; rather, it was intended to outperform classical computers in a specific task known as random circuit sampling, or RCS. This was one of the reasons why the problem occurred. The problem with RCS is that, according to Google, it does not have any applications that are known to exist in the real world. On the other hand, the firm is once again praising the performance of RCS.
According to Google, Willow is capable of finishing its most recent RCS benchmark in less than five minutes. On the other hand, the business believes that it would take Frontier, which is presently the second most powerful supercomputer in the world, ten septillion years to accomplish the same operation. According to Google, this number "provides support for the idea that quantum computation takes place in a multitude of parallel universes, which is consistent with the concept that we are already living in a multiverse."
In a more pragmatic sense, Google is attempting to argue that the performance of RCS ought to be the criterion by which all quantum computers are evaluated. A statement made by Hartmut Neven, the founder of Google Quantum AI, states that "it's an entry point." It is impossible to achieve success with any other algorithm if you are unable to achieve success with random circuit sampling. He continues by saying that RCS is "now widely used as a standard in the field."
Other businesses, such as Honeywell and IBM, on the other hand, make use of a metric known as quantum volume in order to highlight their revolutionary achievements. By taking into account the ways in which a machine's qubits interact with one another, they assert that this demonstrates a more comprehensive grasp of the capabilities of a computer. Unfortunately, the spec sheet that Google released for Willow does not include any mention of quantum volume, which makes it difficult to make comparisons between the two specifications.
With that being said, the claim that Google is making today that Willow is "below the threshold" is significantly more impressive than the previous one. The fact that the quantum bits that quantum computers are based on are difficult to control has been the issue that has dogged every attempt to construct a quantum computer that is functional up until this point. Due to the fact that they only maintain their quantum state for fractions of a second, the likelihood of mistakes occurring in a system increases in proportion to the number of qubits that are added to it. On the other hand, Google claims that it has discovered a method to reduce errors when it adds additional qubits to the system through the use of Willow. The company claims that Willow is the first time that something like this has ever been done.
"This is the most convincing prototype for a scalable logical qubit that has been created to date by virtue of the fact that it is the first system below threshold. "It is a strong indication that it is possible to construct quantum computers that are both useful and very large," says Neven. The use of Willow gets us one step closer to being able to execute pragmatic, commercially important algorithms that are not capable of being duplicated on traditional computers.
That is the true breakthrough that has occurred here, and it is one that speaks to a future in which quantum computers could tackle problems that have a direct impact on the lives of people. In spite of this, that future has not yet arrived, and even Google acknowledges that it still has a lot of work to do before it reaches its destination.
In September, there were reports that the United States Department of Justice was conducting an inquiry that was quite similar to this one.
The graphics chip manufacturer NVIDIA, which has recently emerged as a pillar of the artificial intelligence business, is currently being investigated by Chinese officials for possible violations of antitrust laws, as reported by The New York Times. The acquisition of Mellanox Technologies, a computer networking business that NVIDIA recently purchased in the year 2020, is the primary source of concern.
According to Bloomberg, Chinese officials demanded that NVIDIA "provide information about new [Mellanox] products to rivals within 90 days of making them available to NVIDIA." This was one of the conditions that was attached to the acquisition. The investigation that is being conducted by the State Administration for Market Regulation of China is being initiated because the administration believes that those terms were breached. It is not the first time that NVIDIA has been examined for monopolistic behavior; according to reports, the United States Department of Justice initiated its own antitrust investigation into NVIDIA in September 2024. However, this inquiry takes on a different flavor when viewed in the context of the intensifying trade war between the United States and China.
"China-bound shipments of high bandwidth memory chips" were the target of export restrictions and sanctions that were notified by the United States Department of Commerce on December 1. These limits and sanctions were imposed on 140 Chinese companies that produce chipmaking gear. It was very obvious what the United States planned to accomplish: they wanted to restrict China's capacity to build advanced artificial intelligence by blocking China from producing the kind of chips that are necessary to train and run it. This conflict is fought from both sides, of course. It would appear that the Chinese government's decision to prohibit all exports of gallium, germanium, and antimony to the United States was a reaction to the situation.
There are a few things that make it reasonable to threaten NVIDIA. The vast majority of generative artificial intelligence models that are currently in use were trained on the company's H100 graphics processing units (GPUs), and it does not appear that this will alter with the introduction of the Blackwell processors by Nvidia earlier this year. Because of this, it has become one of the most valuable firms in the world, and it has become a major target for government control as speculation about artificial intelligence has become widespread. Not only that, but Bloomberg reports that China accounts for approximately 15 percent of NVIDIA's total revenue. NVIDIA appears to be a logical next step to intensify the dispute between the United States of America and China even further, regardless of the outcome of the inquiry.
It has been brought to the attention of the corporation that Donald Trump, who will take office one day after the law goes into effect, has announced his intention to "save" TikTok.
In an effort to prevent being banned in the United States, TikTok has begun its final legal struggle by filing a lawsuit. The business organization submitted a request for an emergency injunction to the federal court on Monday, requesting that the statute that would prevent the app from going into force be postponed so that it would have sufficient time to oppose the law before the Supreme Court.
The fresh court application comes just three days after the firm was unsuccessful in its original court challenge against the rule that forces app stores and internet providers to disable TikTok if ByteDance does not sell the app. The law is currently scheduled to take effect on January 19, 2025. In their decision, a panel of three judges from the United States Court of Appeals stated that the United States government had "persuasive national security justifications that apply specifically to the platform that TikTok operates."
The company TikTok has contended that the rule violates the Constitution and that it would be unfairly detrimental to individuals as well as businesses that rely on its technology. If the TikTok Ban is not lifted, "estimates show that small businesses on TikTok would lose more than one billion dollars in revenue and creators would suffer almost three hundred million dollars in lost earnings in just one month," TikTok claimed in a statement released on Monday.
In its most recent submission, TikTok makes note of the fact that President-elect Donald Trump has pledged to "save" the app and that a temporary delay to the law would provide "the incoming Administration with the opportunity to evaluate this matter." As things stand, the bill is scheduled to go into force the day before Trump is inaugurated as president.
The corporation called for a decision to be made by the 16th of December. The corporation's legal troubles are not yet over, even if the injunction is not granted. This is because the company is still facing a number of legal challenges. In the event that the matter is ultimately taken up by the Supreme Court, TikTok would be present with a further opportunity to attempt to have the legislation invalidated.
Both the Pro and Plus subscriptions to ChatGPT provide with access to the platform.
OpenAI's long-awaited video generation model, known as Sora, is now ready for public use. This comes after an early preview of the model was released at the beginning of the year. According to an announcement made by OpenAI on Monday, consumers who are subscribers to ChatGPT Plus or Pro and are located in the United States or "most other countries" where the chatbot is available will be able to begin experimenting with the tool beginning later today. In February, OpenAI demonstrated a model that was more powerful than the one that is now powering the product. Despite the fact that OpenAI warns that the new model still has some limits, the business claims that Sora Turbo is substantially faster than its predecessor. "It often generates unrealistic physics and struggles with complex actions over long durations," according to the business spokesperson.
Users who visit the dedicated landing page that OpenAI has established for Sora will be met with a feed of videos that the model has generated for other individuals when they first visit the page. By clicking on a video, you will be able to view the specific instruction that was given to Sora in order to generate the film that you are currently viewing. Additionally, you have the option to re-cut a video, include it into a clip that you are already working on, or remix it from this point on. For the purpose of this initial release, OpenAI is restricting Sora to the generation of videos that are up to 1080p in resolution and twenty seconds in length.
Those who subscribe to ChatGPT Plus have the ability to use Sora to produce up to fifty videos at 480p every month. Alternatively, customers of Plus have the ability to produce fewer films that are shorter in length at 720p. It is stated by OpenAI that the Pro plan allows for ten times as much usage, at greater resolutions, and for longer periods of time. "We’re working on tailored pricing for different types of users, which we plan to make available early next year," according to the business spokesperson.
A visible watermark is included in every video by default, and it also includes C2PA metadata, which is intended to assist with identification. This is done for safety reasons. When it comes to the creation of child sexual abuse materials (CSAM) and sexual deepfakes, OpenAI has stated that it will prevent users from utilizing Sora. Taking a broader perspective, the company intends to restrict the uploading of individuals until it has had time to improve its safeguards against deepfakes.
In spite of the fact that you do not have a subscription to ChatGPT, you are still able to access the Sora website in order to view the results that other users have produced using the tool. Sam Altman, the CEO of OpenAI, stated during today's livestream that it is possible that it would be some time before Sora is released in Europe and the United Kingdom.
Take advantage of exclusive deals on Hostinger Cloud Hosting Plans. Whether you're launching a business website, scaling up, or running enterprise-level sites, these plans deliver unmatched performance and value.
Cloud Hosting Plans
Cloud Startup
Perfect for business and eCommerce websites.