There is an underlying reason that explains it, according to Bloomberg

The agreement between OpenAI and Apple, announced at WWDC24brought together all the ingredients to become one of the most refined movements in the sector. The artificial intelligence company of the moment landed with its star product in the ecosystem of the technological giant where practically everyone wants to be. ChatGPT began to occupy a privileged position within the devices compatible with Apple Intelligence– The user could take advantage of its capabilities from Siri and Writing Tools, and even link their account to access a more complete experience. And yes, unlike other Apple Intelligence features, like the new Siri, that ended up being delayedthe integration with ChatGPT did end up being deployed on the apple company’s different platforms. What did not materialize, according to Mark Gurmanwere the results that OpenAI expected. The startup led by Sam Altmanexplains the Bloomberg analyst, was confident in an avalanche of new paying subscribers that would translate into billions of dollars in annual revenue, at a time when every dollar counted. Two years after that agreement, even the possibility of legal action begins to appear on the horizon. The agreement, furthermore, was not exactly frozen in that first photo from WWDC24. Bloomberg points out that the integration ended up expanding with new entry doors: users could subscribe to ChatGPT directly from the iOS settings, Apple incorporated it as an option to create images in Image Playground and also used it to analyze on-screen content. On paper, it was no small feat. The problem, always according to the sources consulted by Gurman, is that these expansions were simply not enough. An alliance without money involved, but with many expectations The economic key to the agreement helps to understand why the disenchantment weighs so much. The American media already explained in June 2024 that Apple was not paying OpenAI to integrate ChatGPT into its systems. Nor was it considered, at least at that time, as an operation in which OpenAI would buy its place within the iPhone. The payoff lay elsewhere: Apple offered distribution, visibility, and potential access to a gigantic user base. For a company that needed to turn popularity into recurring revenue, that promise was apparently worth as much as a check. Now, the reality was much less brilliant. Bloomberg says user studies conducted by OpenA suggest that Apple customers turned much more frequently to the standalone ChatGPT application than to the integration within Siri and other system services. That information changes the reading of the agreement. If the user already has the habit of opening the app, the presence in Apple Intelligence stops working as an entrance door and it becomes something more secondary: it is there, but it does not necessarily drive the business. The shape chosen by Apple would not have helped much either. The use of ChatGPT within their systems remained limited. The reproach attributed to OpenAI has a quite concrete logic. Bloomberg notes that the company hoped Apple would put ChatGPT in a more prominent place, not just as an option available if the user knew where to look for it. It also relied on broader integration with in-house apps and more intense promotion within its platforms. But the matter goes further. Gurman assures that OpenAI lawyers are working with an external firm on different legal options that could be executed soon. One of them would be to send Apple a notification by alleged breach of contractwithout this necessarily implying filing a lawsuit from the first moment because it aims to resolve the conflict outside of court. It should be noted that there is no official information, and that Apple and OpenAI did not comment following the request of the American economic media. Apple, for its part, doesn’t seem very comfortable either. The company reportedly maintained reservations about OpenAI, including whether the startup did enough to protect user privacy. Added to this is a more recent tension: OpenAI no longer wants to limit itself to software and works on devices with several former Apple peopleincluding Jony Ive. For Cupertino, according to Bloomberg, the signing of engineers from its hardware teams would also have been especially annoying. That scenario is going to change even more. Already at WWDC24 Apple announced that it would open the door to integrate other AI systems in its ecosystem. So everything seems to indicate that it is a matter of time until OpenAI rivals such as Google with Gemini either Anthropic with Claude land within the apple company’s ecosystem. It would not be an expulsion from ChatGPT, but it would be an obvious loss of prominence. We have to wait to see how this agreement and the relationship between both companies will evolve. Images | Xataka with Nano Banana In Xataka | The ‘Chinese Netflix’ has designed a plan for AI to generate the majority of its content within five years. It sounds risky

Cloudflare is planted in Italy due to blockades. In Spain, the conflict with LaLiga points to the same underlying problem

We are witnessing firsthand how what began as an offensive against unauthorized party broadcasts has transformed into something much broader, a dispute over who can decide which parts of the internet are turned off and how. In Italy and Spain, judicial and administrative resolutions that apply current legislation are endorsing or ordering measures that operate at the network level, measures that, as they are now being applied, may not distinguish between an infringing service and legitimate services that share infrastructure. This scenario has brought to the fore cloudflarea company whose name has been sneaking into the technology conversation for some time. Here we must be clear. What unites the cases of Italy and Spain is not the type of content, but the logic that supports them: to stop the unauthorized dissemination of matches, it has been decided to act where the network becomes vulnerable, in the intermediaries that connect the public with the servers. It is not a button in the hands of a government, but rather a fit between laws, judges or regulators, rights holders and different actors who execute the measure. That strategy allows you to block quickly and with massive range, but it also has collateral damage. Behind every block there is a clear sequence. In Spain, LaLiga takes its requests before a judge and it is the courts that authorize the operators to execute the cuts. In Italy, rights holders enter domains and IPs into Piracy Shield and it is AGCOMthe Italian telecommunications and media regulator, who reviews these signs and converts them into administrative orders that providers must apply. When an authority orders a block, it is not simply saying “close this page”, it is choosing at what point in the journey the connection between the user and the server is interrupted, according to the limits established by current legislation. This can be done by preventing the website name from being translated into a technical address, directly blocking that address, or asking an intermediary to stop serving the data. In this invisible journey there is a particularly sensitive piece, the system that translates website names into technical addresses that computers can understand. Every time we type a URL or tap a link, a DNS resolver responds with the correct IP so the connection can be established. If this translation is interrupted, the page is no longer accessible even if the server continues to function. That is why DNS has become a very attractive lever for blocking, because it allows access to be cut off quickly and without directly touching the content. What is 1.1.1.1 and why is it in the center. Among the many DNS services that exist, there are some open to the public that do not belong to any national operator, and the best known is 1.1.1.1, managed by Cloudflare. It serves as a widely used public DNS resolver that users and applications use to translate domain names into IP addresses. That scale is what makes it especially sensitive in this debate, because any intervention on it is not limited to a country or a specific network, but can have much broader effects. A modem with network cables The company explains For years it has been able to comply with court orders that force it to act on specific clients or on its distribution network, because there it is controlling its own service within a jurisdiction. What it rejects is modifying open tools such as its public DNS by administrative decisions of a single country. In his approach, that would mean that a national authority could change how a basic piece of the internet works for users around the world. Italy, the Piracy Shield system and controversies. The Italian model does not just cut individual pages, but entire pieces of the route along which traffic circulates. Through Piracy Shield domains and IPs are ordered to be blocked and, according to the regulator itselfthe framework also expressly includes public DNS services and VPN providers as obligated parties when they are involved in the accessibility of that content. Cloudflare Global Network Map The problem is not only that the system blocks a lot, but how it does it and with what margin for rectification. Its quick reaction logic prioritizes cutting access while the event is happening, and that increases the risk of affecting third parties when acting on shared parts of the network. AGCOM quotes as balance that since February 2024, more than 65,000 FQDNs, that is, fully qualified domain names and about 14,000 IPs, have been disabled. That clash took concrete form at the end of 2025. In a decision taken on December 29 and recently notifiedAGCOM imposed a penalty of more than 14 million euros on Cloudflare for failing to comply with a previous order issued on February 18, 2025. According to the regulator, the company had to deactivate the DNS resolution of certain domains and the routing of traffic to IP addresses indicated through Piracy Shield, or apply equivalent measures to prevent users from accessing that content. Spain, the judicial path. As we mentioned above, in Spain the system is not based on an administrative regulator, but on a resolution from a commercial court obtained by LaLiga. On December 18, 2024, the Commercial Court No. 6 of Barcelona authorized blocking measures against addresses used to broadcast matches without rights. On March 26, 2025, that same court rejected the challenges and left the order in force. That is what allows access operators to execute these blocks during matches under the direct legal coverage of a judge. The way that order is executed in practice explains many of the complaints that have arisen in Spain. Access providers block entire IP addresses, not just specific domains. This mechanism explains why so many legitimate services end up dragged down by these blocks. Instead of deactivating a specific domain, operators sever an entire IP address, which is often shared by hundreds or thousands of websites. It’s a bit like boarding up the entrance to a building … Read more

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.