An economic science fiction text has sunk Visa and Mastercard in the stock market. The reason is more disturbing than the story itself

Citrini Research, a hedge fund American published this week a text written as if it were a macroeconomic memorandum from June 2028. It is not a prediction, its authors warn. It is a speculative exercise. A feasible scenario. It has achieved 24 million impressions, and counting. It is not an anecdotal tweet. The markets they have responded by sinking. Visa has fallen 4.4%. Mastercard, 6.3%. American Express, almost 8%. And Capital One, 8%. This deserves an explanation. And it’s not what it seems. Between the lines. The market reaction is not explained by the specific content of the Citrini Research report, which includes arguments as debatable as that AI agents will abandon cards to pay with stablecoins in Solana. Antonio Ortiz, technology analysts, has pointed it out precisely: part of the argument “it is from the first of Twitter AI-hype“. The idea that an agent will compare twenty food delivery apps vibecodeadas to find the cheapest one smells like a caricature of the future. But the panic is not irrational. It is precisely the panic of not knowing where the limit is. Why is it importantand. What has moved the market has not been so much the thesis about payments but the thesis about the destruction of value. And that is solid: many billions of dollars of market capitalization have been built on a single foundation: that humans are slow, impatient, forgetful and loyal out of inertia. That we do not compare prices. That we renew subscriptions that we do not use. And that we pay commissions that we do not negotiate. An AI agent has none of those weaknesses. And that changes everything. The backdrop. Citrini’s report comes at a time when the so-called “saaspocalypse“is no longer a metaphor. WSJ states that investors are terrified by the possibility that AI ends up doing the work that large software companies bill for today. ServiceNow, Salesforce, business management platforms… all built on the premise that companies need software for their employees to do their jobs. But… what happens when employees disappear? What if the software itself can be replicated in weeks with agentic coding tools? Citrini’s fiction begins exactly there, in early 2026, when a competent developer can reproduce the core functionality of a mid-market SaaS in a few weeks, and constructs a scenario of systemic collapse. The big question. The report’s most disturbing argument is that in every previous technological cycle, job destruction created new jobs that only humans could do. This time, AI is already occupying those new positions as well. If that’s true—if AI improves faster than workers can reorient themselves—the self-correcting mechanism that has always kept creative destruction from turning into outright destruction wouldn’t work. That is the scenario that the markets have discounted this week, even if only partially and speculatively thanks to a creepypasta financial. Yes, but. The scenario requires assuming a speed of adoption that is not guaranteed, a completely absent political response and a total absence of new economic sectors. None of the three conditions are set in stone. Furthermore, as Antonio points out, there is some collective hysteria in the reaction: each announcement or “scary story catches attention and moves investors.” Markets are trading in panic over the unknown. But there’s an important difference between saying “this scenario won’t happen” and saying “this scenario is impossible.” And that difference is exactly what has the market nervous. The alarm signal. The most striking thing this week is that a speculative text, written in economic science fiction format, has been enough to move billions in market capitalization. That says a lot about the state of certainty in the markets regarding AI: it is practically non-existent. Nobody really knows how much a company whose moat It is human friction in a world where that friction is disappearing. The canary is still alive. But investors have stopped trusting the canary. In Xataka | AI promised to revolutionize all sectors. It has only revolutionized programming while the rest is still waiting Featured image | Avery Evans

A 28-page US document has brought peace in Ukraine closer than ever. The problem is that it is the translation of a Russian text

And suddenly a 28 page document unpublished to date has suddenly entered as a missile in the negotiations of the war in Ukraine. Promoted by Washington, it has unleashed a diplomatic storm in Europe and in kyiv because, far from having been prepared with the main parties involved, it had been conceived in discreet negotiations between the American businessman Steve Witkoff and the Russian envoy Kirill Dmitriev, with the participation of Jared Kushner and the late endorsement of Trump. The origin of a plan. The result of these meetings was a text that Europe and Ukraine had not seen and that, to further alarm (according to one Bloomberg exclusive), preserved the linguistic structures typical of an original written in Russian, confirming the suspicions that Moscow had achieved filter your vision of the war in a document presented as a US initiative. The pressure exerted by Dan Driscoll (a close ally of JD Vance) on European and Ukrainian diplomats, urging them to accept territorial concessions in a matter of days, ended up setting off all the alarm signals. For European governments, which considered themselves central partners in any peace negotiations, the origin of the plan became a strategic question: they needed to know who had written it and with what objectives before sitting down to discuss. This information gap triggered a race against time to stop the imposition of a text that, in its initial form, was not only surprising for its demands, but also for its obvious alignment with Moscow’s interests. Territory, legitimization and a threat. The most explosive section of the American plan required that Ukraine will withdraw of the fortified urban centers that it still maintains in Donetsk, breaking the “belt of fortresses” that has slowed the Russian advance since 2014. This withdrawal would not only imply the displacement of tens of thousands of Ukrainian citizens, but it would open a corridor that would leave exposed to key cities like Dnipro and Zaporizhzhia. To make matters worse, the document proposed that the occupied areas be recognized as “de facto Russian”a more favorable formula for Moscow than the already problematic “de facto under Russian control”, and which, in practice, brought the international community closer to accept territorial changes achieved by force. Added to this was the idea of ​​converting the evacuated territories into a demilitarized zone whose violation by Russia (not an implausible scenario given recent history) would allow Moscow to open a new, even deeper offensive in the future. From the Ukrainian perspective, accepting this point would be sowing the conditions for a future war in worse terms, reinforcing the impression that the document did not seek a stable peace, but rather formalized a strategic result that Russia has not been able to obtain through military operations. Security cut and promises broken. The security guarantees included in the plan were vague to the point of irrelevance: they promised “reliable protection” without detailing mechanisms, but simultaneously prohibited Ukraine from entering never in NATOprevented the stationing of allied troops in its territory and forced kyiv to modify its Constitution to renounce accession. For a country marked by the experience of Budapest Memorandum (formal guarantees that prevented neither the annexation of Crimea nor the 2022 invasion), accepting an even more ambiguous framework would amount to to be left helpless facing an aggressor who has systematically broken all previous agreements. Red lines. The absence of a commitment type Article 5 and the refusal to allow training missions or deterrence forces on Ukrainian territory reinforced the conviction that Ukraine would be trapped between a strengthened Russia and a West that would reserve the right to “diplomatically support,” but not to intervene. This component fueled rejection in European capitals, which consider it essential that Ukraine keep an army strong as a land barrier that protects the continent. Limit to 600,000 troops to the only country in Europe at war, without imposing a similar restriction on Russia, was perceived as covert disarmament and a prelude to a future Russian offensive. Amnesty and frozen assets. One of the most shocking elements of the plan was the proposal of a general amnesty and Ukraine’s renunciation of any legal claim about war crimes, deportations or deliberate destruction of infrastructure. For an exposed population to documented atrocitiesthis clause meant not only the denial of justice, but also the elimination of the legal basis that allows Europe to advance the reparations loan backed by frozen Russian assets. That loan, of 140,000 million of euros, is considered by the EU as the more solid path and less expensive to sustain Ukraine during the postwar period. The US plan not only made it unviable, but also redistributed those funds in an unusual way: 100 billion would go to a US investment vehicle that would deliver half of its profits to Washington, another 100 billion would be contributed by Europe and the rest would go to a joint fund with Russia. For Berlin, Paris or Warsaw, the message was clear: Russia would obtain indirect financial relief while the Europeans would see their most effective tool of strategic pressure weakened. The attempt to force kyiv to renounce all moral and legal responsibility for the aggressor reinforced the perception that the plan sought to resolve the war “quickly,” not “fairly.” The Russian strategy. Since the beginning of the invasion, Moscow has not changed their fundamental demands: more territory in the east, military neutralization of Ukraine and permanent veto on its accession to NATO. This strategic immobility, together with gradual advances on the front, has allowed it to capitalize on Western fatigue, the political fractures in kyiv and transatlantic tensions. For the Kremlin, the leaked plan demonstrates that its commitment to prolonged resistance, military pressure and the erosion of Western will is bearing fruit. Putin openly celebrated it, affirming that the document could serve as a basis and that rejecting it would only lead to new Ukrainian defeats. Likewise, Moscow has hinted that even a signed agreement could be used as leverage to resume the … Read more

How to create a podcast from a text to study, investigate or simply listen if you don’t feel like reading

Let’s explain How to create an audio from text wearing artificial intelligenceso that you can have a kind of podcast to study or review something when you do not feel like reading. Thus, instead of having to be reading a text, you can simply relax while listening. We are going to start the article with a series of previous tips and recommendations to take into account before putting the task. Then, we will tell you how you can do it with one of the best tools available for it, and we will continue with two ways to do it with assistants of AI as Chatgpt And some other alternative that is also interesting for Generate voice with artificial intelligence. Before starting, things to take into account Before starting to get to work with this, there are some things that you must take into account. First, you must define whether this audio or “podcast” that you want to create is educational, informative or only entertainment. Try too Be clear if you are going to listen to only you or also some other person, in order to polish how well finished it is. A written text does not always sound natural when you listen to it, especially when you have longer and somewhat intricate phrases. That is why it can be recommended that The phrases are short and close languagenot something that makes you heavy listening. In cases where you are going to be able to do it, Decide tone and voice stylechoosing between one more formal or informal. This may depend on the content you want to consume in audio, and also if you are going to listen to only you or other people. It is also recommended that You take care of the audio duration. The recommended duration may vary, because if it is something dense or to study it may be that from 20 minutes begin to get tired, but if it is for leisure or something narrative you should not have problems with longer durations. If it is something dense, try to segment the content well. And finally, Always check the result and do not be afraid to try again If you don’t convince you. Because sometimes the voices of AI may sound unnatural or there may be errors, and that is why it is important to check it even if it is a bit. In the end, The most important part is in the script That you will create to be narrated, and in this it is what you will have to spend most of the effort. Try to be natural, with good scores and well structured. Audios from your notes with notebooklm The first tool you can use is Notebooklm from Google, a service with which you can use artificial intelligence to organize your sources, and even Create an audio summary As if it were a podcast. It is a kind of chatgpt, but in which everything you do will be based on the sources you have added by hand. You can use notebooklm Through its official website or by mobile applications. The official website is Notebooklm.google.comor also in their apps In Google Play For Android and In the App Store For the iPhone. You can use it for free, but its payment version will give you more audio summaries. The first thing you should do with this tool is to create a notebook or work space. Inside, In the left column you can add sourceswhich are the files that the AI ​​will use to obtain the information. They can be text documents, slides, PDF, YouTube videos or links to web pages or online items. Once you have added the sources, Go to the section of Studiowhere the tool will create an audio summary of the content of all files. With this you will create your personal podcast To get information about everything you want. How can you do it with chatgpt and others AI Another option is to use a generative AI such as Chatgpt, COPILOT, Gemini either Deepseek. These tools They do not allow you to create a downloadable audio file although this is something you can do with other third -party tools. What you can do is listen directly to the AI ​​app. What you have to do is Create a summary script of an article. Then, this script can be heard directly in Chatgpt, Gemini or other platform, or take it to another AI that generates a downloadable audio. Let’s explain everything step by step. To start, you have to tell Chat GPT to make you summarize an article or a web page. For that, You must attach it and include the prompt or the instructions to generate your script. The article from which you want the summary can be included by uploading the text file or PDF, or directly putting the website. The prompt that we have used is as follows: “I want to create an audio to listen to a summary of the content of this website. I want you to generate the script to then copy and paste it into a program that passes from text to audio. The script has to be narrative, without structure, you simply have to write it to read it from there. (Link)” As you can see, in the prompt we have reiterated that the text generated by chatgpt It must be natural and should be read directlybecause sometimes if you do not mention it, you can tend to generate a script scheme with things to be filled, and what I want is to generate a text to copy and paste. Here, It is you who decides how and what do you want to make the summarybeing able to be one or several files that you raise or add. Just remember that Everything you upload or vintages will be saved on servers of the company that owns the AI, so be careful in the event that you are adding sensitive data that you … Read more

How to generate one in voice artificial intelligence from a text: Create your personalized announcer

Let’s tell you How to generate a voice with AIto be able to tell things from a text you write to him. This will allow you to have a personalized announcer or broadcaster to create your own podcasts or audiobooks, and to read you anything you want. First, we are going to give you a series of previous tips for you to know what you should have in mind to use these tools for artificial intelligence. Then we will tell you four free tools you can use To move from text to voice and use voices by AI to create audios. Before starting, some tips Before put to create locutions with a voice made by artificial intelligence you must think about What do you want to get with her. Think about whether you want it to be a natural voice or a robotic narrative, and also in the language, the tone or accent you want to use. It is very important have written a well structured text so that then the voice of the say it. Use the punctuation marks well so that the intonation sounds natural, and if it can be, divide the text into short phrases to also improve naturalness. Tests before getting seriouslyUse some short phrases to test your voice and see if you are using a tool or configuration that convenient you. And before that, Listen to several voice examples If the tool has several available to better choose the one you want to use. Finally, be aware of the limitations of the platform you use. Most are paying, and In their free ways they have limits by number of characters or minutes of audio generated. Therefore, if you see it necessary, you may have to divide a long text into different fragments instead of trying to make it whole. Generates phrases with notebooklm The first tool you can use is Notebooklmwhich is free and from Google. In this case, You can only make audio summaries Of one or more issues that you upload on the web, you will not be able to write what you want me to say the voice. But it is totally free. To use it you have to enter Notebooklm.google.comand on the main page click on the option of Create notebook. It is also available In Google Play For Android and In the App Store For the iPhone. Notebooklm is divided into notebooks, which are work spaces. The first thing you will have to do, therefore, is to create a new one. Once inside, You will have three columns or sections. On the left you have the fountains. Here, You can add one or more sources. Everything you upload will be what analyzes Google’s artificial intelligence when you ask you questions or ask for content. They can be text documents, slides, PDF, YouTube videos or links to web pages or online items. Then you have the section of Chatwhich is where you will be able to ask Google’s AI all the questions you want and get answers based on the sources. And then you have the section of Studio. In it, you will be able to create an audio with a summary of the sources provided in this project. If you click on the Customize Audio Summary buttonthen you will be able to determine how the summary that your voice will make, giving indications on the subject to speak or the source to focus, and you can also have the way this voice speaks. Elevenlabs Elevenlabs is a platform with several artificial intelligence resources, including one for AUDIO TEXT. You can find it on the web Elevenlabs.io/es/text-to-speechand although you can try it without registering you will need to create an account to download the audio you generate. It is very easy to use. First you write or fight the text you want to narrate, and then you will have to Choose the voice and language What do you want to use, being able to determine the model to be used and the speed at which it is discussed. There are voices in Spanish from Spain, and also Latinas. Then you press in Play And ready, reproduce. Here, the bad news is that The free account has limit In the number of characters to process every month. Specifically, you can create 10 minutes monthly of high quality text -based audio. TTSMaker This is another tool to move from text to voice with artificial intelligence, which is characteristic by Do not need to create account To use it for free. You just have to write the text, choose the language and ready, generate what you have written in voice. You can use it from the web TTSMaker.com. Here, you should know that each audio has a maximum of 1,000 characters, and that you can Use 20,000 characters a week being a free user. For more, you will have to pay. They are very generous margins, although the web has enough advertising. You also have options how to choose the audio format you want to generate, or listen only to the first 50 characters before creating the audio to make sure you are to your liking. You can also configure the speed, volume, audio quality, the length of each pause. Many things. Clipchamp This is a Microsoft tool to create videos, including the option to do so for the. However, You have an option of Text to voice When creating, with which you can write a text you want and choose the voice to use. You will need to enter the website clipchamp.com and log in with your Microsoft account (Hotmail or Outlook). Once the text option is chosen to voice, You will have the creation options in the right column. There, you can hit the text you want, choose the language and voice. In addition, there are advanced options to choose the voice passage and the rhythm you use to read what you write. And ready, with this, what you write will … Read more

How to translate a sign, sign or any text with chatgpt, even asking you to explain it to you

Let’s tell you How to translate texts and posters with Chatgptand thus take advantage of its image recognition characteristics. This is something that better or worse you can also do with other alternatives, such as COPILOT, Gemini, Grok either Deepseekbut we are going to focus on Openai’s. We have done the tests in the mobile version, but you can also do it in chatgpt in the browser or in the computer application. In addition, it is a function that is available for both free and payment users, although perhaps payment receives more advanced summaries and explanations. In any case, now that artificial intelligence is becoming a tool that we get used to carrying with us, in this context it can be very useful when you travel. You can translate texts with a large amount of languages, and even ask for summaries of them. All you need is have Internet connection To be able to use it. Translate texts and signs with chatgpt The first thing you have to do is click on the option to attach content to the message you are going to write in Chatgpt. Here, you can Add the photo of a text or poster or use the camera To take it out in the event that the text is in front of you. When you add an image, do not send it directly to the chat. You have to accompany her from a petition Or prompt asking him to translate what he puts. Remember that you have to use natural language, you can ask you to translate what you put in the photo in several ways. When you ask him to tell you what he puts in the photo you have sent him, Chatgpt will explain everything. It will give you A translation of the textand depending on the version you use will give you context and will detail everything you see necessary for you to understand it well. You can also Ask him to make you a summary Fast or detailed what it puts in the text. Thus, in addition to translating it you can also summarize it in the event that it is a very broad text, and you will understand it well. In Xataka Basics | Seven tips to take advantage of the chatgpt photo editor

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.