The artificial intelligence of xformerly known as Twitteris now available to all users of the social network, regardless of whether they are paying subscribers or not. Grok, this is the name of the chatbot developed by Xaiis accessible through a new icon in the app, ‘/‘, and the option Grok in the web version, but with limited use for free account users: 10 responses every two hours.
Xai is the company that Elon Musk announced last summer to develop large language models (LLM) and that already In November it made the first version of this LLM available to premium users. The movement was striking because by then the millionaire had months warning of the dangers of artificial intelligence and the hasty development of this technology. Then he jumped on the bandwagon with an AI that boasted of not to be woken and that it had a more sarcastic tone than ChatGPT or Gemini. A year later, it is available to all users with a new, more advanced language model, Grok 2. It has gained speed, Internet access, the ability to generate images and is multimodal.
What you can do with Twitter AI
Grok is not particularly different from ChatGPT, Gemini, Claude or other LLMs, although it contains fewer safeguards and offers both more creative freedom as the ability to generate content that may be controversial.
The user can use it to perform a wide variety of queries, generate images, code, etc. It has a ‘fun mode’ which is accessible via a slider in the web version menu, but does not yet appear in the app. Grok is connected to the Internet and searches in real time for information about what is asked and maintains a record of all the conversations that the user has and the images that the chatbot generates.
You can also ask about the latest trends on Twitterwhich in the web version appear in the Grok section in addition to their usual location. Just click on any of them to obtain the necessary information. And you can also use it to perform searches on Elon Musk’s platform.
When accessing the chatbot, some of the presentation texts are displayed in English, but don’t worry. Grok understands Spanish and responds in this language when asked using it. The first time it is accessed, the chatbot warns that it may make a mistake and that X will use your data from X and your interactions with Grok to train the model and personalize the user experience. If you prefer not to be like this, you must access the web version of X, follow the path More options > Settings and privacy > Privacy and security and disable the option Grok and external collaborators.
Grok and Aurora, a new language model to generate photorealistic images
With Grok 2, in addition to the usual functions of a chatbot, you can generate images with the FLUX Beta language model from Black Forest Labs. This Monday, X announced that it was incorporating a new model into Grok 2 to generate more realistic images. Aurorahowever, it only remained in the app for a few hours and the option to use it disappeared from the X menu.
Xai subsequently announced Aurora on his blog, presenting it as ‘an autoregressive expert-mixing network, trained to predict the next token from intertwined text and image data.’ We train the model with billions of examples from the internet, giving it a deep understanding of the world. As a result, excels at photorealistic generation and precise following of text instructions. Beyond text, the model also has native support for multimodal inputallowing you to draw inspiration from user-provided images or even edit them directly.’ The company stated that It would be rolled out to all users throughout this week. Given the quality of the images that free account users are publishing and that are already trending, it seems that you won’t have to wait even that.