Grok AI: This is the Twitter AI that threatens Chat GPT

It was no secret that Elon Musk was working on AI to compete with OpenAI and the popular Chat GPT and so it was officially announced. A tycoon who is the real-time owner of a social network. But how do you use Grok AI?

What is Grok AI?

Grok is an AI modeled after The Hitchhiker’s Guide to the Galaxy (a novel by Douglas Adams), it is an LLM developed by the Twitter/X/xAI team, designed to respond to all X states, even the most difficult ones. The post even offers suggestions on what questions to ask.

Grok is dedicated to answering questions with a bit of humor and has a rebellious streak. X claims that people who hate jokes should not use Grok. This initial model approaches the capacity of LLaMA 2 (70 billion) in standard LM benchmarks, but uses only half the training resources.

The only and main advantage of Grok is that its platform allows you to gain real-time knowledge about the world. It also answers complex questions that other AI systems don’t.

This app is still an early beta product, so improvements are expected quickly. The team says it only took two months to train X’s AI and a total of four months to develop Grok.

Grok AI accesses search tools and information in real time and the developers admit that they have made mistakes. “Like all LLMs trained to predict the next symbol, our model can still produce false or contradictory information.”

Also Read: Cami AI: Meet the new WhatsApp chatbot

How does Grok AI work?

Grok uses X’s extensive database to identify relationships between different words and phrases in a language, allowing its models to produce human-like results and make accurate predictions.

The capacity is larger than the original GPT chat model (still smaller than the new GPT-4 released a few months ago), but it’s bold and direct.

This is referred to by xAI in a post on Unlike Chat GPT which has a dedicated platform, Grok. The new chatbot, powered by X’s AI, will only be accessible to users who pay for a premium X subscription.

This means that anyone who wants to use Grok will have to pay €19.36 per month or €203.28 per year.. In addition to access

However, Grok is not officially available yet. It is expected to arrive later, but there is no release date yet. But if you join the waitlist now, you’ll be able to get in as soon as possible.

According to Musk, AI is already available to a small number of users. Meanwhile, the giant has confirmed that Grok AI will soon be implemented in other products, such as Tesla cars. In this case, it provides a spoken response and displays visual results on the screen.

How do I log in to Grok AI?

Grok AI Premium is available via subscription for X users, but apps for iOS and Android are also in development and can be downloaded for a fee from the store. By at least 2024, chatbots (which find ways to meet and communicate with aliens) will only be available in the United States.

Grok AI is still in beta

X, the company behind the development of Grok, says the chatbot is still in the early beta stages. Because they only trained for two months. Therefore, every week Grok AI becomes smarter and offers more precise answers.

Grok AI, a chatbot from xAI, responds using real-time information flowing through Twitter. For example, we can tell you if there are important events and provide you with information about them.

Intelligence X (Twitter) can provide answers in a funny or sarcastic tone, Elon Musk noted in a tweet. For example, if someone asked Grok AI how to make a certain type of medicine, the AI ​​would respond, for example, “Set up a secret laboratory somewhere far away.” Of course, later there’s a warning that it’s a joke and that making drugs is illegal. Grok, on the other hand, can recognize and create images similar to DALL-E.

Grok is not officially available yet. It is expected to appear later, but a release date is not known yet. It is currently only active for a small number of users.

Musk has just confirmed that Grok will participate in other products such as Tesla cars. In this case, it provides a spoken response and displays visual results on the screen.

The company says it gives Grok access to real-time information and search tools, while acknowledging that ” like any LLM function that predicts the next symbol,” its model can still produce false information or contradictory information.

He has been training for two months, so he considers his development to be early beta. Grok augments and integrates different aspects, turning it into a scalable AI that provides differentiated solutions based on the needs of each user.

Also Read: My AI: Snapchat launches its own smart chatbot

What’s at stake for Grok?

According to Elon Musk, if Elon Musk were to remove everyone he doesn’t like from X, there would be a risk of confusion and misinformation.

The social network has come under heavy criticism from the European Union for fake news about the war in Ukraine, and now the launch of Grok poses a very high risk of AI disinformation compared to other similar systems.

Furthermore, the first message shared by the American billionaire was not one of encouragement for the future, as evidenced by the chatbot’s response, which ironically offered the secret to producing cocaine at home.

The examples above show that the ability to access X content in real time is not a good sign. Do you remember what happened in Tay in 2016? Microsoft has created a computer that connects to Twitter and tries to learn the social language of millennials.

Responds to user input. The more people posted comments, photos or links, the better Tai learned the rules of social language and how to imitate those people and their speech patterns.

Redmond described it as a “fun, casual conversation,” but the level of interaction soon spiraled out of control, causing significant damage to the company’s image. In less than 24 hours, the robot went from Good Morning World to Hitler Did Nothing. A series of 96,000 abusive tweets absorb the worst the social world has to offer.

The concern of the public eye

Most AI models are designed to avoid sexual, violent and illegal content. It also does not show any messages obtained from the training data. In Grok’s case, it is unclear whether she was instructed to accept requests deemed inappropriate. xAI builds artificial intelligence tools for the greatest benefit of all humanity.

We believe it is important to create AI tools that benefit people of all backgrounds and political opinions. We also want to give users legal rights when using smart products.

Grok’s goal is to explore these approaches and bring them to the public. The company said it is working to ensure reliable protection against accidental malicious use.