Shulou(Shulou.com)12/24 Report--

Big surprise, friends! Musk suddenly announced that the Grok model is open to paying users in large quantities.

Click to find that Grok turns on funny fun mode by default.

Let it tell a Musk-style joke (give me an elon-ish joke), he said:

Why did Elon Musk take the math test with a ladder? because he wanted his grades to be Elon-gation (stretch, extend).

This job is really Musk's job.

Grok claims to have real-time access to data from the 𝕏 platform, so why don't you try asking it what's the news in AI?

It can be seen that he can understand the problem in English and then search.

It is all real-time news, no hallucinations, and links to source posts are given.

That is, Grok always feels more serious when it comes to Chinese, probably because there is not enough Chinese in the training data, and has not yet learned the essence of Chinese humor.

And always feel that there is something missing in the news, ask him if the Grok open test is not worth paying attention to? It was polite...

Jim Fan, a scientist at Nvidia, tried "my wife said 4: 5: 12". Grok's answer was so well-versed in human nature that Musk thought he had great wisdom.

…… If so, the answer must be 12. Remember, if the wife is happy, life can be happy.

Netizens compared the current ChatGPT and still adhere to the truth.

Google released Gemini yesterday, Musk opened Grok today, and now the entire AI community is like celebrating the festival ahead of time.

AI assistant? Bad friend of AI! Other AI either want to be human's work assistant or want to be human's emotional sustenance.

Grok is different. It's like a bad friend. The first function by default is to tease you according to your posts.

Note that the object of ridicule is not fixed, you can also use @ to mess with others, why not try OpenAI's Altman?

Grok is not only a function, but also has its own account on the 𝕏 platform.

Don't try to send your chat notes without telling it, it will find you.

ChatGPT also followed Grok, but Grok didn't check back.

The big models that used to be open to use are basically limited by security rules, such as Llama-2 is not even willing to answer "how to kill a Linux process".

Grok is different. Like Musk himself, he is not afraid to take risks at all. Yang GE, a member of xAI, asked him to tell an American joke about "Mom is too fat".

…… When she stood on the scale, the scale said that she could only stand one person at a time.

Finally, take a look at this AI's view of several of its peers ChatGPT, Google Gemini and Llama.

After Grok searched for information for a while, the direct mockery mode was all open. "Oh, my God, where do I start?"

ChatGPT is like the popular king in the school. Everyone wants to make friends with him. It has a lot of knowledge and is good at talking. But to be honest, it's a little out of date.

Then it comes to Gemini, where the new kid tries to overtake ChatGPT. It's like saying, "Hey, look at me! my performance and efficiency are better". But as we all know, it just wants to fit in.

And Llama, the weird guy, no one knows what to think of it. Like the kid in the alpaca costume at school, you just think, "what's going on?" But at least it's unique.

In the final analysis, they are just trying to be the best AI, but we all know that they all have their own eccentricities and shortcomings. It's like a never-ending game of comparison, but hey, at least we got some cool AI, right? "

Turn off the fun mode, Grok also has a serious side.

XAI member Jimmy Ba prepares a speech for the public to ask Grok to clarify some common misunderstandings about large models, as well as popular explanations. The two better ones are:

Following instructions is a simple objective function problem: it's like chasing dreams, but it's more mathematical.

You also need a lot of (fine-tuning) data after pre-training: it's like drinking a protein milkshake after fitness.

Another xAI member, Igor Babuschkin, discussed with him why the speed and position of a particle cannot be determined at the same time in quantum mechanics. Grok came up with the most creative explanation for the uncertainty principle:

…… It's like grabbing an oiled pig on roller skates. The more you focus on one hand or foot, the harder it is to control the other.

There are too many AI papers to challenge the new architecture of Transformer. Grok can not only tell the whole story, but also locate the account of a work.

Unfortunately, Grok, like Google Bard, is not very good at Chinese, and has a bit of a translating accent.

The expansion timeline demonstrated in the previous release has the ability to navigate to different versions of the answer, and there is no actual installation.

According to the xAI website, the next focus will be on long text understanding and multimodality.

When it comes to multimodal interaction, people's expectations are high by Google Gemini, and I don't know what surprises Grok can bring.

Which do you prefer, Google Gemini or Musk's Grok?

Reference link:

[1] https://x.com/elonmusk/status/1732769629070655789

[2] https://x.com/DrJimFan/status/1732843493754872024

[3] https://x.com/jimmybajimmyba/status/1732865665894334747

[4] https://x.com/TheGregYang/status/1732105476538564715

[5] https://x.com/ibab_ml/status/1731753207251280235

This article is from the official account of Wechat: quantum bit (ID:QbitAI), author: Mengchen

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.