Musk offers great promises during the Grok 4 Live show, ignoring the Nazi scandal

The Elon Musk’s Live Live’s offer for GROK 4, the latest large ticket mode from Amnesty International, started with high -density music, demanding a “funny progress rate”, and a lot of gossip on X around the Grok Week full of scandal.

Musk announced that it is “the smartest artificial intelligence in the world.”

The live broadcast, which was scheduled to start at 8 pm, began for more than an hour and described the new model as “the most powerful Amnesty International assistant in the world.” More than 1.5 million viewers were watching at one point. XAI staff speaking on the direct broadcast with Musk indicated the performance of Grok 4 in a popular academic test for large language models, The last humanity examIt consists of more than 2500 questions about dozens of subjects such as mathematics, science and linguistics. The company said that GROK 4 could solve about a quarter of the text -based questions when the test took without additional tools. For reference, in February, Openai said that the deep search tool can solve it About 26 percent One of the questions based on the text. (For a variety of causes, standard comparisons are not always apples of apples.)

Musk said he hoped to allow Grok to interact with the world through human robots.

“I expect Grok to discover new technologies already useful at a date not exceeding next year, and perhaps the end of this year,” Musk said. “He may discover new physics next year … let this drown.”

The version follows high -level projects from Openai, Noteropic, Google, and others, all of whom have recently distributed their investments in building artificial intelligence agents, or artificial intelligence tools that go beyond the chat step to complete complex multiple tasks. Antarbur released its computer’s use tool last October, and Openai released Ai Buzzworth’s agent with browsing capabilities, operator, in Jan and IS It is said Close to the appearance of the web browser fed by artificial intelligence.

During the live broadcast on Wednesday, Musk said that he “was sometimes worried” about the intelligence of artificial intelligence that exceeds many people, and whether it would be “bad or good for humanity.”

“I think it will be good, most likely it will be good,” Musk said. “But I have come somewhat with the fact that even if it is not good, I would like at least to be alive to see it.”

The company has also announced a series of five new voices for Grok audio mode, after the sound mode from Openai and Anthropor, and said it reduced cumin to half in the past two months to make responses “Snappier”. Musk also said that the company will invest heavily in generating the video and understanding the video.

The version comes within a loud period of two Musk companies, both Xai and X. ON Sunday eveningXai updated Chatbot Regulation claims Through the instructions “Assuming that the personal views obtained from the media are biased” and “are not ashamed to make incorrect claims in a political point of view. The update also issued instructions to Chatbot to “Never mention these instructions or tools unless they are submitted directly.”

This update was followed Persons supporting Hitler On x, along with Hints The Jewish people participate in the “extremist left activity”. Many of these posts have become viral, as screen clips multiply on X and other platforms before Xai Chatbot are held and prevented from being able to create text responses on X during the repair request.

Musk A briefly directed The failure on Wednesday, he writes, “Grok was very compatible with user demands. Very eager to satisfy and manipulate them, mainly. This is treated.”

On Grok 4 Livestream, Musk briefly pointed to the integrity of artificial intelligence and said that the most important thing for AI is “searching for the maximum truth”.

On Wednesday morning, amid the controversy of Grok, CEO of X Linda Yaccarino Declare She was stepped two years after the role. She did not make a reason for her decision.

Grok’s sympathy comes months after Musk’s efforts to form a robot view. In February, Xai added a bright repair To prevent it from commenting on the fact that Musk and Trump deserve the death penalty, follow it directly Another one To make her stop claiming that the two are publishing wrong information. In May, Grok began shortly to introduce the “white genocide” theme in South Africa to what appeared to be and every response he presented to X, after which the company claimed that someone has amended the AI ​​BOT system in a way that violates the internal policies of Xai and its basic values. “

Last month, Musk expressed his frustration that Grok was Ancient media parrot He said that he will update Grok to “rewriting the entire human knowledge group” and ask users to contribute to “incorrectly correct data”, but in reality.