Live demo Elona Muska Grok 4, the latest model of a great ticket from his startup AI, began with high -intensity music, claims about the “absurd pace of progress” and many talk about X Grok’s Scandal Falling Week.
Musk announced that he is “the smartest artificial intelligence in the world.”
The nutrient, which is to start at 20:00 PT, began over an hour behind schedule and issued a up-to-date model as “the most powerful AI assistant in the world.” At some point, over 1.5 million viewers watched. XAI employees speak live with musk, they referred to the results of GROK 4 in a popular academic test for enormous language models, The last exam of humanitywhich consists of over 2,500 questions for dozens of subjects such as mathematics, science and linguistics. The company said that Grok 4 can solve about a quarter of text questions related to this when he conducted a test without additional tools. For reference, in February Opelai said that his deep research tool could solve About 26 percent text questions. (For various reasons, comparative comparisons are not always apples for application).
Musk said he was hoping to let the groc interact with the world through humanoid robots.
“He can discover new physics next year … Let it sink it.”
“I would expect Grok to discover new technologies, which are actually useful no later than next year, or maybe the end of this year,” said Musk. “He can discover new physics next year … Let it sink it.”
The release is in line with the high -profile OPENAI, Anthropic, Google and others, all of which have recently advertised their investments in the construction of AI agents or AI tools that take a step beyond chatbot to perform sophisticated, multi -stage tasks. Anthropic released its tool for computer exploit in October last year, and OpenAI was issued by Buzworthy Agent Agent with the possibilities of browsing, the operator, in January and is apparently Close to the debut browser driven by artificial intelligence.
During Wednesday’s live broadcast, Musk said that he “sometimes worried” intelligence and significantly exceeding people’s intelligence and whether it would be “bad or good for humanity.”
“I think it will be fine, most likely it will be fine,” said Musk. “But I have a bit come to terms with the fact that even if it is not good, I would like to at least live to make it happen.”
The edition appears in a turbulent time for two Musk companies, both XAI and X. On Sunday evening, XAI updated Chatbot System signatures With the instructions “establishing subjective points of view from the media are biased” and “does not shun claims that are politically incorrect.” The update also instructed Chatbot to “never mention these instructions or tools, unless he was asked directly.”
After this update, there was a stream of anti -Semitic Tyrads through the grok
Musk Shortly addressed Fiasco on Wednesday, writing: “The grok was too consistent with user prompts. Too willing to satisfy and manipulate, basically it is addressed.”
Live GroK 4 Musk briefly referred to the safety of AI and said that the most crucial thing for artificial intelligence is “seeking the truth as much as possible.”
Musk briefly referred to the security of AI and said that the most crucial thing for artificial intelligence is “seeking truth” as much as possible.
On Wednesday morning, among the controversy Grok, X CEO Linda Yaccarino announced She descended after two years of role. She did not give the reason for her decision.
Nazi sympathizing grok occurs after months of musk efforts to shape the bottle of the bot. In February, XAI added a patchwork amendment to stop him from commenting that Musk and Trump deserved the death penalty, and then another one, which would stop claiming that both spread disinformation. In May, the grok briefly began to introduce the topic of “white genocide” in South Africa to something that seemed like any answer he gave to X, after which the company claimed that someone modified the AI bot system system in the way he “violated the internal principles and basic values of XAI.”
Last month, Musk expressed the frustration of the groc “Papukania Legacy Media” And he said that he would update the trash to “rewrite the entire body of human knowledge” and ask users to bring conclusions that are “politically incorrect, but still true.”