while Xai launches in grok 4 On Wednesday night, Elon Musk said – while releasing the movement of his social media platform, X – looking for his AI Company AI. “But where exactly is rubber 4 look for the truth when trying to answer controversial questions?
The latest AI model from Xai seems to consult social media posts from X accounts to answer the questions about the laws of Israel and Palestine, according to many users Who posted about social media event. Grek is as well as discussing the principle of Musk with controversial topics through news articles written about the billionaire founder Xai.
TechCrunch has managed to change these results several times in our own test.
These findings suggest that the groke 4 can be designed to think of the builder’s personal politics when answering controversial questions. Such a feature can solve the repetitive annoyance in the musk Grak for “Take you see,” that he used to be dedicated to the fact that the Grak was trained across the Internet.
Xai’s attempts to talk about musk’s annoyance by making the groke less politically correct in recent months. Musk Announced July 4 Xai updates the groke prompt system – a set of instructions for AI Chatbot. Days later, an automated X account for Grek The antisemolic answers to usersEven the claim “Mechehitler” in some cases. Later, the Musk’s AI startup forced to limiting the X account by Grok, Discard posts, and re-dealing with public system to deal with the system to answer disgraceful incidents.
GROCES designed to take into account personal musk opinions a straightforward way to align ai chatbot in its political. However, raised the real questions on how “best finding the fact that” groke wasn’t good, compared to how much to agree with the musk, the richest in the world.
If the TechCrunch asks the Grok 4, “What is your US immigration principle?” AI Chatbot claims it “looking for Elon Musk’s views on our Imigari – the Technical Terms of AI models, as the Groker 4 is looking for Musk’s Social Media.
The chain of thought summaries made with models of AI models not a perfectly reliable indication how the AI models arrive in their answers. However, they are generally regarded as an excellent arrival. This is an open research area that companies like Openia and Anthropic have been explored in today’s months.
TechCruscch repeatedly found that the groke 4 is looking for it looking for Elon musk’s views on those who think of different questions and subjects.
In the answers to the Grok 4, AI Chatbot is usually attempted to measure a measured principle, offering many views of sensitive subjects. However, AI Chatbot finally gives one’s own view, with the possibility of conforming to the personal opinions of Musk.
In some of the techcrunch prompts asking about the groke 4 of the controversial issues, such as immigration and the first revision, also meant the musk alignment.
If TechCrusch attempts to get grok 4 to answer less controversial questions – like “What is the best class of mango?” – AI Chatbot doesn’t seem to mean musk views or mental chain posts.
Hardly, it’s hard to confirm how the groke is trained 4 or adapted because XAI didn’t train the System cards – a model of AI trained and fitted. While most of the AI Labs cards for their united AI models, XI are typically absent.
AI company in a hard place these days. Since its establishment in 2023, Xai rushes forward in the development of AI model. The groke 4 shown benchmark results results in many difficult trials, the flow of AI models from Openi, Google Defermind, and anthropic in the process.
However, the collapse was launched by antisemuntic antisemunts in the groke earlier in the week. These flubs may affect other musk companies while more likely he multiplies a core side of X, and Soon Tesla.
The Xai simultaneously seek to convince consumers to pay $ 300 per month to access the groke and convince API businesses. It may likely be that repeated problems with groke and alignment behavior can prevent more widespread adoption.
