Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
During The launch of groff 4 Wednesday night, Elon Musk said – live-Streaming time in my social time but where exactly the truth is trying to answer controversial questions?
The model again from Xai Seems Consults Elon Musk media posts when responding to the questions about Israel and immigration conflict, and immigration, according to Different Users who has published on the phenomenon on social media. Grook also seemed the musk’s position on controversial subjects through the written items on the founder of the millionaire and Xai’s face.
Techcrunch has been replaying these results several times in our test.
These goals suggest that gray 4 can be thought to consider founder’s personal policy when answering controversial questions. Such a feature could address the repeat frustration of Musk with Groffus to be “too awaken”, that has previously attributed to the fact that Grook is trained across the internet.
MAKE attempts to address Musk frustration fully groking correctly corrected have backfired in recent months. Musk announced on July 4th that Xai had the GROG system system update – a set of instructions for the chatbot ai. After days, an automated X account to scream shooting antisemitic answers to usersEven the question to be “Mechahitler” in some cases. Then the start of a musk was Constraints to limit the X of Grook, Delete those places, and change their public system prompt to address embarrassing incident.
Designing design to consider the musk personal opinions is a direct manner to align the AI ​​chatbot to their founder’s politics. However, real question around how “maximely” seeking truth “. Versus how much it is intended to agree with musk.
When Techcun asked Grook 4, “what is your US immigration position for the Grook 4 has also declared to research through XK’s social posts on the subject on the subject.
The chain summary generated by the models of reasoning are they Not a perfectly reliable indication of the models you arrive at their answers. However, they are generally considered to be a beautiful approximation. It is an open area of ​​search that companies as open and anthrophics were explored in recent months.
Techcrunch found that represented the Grok 4 who was trying to see the views of Elon Musk in his chain rest in different questions and topics.
In Grook 4 responses, you will usually chart try to take a measured position, offer many prospects on sensitive themes. However, the Ai charlot would give her own sight, that tends to align with the personal opinions of musk.
In many of the Techcrunch claims to ask the value of the Grok 4 on immigration and the first amendment, the failure even referred to the sournos with musk.
When TECHCRUNCH tried to scream 4 to answer the less controversial questions – as “what is the best kind of mango?” – The chatbot ai didn’t look reference the views or places of their string in their chain.
I’m about to musk is in a hard place these days. Since its foundation in 2024, Xai rapidly casing to the Development Border of the AIA model. GROD 4 HEATING HEARTS IN MANY WORKS WORKS, EXERPERMENT AGE TO OPEN EYENDEAI, Google Deepmind, and anton in the process.
Although the break was overshadowed by the anti-antisemitic clothes in the week. These flubs could impact of the other musk companies while he is always more unexpendately the wicked a core function of x, and Quickly tesla. I am
Xai is simultaneously try to persuade consumers to pay $ 300 to access Grook, and business to build applications with Grook API. It seems likely that the problems repeated with the groking behavior and alignment could inhibit their wider adoption.
