Friday, February 14, 2025

Elon Musk Says His New AI Chatbot Is 'Scary Smart'—And Arriving in Weeks


Elon Musk Says His New AI Chatbot Is 'Scary Smart'—And Arriving in Weeks


Elon Musk announced that the next generation of his company's AI chatbot 

Grok may be just weeks away from release, describing it as "scary smart"

 and claiming it had already outperformed every other AI model in testing.

The xAI CEO made these remarks during the World Governments Summit 

in Dubai on February 13.

"At times, I think Grok-3 is kind of scary smart," Musk said. "It comes 

up with solutions that you wouldn't even anticipate—you know, not obvious 

solutions."

The chatbot developers utilized unique training methods for Grok-3. Instead of using 

real-world data like ChatGPT, Grok-3 relied on synthetic data and employed a self-correcting

 mechanism to maintain logical consistency. It got so accurate, Musk claimed, that even when

 it encountered incorrect information, the system reflected on the data and removed content 

that didn't match reality.

The computational demands for training Grok-3 were massive. Experts calculate that it 

required 200 million GPU hours, dwarfing its Chinese competitor DeepSeek-V3's

 2.7 million hours. It ran on xAI's Colossus supercluster with 100,000 Nvidia H100 GPUs

ten times more computing power than its predecessor. Even without fine-tuning, Musk 

claimed the base model performed better than Grok-2.

Grok-3's integration with X, Musk's social media platform, gave it the advantage of being able 

to scrape the social media app in real time instead of relying on browsing the web. The system 

can pull real-time data from X, and features what the company called "Unhinged Mode"— 

which, according to xAI’s own FAQ, is “intended to be objectionable, inappropriate, and 

offensive.”

More....

No comments: