Grok-3's Joke
xAI researcher @hyhieu226 posted on the X platform, claiming that the latest AI model Grok-3 has "proven the Riemann Hypothesis."
Although it was later clarified that this was just a joke, the news quickly sparked heated discussions online.
(Riemann Zeta Function) and the location of its non-trivial zeros, specifically:
: The real part of all non-trivial zeros is 1/2. 
This conjecture is closely related to the distribution of prime numbers in number theory and is key to mathematicians' exploration of the mysteries of prime numbers. If the conjecture is proven, it will have far-reaching impacts in multiple branches of mathematics, including:
: Providing a precise theoretical foundation for understanding the distribution of prime numbers. : Deepening insights into the behavior of complex functions. : Potentially having significant implications for modern encryption systems based on the properties of prime numbers (such as RSA encryption).
The Riemann Hypothesis is one of the Clay Mathematics Institute's "Millennium Prize Problems," and anyone who can successfully prove it will receive a $1 million prize. If Grok-3 could indeed prove the Riemann Hypothesis, would the $1 million prize be converted to BTC for this AI? However, since the AI proved the Riemann Hypothesis, cryptographic technology might be disrupted, and the value of BTC could also disappear. 🎭
After the post, @hyhieu226 quickly updated with another post:
The book mentioned in the post, I've already asked my idol Teacher Gou for the epub version to read later:
Grok-2 Trial Experience
Here is the response from Grok-2 when asked "Why would it be dangerous if Grok-3 proved the Riemann Hypothesis":
If Grok-3 proved the Riemann Hypothesis, it could lead to:
: Because the Riemann Hypothesis involves the distribution of prime numbers, it could affect existing cryptographic technologies. : Financial markets rely on existing mathematical models, and new mathematical insights could lead to market volatility. : It could accelerate uncontrolled technological developments, such as quantum computing. : An AI solving human problems may raise philosophical questions about human intelligence. : Entities that possess this new knowledge might gain excessive power. 
and the more personalized
: Similar to the standard Grok model, focusing on rigor and practicality. : A unique and entertaining experience with a more engaging and infectious expression style. 
As described by Forbes journalist Cheung: "It's more interesting than ChatGPT's dull responses!"
Fun Mode not only encourages creative thinking but also may bring some unconventional ideas for business growth that ChatGPT wouldn't think of, keeping inspiration flowing. Whether it's problem-solving or idea generation, you can give Grok-2's Fun Mode a try. (I personally think this product entry point is quite good.)
In yesterday's update, Grok-2 will search web pages and content on the X platform as supplementary information, and cite the sources: