Artificial Intelligence thread

Bellum_Romanum

Bellum_Romanum

Brigadier
Registered Member
KYli said:
Please, Log in or Register to view URLs content!
Click to expand...
Please, Log in or Register to view URLs content!

A pretty impressive accomplishment for a company that was just started back in March of this year. I wish Kai-Fu-Lee the best of luck on this endeavor. I find it quite interesting as well that the A.I. LLM listed on this chart are all Chinese companies with the exception of Llama (Meta) and Falcon (TII of Abu Dhabi)

1699234882896.png
 
W

Wuhun

New Member
Registered Member
Wuhun said:
Deepseek AI, a newly established Chinese AI start-up, has launched their own Code Language Model with with superior performance than Meta.

View attachment 120960
Click to expand...

This is a serious start-up. As you've heard that xAI (elon musk company) has released their own foundation model yesterday. However, this small Chinese company with limited resources, literally built by PhD students, simply outperforms xAI. Remember that three of the co-founders of xAI are Chinese, who may returns back to China in the distant future as all of them are in their late 20s, and AGI is 20+ years away..

On HumanEval, xAI 33B = 63.2
DeepSeek 33B = 79.3

On Math, xAI 33B = 23.9
DeepSeek 33B = 35.3
 
Last edited:
O

Overbom

Brigadier
Registered Member
KYli said:
Please, Log in or Register to view URLs content!
Click to expand...
I know that people joke about a lot about such startups, but this one specifically is a serious company that people should pay attention IMO

Just imagine, they seemingly have state of the art LLM and they only started operations in June...

Also very smart to release open source models with reduced parameters so that they can tap into the open source AI community. Llama has tremendously benefitted from that
The size of the just-launched AI system, 34 billion parameters, was carefully chosen so that it can run on computers that aren’t prohibitively expensive. The company is also releasing a 6B model to appeal to a broader swath of developers. “It’s a highly calculated decision,” Lee said. “The world doesn’t need another arbitrary model, the world needs us.”
Click to expand...
 
B

BlackWindMnt

Captain
Registered Member
Overbom said:
I know that people joke about a lot about such startups, but this one specifically is a serious company that people should pay attention IMO

Just imagine, they seemingly have state of the art LLM and they only started operations in June...

Also very smart to release open source models with reduced parameters so that they can tap into the open source AI community. Llama has tremendously benefitted from that
Click to expand...
Its only because of the name Kai-Fu Lee that i think interesting! He gave some really good insight into the competitive nature of the Chinese big tech market in his book.
 
You must log in or register to reply here.
Top