Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The co -founder and executive director of Google Deepmind, Demis Hassabis, offers a conference during the Mobile World Congress (MWC), the largest annual meeting of the telecommunications industry, in Barcelona on February 26, 2024.
Barrena Pau | AFP | Getty images
Paris – Deepseek’s model “is probably the best work” by China, Demis Hassabis, Google Deepmind CEO said on Sunday, but added that the company showed no new scientific advances.
Last month, Depseek from China published a research work that shook global markets after stating that its AI model was trained at a fraction of the cost of the main AI players and with less notice Nvidia French fries.
Deepseek’s announcement caused a liquidation of aggressive actions and caused a considerable debate on whether large technology companies are spending too much on AI infrastructure.
Hassabis praised Deepseek’s model as “an impressive job.”
“I think it’s probably the best work I’ve seen from China,” Hassabis said in an event lodged on Google in Paris before AI Action Summit That is being organized by the city.
The Deepmind CEO said the AI model shows that Depseek can do “extremely good engineering” and that “things change at geopolitical scale.”
However, from the point of view of technology, Hassabis said it was not a great change.
“Despite the exaggeration, there is no new real scientific advance … it is using known techniques (in AI),” he said, he added that the exaggeration of Deepseek has been “exaggerated a bit.”
The Deepmind CEO said the company Flash Gemini 2.0 modelsthat Google threw everyone this week, are more efficient than the Deepmind model.
Deepseek’s statements around its low cost and the chips they use have been questioned by expertsThose who think that the cost of development for the models of the Chinese company is higher.
AGI five years away
The world of AI has been debating for years when the arrival of artificial general intelligence, or AGI will occur. AGI refers widely to AI that is smarter than humans.
Hassabis said that the AI industry is “on the road to AGI”, which he describes as “a system that exhibits all the cognitive abilities that humans have.”
“I think we are close now, you know, maybe we are just, you know, maybe 5 years or something far from a system like the one that would be quite extraordinary,” said Hassabis.
“And I think society needs to prepare for that and what implications will have.”
Hassabis’s comments reflect others in the industry that have suggested that Agi could be closer to reality.
The OpenAi CEO, Sam Altman, said this year that he is “sure we know how to build AGI as we have traditionally understood.”
Even so, many in the industry have also marked multiple risks associated with AGI. One of the biggest concerns is that humans will lose control of the systems they created, an opinion shared by the outstanding scientists of the Max Tegmark and Yoshua Bengio, who recently He shared his concerns with CNBC on this form of AI.
