Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Google Deepmind CEO and CEO Demis Hasabis gives a conference during the Mobile World Congress (MWC), the largest annual collection in the telecommunications industry in Barcelona on February 26, 2024.
Barrena Pau | AFP | Ghetto images
Paris – the Deepseek AI model “is probably the best job” from China, Demis Hasabis, Google Deepmind CEO on Sunday, but added that the company did not show new scientific achievements.
Last month China Deepseek launched a research book that shook the global markets after saying that its AI model was trained in some of the cost of leading AI players and less advanced Nvidia chips.
Deepseek’s message has caused Aggressive sale of shares And it caused a significant debate about whether large technology companies spend too much on AI infrastructure.
Hasabis praised the Deepseek model as an “impressive job”.
“I think, AI Action Summit This is the host of the city.
Deepmind’s CEO said the AI model shows that Deepseek can make “extremely good engineering” and that “changing things on a geopolitical scale”.
However, from the technology point of view, Hasabis said it was not a big change.
“Despite the overweight, there is no real new scientific progress … He uses known techniques (in AI),” he said, adding that overwhelmingly about Deepseek is “exaggerated a little”.
DeepMind CEO said the company Flash models Gemini 2.0Which Google this week has released to everyone is more effective than the DeepMind model.
Deepseek’s claims around its low price and the chips it uses questioned by expertswho believe that the cost of developing the models of the Chinese company is higher.
AGI at five years old
The AI world has been discussed for years when the arrival of artificial general intelligence or AGI will happen. AGI is widely referred to AI, which is more inherent than people.
Hasabis said the AI industry is “on the way to AGI”, which he describes as a “system that exhibits all the cognitive opportunities that people have.”
“I think we’re close now, you know, maybe we’re just, you know, maybe 5 years or something away from a system like the one that would be quite extraordinary,” Hasabis said.
“And I think society has to prepare for this and what consequences they will have.
Hasabis’ comments reflect those of others in the industry, who suggested that AGI could be closer to reality.
Openai CEO Sam Altman Sam this year said he was “confident that we know how to build AGI, as we traditionally understood it.”
However, many in the industry have also marked many risks associated with AGI. One of the biggest concerns is that people will lose control of the systems they created, the point of view shared by the prominent AI scientists Max Wagmark and Yoshua Belogion, who recently shared their concerns With CNBC over this AI form.
