Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The General Manager of Google Deepmind and Demis Hassabis provides a talk at the Mobile World Congress (MWC), the largest annual meeting of the telecommunications industry in Barcelona on 26 February 2024.
Barrena Pau | AFP | Getty Images
Paris – Deepseeeek’s model “It is said in Demis Hassabis, Google Deepminind, he said on Sunday, but has added that the company had no new scientific progress.
Last month, the Chinese Deepse released a research paper after claiming the AI model, he was trained in a division of the cost of the main Aiko player and was less advanced Nvidia chips.
Deepseeek’s announcement turned on To sell aggressive stock And he sparked great discussion when large technology companies are spending too much in AI infrastructure.
Hassabis, dependingeek model “spectacular work”.
“I think I’ve probably seen the best work I’ve seen from China,” Hassabis said in a Parisian Google event before AI Action Peak It is being hosted by cities.
The Director of DeepMind says that the AI model shows that Deepsees can do “very good engineering” and “changes things in a geopolitical scale.”
However, from a technology point of view, Hassabis said it was not a great change.
“Although hype, there is no new scientific advances … Knows the well-known technique (Ain),” he said, deepened the depths of the depth.
Deepmind CEO said that the company Gemini 2.0 flash modelsGoogle released everyone this week, more effective than Deepmind’s model.
Deepsees have been claims about its low cost and chips used Asked by expertsThose who think the development cost of China’s models is higher.
Agi five years away
AI The world has been discussed in the years when the arrival of general intelligence or Agi will occur. Agik refers to AI that is more intelligent than humans.
Hassabis said that the AI industry is “agi” that describes as a system that shows all the systems that contain all the cognitive abilities of humans “.
“I think we’re now close, you know, maybe we’re alone, perhaps away from a system like 5 years or something extraordinary,” Hasshabis said.
“And I think society must be ready for this and what they will be. And you know that we will come from that and that whole society is beneficial…
Hassabis’s comments Mirror reflects the mirror of others who could approach agias to reality.
This year’s Openai CEO Sam Altman said, “Because we know that we know that agi is trusted”.
However, many in the industry has also made many risks related to documents. One of the greatest concerns is that humans will lose control of systems created by AI, the most important scientists Max Tegmark and Yoshua Bengio, recently shared view they shared their concerns With CNBC ai in this form.
