Watch: Google DeepMind CEO and AI Nobel winner Demis Hassabis on CBS’ ’60 Minutes’

A segment in a weekly deep TV news program in CBS 60 minutes last night (Here also shared on YouTube) Google has offered a view that includes vision of Doepmind and its co-founder and Nobel Prize CEO, the legendary AI researcher demis Hassabis.

Interview Deepmind’s aggression to achieve rapid progress and artificial intelligence (AGI) in artificial intelligence – to achieve a manual and superhuman scale with a man.

Hassabis described today’s AI trajectory as the “improvement of the improvement”, the growing interest, talent and the field was gas.

Before the previous one two years later 60 minutes Interview Heralded Chatbot Era, Hassabis and Deepmind, now the language is not understanding, but also the language is also followed by more skillful systems designed to develop the physical world around.

The interview came Google’s Cloud next 2025 conference is at the beginning of this monthThe search giant presents a centralized features around the new AI models and its twins 2.5 Multimodal AI model family. Google left that conference, a visible leadership Compared to other technological companies to ensure a powerful AI for the enterprise, it uses the cases of use in the most affordable price points that exceed Openai.

More information about Google Deepmind’s ‘Project Astra’

One of the focus points of the segment, Astra, Deepmind was the next generation chatBot project. Astra is designed to interpret the visual world in real time.

In a demo, identified paintings, evaluated emotional situations and created a story around a hopper painting with the line: “Only moving over the flow of thought.”

When asked if he was bored, Astra responded thoughtfully, found the degree of sensitivity to tons and interpersonal nuances.

Product Manager Bibbo shu Astra’s unique design: “An AI that can see about everything, hear and chat

Gemini: to the right AI

The broadcast was also shown TwinsThe DeepMind AI system teaches you to comment on the world, not to interpret the world, but also online bookings and shopping online.

Hassabis said that Gemini is a step towards AGI: as a human being as a human being, manage and operate in complex environments.

This 60 minutes The team tested a prototype in glasses by demonstrating visual recognition and voice responses in real time. Also can pionering be in the return of the upcoming return, but eventually turn off the early expanded reality glasses known as Google Glass? Debuted in 2012 previous Retire in 2015?

Special Gemini Model versions Gemini 2.5 Pro or Flash The segment was not celebrated, recently Google’s wider ecosystem could recently reflect these models for the use of enterprises, parallel development efforts.

This integration supports the growing ambitions in the EU, which is applied in the AI, whether they fall outside the directly covered in the interview.

AGI Soon since 2030?

When a schedule asked, Hassabis can come as soon as possible with the 2030s, the systems that understand their environments in the best and deep-way systems.

The interview also solved the possibility of self-identification in the EU. Hassabis said that the current systems are not conscious, but future models can demonstrate signs of self-concept. Again, emphasized philosophical and biological division: Although machines imitate conscious behavior, are made of the same “Squishy Carbon Matter” as people.

Hassabis also predicted great developments in the robotics, the neurts could come in close years. The segment presented robots that determine the green block that forms a green block that occurs by mixing the abilities of yellow and blue-blue-blue rising justifications in physical systems.

Good luck and security concern

The segment, the AI ​​model, which predicts the structure of more than 200 million proteins, has been revised to Deepmind’s favorable achievement.

Hassabis and colleague John Jumper, awarded 2024 Nobel Prize in Chemistry for this work. Hassabis stressed that this progress can accelerate the development of drugs, reduces time for up to a few weeks. “I think a day maybe we can treat the whole disease with the help of AI,” he said.

Despite optimistic, Hassabis was clearly concerned. He referred to the two great risk: the abuse of the growing autonomy of the AI’s evil actors and people out of control. As he could teach a child, stressed the importance of the building, which is teaching Karovdrail and value systems. He noted that the EU’s influence will touch each country and culture and called international cooperation.

“One of my great concerns,” he said, “The race for the AI ​​dominance can be racing to the bottom for safety.” He stressed the need for leading players and nations to coordinate the ethical development and control of leading players and nations.

The segment ended with a meditation in the future: AI tools are thinking about a world and eventually knowledge, consciousness and even the meaning of life, which can change their efforts. As Hassabis puts, “New great philosophies to understand the effects of this system …