Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Demis Hassabis, CEO of Google Deepmind, could have won a Nobel Prize For his work on Alphafold 2, an AI model that can predict protein structures, but the solution to the problem he really wants to solve the escape.
The problem in question is infinitely easier to grasp and more relatable than Hassabis’ work in the field of chemistry. “What I really want to work is a new generation email,” he said, speaking to Sxsw London on Monday. “I would like to get rid of my email.”
On the basis of the crowd’s reaction, it was a popular feeling in the room, where earlier, former British Prime Minister Tony Blair admitted that he had sent a single e-mail for the entire period of 10 years that he was in office.
There is a certain irony in the quest for Hassabis. The award -winning scientist is responsible for developing some of the most complex and sophisticated AI Models The world has never seen, all this to work on remedies for diseases beyond everything we have access today. Its mission to make emails (probably Gmail?) – A discomfort of our own human invention – obsolete looks like small fry in comparison.
But this also exposes the duality of Hassabis’s responsibilities at Google. He is and has always been deeply committed to pursuing AI for the benefit of humanity. “My personal passion is to apply [AI] At the borders of science and medicine, “he said. At the same time, he is indebted to the interests of Google companies, which acquired Deepmind in 2014.
Hassabis has always imagined that the development of AI is more a “business managed by scientist”, led by computer equivalent to CERN, the Fénéré laboratory of particles physics in Switzerland. But technology has become different, becoming commercially viable much faster than expected. From there, he said, “The capitalist engine did what it does best.”
Hassabis speaks almost as if he were separated from the “capitalist engine”, but of course, he is deeply rooted. Deepmind belonging to Google means that as well as the continuation of its project to heal the disease with AI – undoubtedly the noblest use of AI – He must divide his attention to ensure the products of Google AI, Gemini has I see And everything else The company announced to E / S last month are up to a competitive market.
In pursuit of AG
The competition is “fierce” and it is a heavy work schedule for a man, who says he sleeps very little and does not expect “we arrive at AG” or artificial general intelligence. In addition to developing the basic models of Deepmind AI and translating them into science, it continues to continue the development of AGA or AI which fully (or exceeds) human intellectual capacities. “My feeling is that we are about five to 10 years old,” he said.
Her vision of the act is that she will unlock a world in which “we can cure many diseases – or perhaps even all diseases” and “unlimited renewable energies”. In some ways, Google products stop the points on the way.
One of the reasons why Deepmind has built Veo 3, his latest video generation software, said Hassabis is that AGI must have a physical understanding of the world around him. The global models built for Veo 3 are essential for this understanding. In turn, these world models will be essential for a robotics breakthrough, which, according to Hassabis, is due in the “next years”.
Although it is sometimes not clear where the mission worthy of Deepmind ends and that the commercial priorities of Google come into play, it is clear that Hassabis finds ways to make it work for it and its long -term prosecution of an act.
Despite the seismic change, he predicts that this will cause, even he is skeptical about the media threshing around the AI in the short term. “I mean, it couldn’t be more excited,” he said. “Therefore, it’s a bit over-type.”
