Last year, CHATGPT shocked the world, but this year it was replaced by Sora. They all originate from OpenAI, the open artificial intelligence research center in the United States, and they both have advanced technology that is good enough to pass for the real thing. Last year, AI learned human language, and this year, it learned how to understand what humans see. Some experts predict that artificial general intelligence (AGI) can be achieved within 2-3 years.
While some companies are worried about not being able to buy advanced gpus, top AI companies are worried about power: in 2018, computers used less than 2 percent of the world's electricity, according to experts at the University of Pennsylvania But today that share has risen to 10 per cent, and by 2030 computers are expected to consume a fifth of the world's electricity. These predictions point to a common conclusion: in the ultimate phase of artificial intelligence, energy independence is key. Therefore, the preemptive layout of the energy sector, for the grasp of the future development of science and technology is crucial. The limit of artificial intelligence lies in energy! The limits of computing power lie in electricity, including photovoltaic, energy storage and nuclear fusion. In the absence of major advances in energy, the development of artificial intelligence will not go beyond the theoretical and technical levels. In the future, computing and energy will be the two key “Currencies” that can be converted into each other, says Sam Altmann. In this context, the deep integration of artificial intelligence with photovoltaic and energy storage will become the new chapter of the next mainstream technology story.