OpenAI co-founder and CTO Greg Brockman shared his predictions on when artificial intelligence will be able to reach the level of human intelligence (AGI).
According to him, this could happen as soon as within the next three years, tentatively before 2028. Brockman added that if AGI does not appear before 2030, it will be a signal that “something has gone wrong.” He specified that AGI means systems capable of doing economically significant work as well as a human being.
It’s worth noting that Brockman’s views on AGI have changed over time. Earlier he perceived AGI as a final goal, but now he sees it as a continuous process in which AI is gradually approaching the human level in various spheres.
“There is a point where AI will be able to compete with humans in valuable work. This is an important stage, but not the end of the road,” noted Greg Brockman.
Brockman cited a shortage of computing resources as one of the biggest challenges. He admitted that the distribution of the graphics processor within the company has become a real problem: there are a lot of ideas, but not enough capacity. According to him, this very factor is now considered a weak point in the development of new models and can become the main brake on the way to general intelligence.