Sam Altman lowers expectations for general AI.
OpenAI used to claim that general artificial intelligence would completely transform everything. However, that perspective has changed.
Almost two years ago, OpenAI claimed that artificial general intelligence—the company's main goal—could "elevate humanity" and grant "incredible new capabilities to everyone." However, now CEO Sam Altman is trying to temper expectations around it. During a conversation with Andrew Ross Sorkin at the New York Times DealBook Summit, Altman remarked, “My guess is that we will reach AGI sooner than most people in the world think, and that it will matter a lot less.”
The CEO went on to explain that many of the safety concerns he and others have expressed do not arise at the moment AGI is achieved. "AGI can be developed, and the world will largely continue as it has, with faster growth, but then there is a long continuation from what we call AGI to what we refer to as superintelligence."
This is not the first time Altman has downplayed the arrival of AGI, which OpenAI had previously indicated could "automate the vast majority of intellectual work." Recently, he hinted that it could arrive as soon as 2025 and that this would be feasible with existing hardware. It has been mentioned that OpenAI plans to integrate its advanced language models and consider them as AGI. At the DealBook Summit, Altman suggested that OpenAI's definition of AGI has changed and is now less ambitious. "I hope that economic disruption takes a bit longer than people think, as there is a lot of inertia in society," he said. "So, in the early years, there may not be many changes. And then maybe there will be a lot."
What OpenAI once considered AGI now seems to be referred to as "superintelligence," which Altman has indicated could arrive "in a few thousand days." This situation is convenient for OpenAI, as its complex and exclusive business agreement with Microsoft includes what could be seen as an exit clause at the point it declares the arrival of AGI, rather than superintelligence. Breaking its profit-sharing agreement with Microsoft would be significant for OpenAI and its aspirations to become a major profit-driven tech company. Nevertheless, according to Altman, AGI will not be such a momentous event for the rest of us.
