On Monday local time, Sam Altman, CEO of OpenAI, outlined his vision for AI-driven technological advancements and a future of global prosperity in a new personal blog post titled "The Age of Intelligence." This article paints a picture of artificial intelligence accelerating human progress, with Altman suggesting that superintelligent AI could emerge within the next decade.
Altman stated, "We might see the emergence of superintelligence within a few thousand days; although it could take longer, I am confident in achieving this goal."
He noted that OpenAI's primary goal currently is to create artificial general intelligence (AGI), a technology that can match human intelligence without specialized training. In contrast, superintelligence is a higher level above AGI, capable of surpassing humans in intellectual tasks to an unimaginable extent.
Superintelligence (sometimes referred to as "ASI," or "Artificial Super Intelligence") is a popular yet sometimes fringe topic within the machine learning community. Ilya Sutskever, a former OpenAI co-founder and chief scientist, left OpenAI in June to start a company with "safe superintelligence" in its name. Meanwhile, Altman himself has been discussing the development of superintelligence since at least last year.
So, what exactly is "a few thousand days"? There is no definitive answer. Altman chose a vague number, possibly because he is not entirely sure when ASI will arrive, but it sounds like he thinks it could happen within a decade.
Let's roughly calculate: 2,000 days is about 5.5 years, 3,000 days is about 8.2 years, and 4,000 days is roughly 11 years. In other words, Altman might roughly believe that superintelligent AI could emerge within a decade.
Although Altman's time expectations sound somewhat vague, as the leader of OpenAI, he likely has access to a wealth of cutting-edge AI research information. He acknowledges that not everyone agrees with his optimistic outlook. Computer scientist Grady Booker has questioned his "few thousand days" prediction, considering it mere hype with no basis in reality.
In the article, Altman emphasized the importance of reducing computing costs. He pointed out that to make AI accessible to more people, sufficient infrastructure must be built. He warned that if we do not take action, AI will become a scarce resource, potentially leading to conflicts and serving as a tool for the wealthy.
He also referred to the current era as the "Age of Intelligence," marking another significant technological revolution following the Stone Age, Agricultural Age, and Industrial Age. He believes that the success of deep learning algorithms is the catalyst for this new era. In the future, AI assistants will become increasingly powerful, even capable of forming "personal AI teams" to help people with various tasks.
Despite mentioning the potential labor market impacts of AI, Altman remains optimistic about its overall influence. He stated that while prosperity does not always bring happiness, it will undoubtedly improve the quality of life for people worldwide.
Finally, he concluded the article with an example of an old profession. He mentioned that if a gaslighter from a few centuries ago were to see today's world, they would find the current prosperity unimaginable.
Key Points:
🌐 Superintelligent AI could be achieved within a few thousand days, with OpenAI working tirelessly towards this goal.
🔍 Reducing computing costs and infrastructure development are key to AI普及.
🚀 Altman calls the current era the "Age of Intelligence," promising profound social changes.
