The advances in AI have not only surprised most people, but also surprised scholars and AI researchers.
Max Tegmark, a professor at MIT, says that even his colleagues were completely wrong about their predictions about AI advancement. Tegmark is a physicist, cosmologist and machine learning researcher, known for his pioneering work on ideas about the nature of reality and the nature of artificial intelligence. He says that the scientific community, even a few years ago, had no idea how fast AI would develop.
“I remember most of my colleagues here at MIT very vividly, like six or seven years ago. In general, most of my AI colleagues were pretty certain that we had been decades away from the construction of machines that allowed us to acquire language and knowledge on a human level.
“They were too pessimistic because (these things) were already happening. You can quibble whether it happened with GPT-4 (by passing the Turing test) or whether it was accurate, but if people could be very wrong about it about it, maybe they were even more wrong. In a short year, after-predicted predictions were crushed and things happened faster,” he added.
“I think we've come from an exaggerated administration to an unexaggerated administration, and this is why so many people get to a broad human level in two to five years, depending on the CEO of the technology you're talking about.
Humanity has been blazing past the Turing test, once considered one of the most challenging problems in computer science. Openai CEO Sam Altman says future historians cannot agree when the Turing test passes accurately. There were also other predictions that less than 10% of experts who believed that AI would win gold medals in the International Mathematics Olympiad by 2025, both Google and Openai already achieved the feat. And if AI can continue to advance at this rate, we don't know where it will take years.
