Have we reached Peak Human?
I’m not talking about peak population or peak lifespan or peak performance. When I refer to Peak Human, I’m talking about whether we have reached the point where AI systems can outthink the majority of individual humans, thus defining the moment in time when we “peaked” as an intellectual force on planet earth.
After all, once we pass this milestone, we will steadily lose our cognitive edge until AI systems can outthink all individual humans — even the most brilliant among us. At that later point, we will say that AI has achieved the more dangerous milestone of Superintelligence. And while most people focus on the second milestone, I believe we should track the first because it could happen within months, not years.
Until recently, the average human could easily outperform even the most powerful AI systems when it comes to basic reasoning. There are many ways to measure reasoning, none-of-which are considered the gold standard, but the best known is the classic IQ test. A journalist named Maxim Lott has been testing Large Language Models on a standardized Mensa IQ test. Last week, for the first time, an AI model significantly beat the median human IQ of 100. The model that crossed the peak of the bell curve was Open AI’s new “o1” system — it reportedly scored 120 IQ.
Not so fast. It is not valid to administer a standard IQ test to an AI system because the data it trained on likely included the tests and answers. To address this, Lott had a custom IQ test created that does not appear anywhere online and therefore is not in the training data. He gave that “offline test” to Open AI’s new “o1” model and it scored an IQ of 95.
.jpg)
