Vitalik Buterin, a co-founder of Taifeng, recently spoke of the development of artificial intelligence, saying that AI is gradually taking over more tasks than humans in a way that is not necessarily noticed. In his view, it was not accurate that a single standard had been used in the past to measure the gap between human beings and AI.
Not just smart.
Buterin said that, judging by AI progress, it cannot be asked whether it is “smarter” than humans, nor can it be based solely on economic output. In his view, it would be more appropriate to look at how many types of tasks AI can do and what the capabilities cover.
He cited, for example, the diversity of people in different fields. The same goes for AI. It may be stronger in certain cognitive tasks, but there are still differences in physical, emotional understanding or complex judgement.
From a few areas of expertise to more.
Buterin also mentioned the machine age around 1500 years. At that time, water and windmills had outperformed humanity in only a few specialized tasks, but human beings still prevailed in most areas. He thinks that today's AI is following a similar path, except for a faster expansion and wider coverage.
He stated that AI was not a one-time substitute for human beings, but was entering new areas and approaching and even surpassing human beings in one capacity after another.
- Buterin believes that the progress of AI should be measured by mission coverage.
- He emphasized that a single “smart” indicator was not sufficient to describe AI changes.
- He compared the current AI expansion to the historical evolution of machine capabilities.
Progress measured by number of tasks
In Buterin, it seems that the key to AI's evolution is not whether it's more comprehensive than humans, but whether it can continue to enter more scenes where humans are best suited. This scope will continue to expand as model capacity increases.