Vitalik Buterin suggested evaluating AI by the number of successfully completed tasks
7/25/2026, 06:00 AM • Евгения Слив

Vitalik Buterin, co-founder of the Ethereum network, has proposed a new approach to evaluating the capabilities of artificial intelligence. The programmer argues that the effectiveness of machines should be measured by the growing number of successfully completed tasks, rather than by direct comparison with human intelligence. In his opinion, using a single standard to measure the mental abilities and economic productivity of people and algorithms is incorrect. This method overlooks the diversity of skills because it does not take into account the unique strengths of each party. The developer insists on the need to evaluate technologies based on their ability for physical activity, manual labor, strategic thinking, empathy and mathematical calculations.
To better illustrate technological progress, Buterin gave an example from the history of human development. In 1500, water and windmills surpassed humans solely in the efficiency of converting natural energy. In all other activities, human skills were significantly ahead of artificially created inventions. In the modern world, the development of neural networks follows a similar path. Artificial intelligence is not able to quickly and completely replace humans in all areas, although it gradually covers an increasingly wide range of physical, analytical and creative tasks, expanding the boundaries of its capabilities.
Recently, the co-founder of Ethereum has increasingly addressed the topic of artificial intelligence in his publications. He recently admitted that this technology can significantly simplify the search for vulnerabilities in smart contracts and automate the writing of software code, which in the future will increase the overall security of the blockchain and decentralized applications. In addition, in June, Buterin launched an experiment among his subscribers, offering to find an anonymous document written by him using algorithms. The winner of the competition was analyst Franklin Wang, who calculated authorship not by stylistic features of the text, but by unique patterns of thinking.
