Vitalik: Slowing the hyperscale expansion of compute is overall more beneficial than harmful

Odaily Report: Vitalik Buterin posted on X platform that he disagrees with the strong negative reaction to the recent anti-data-center movement, and believes that slowing the hyperscale expansion of compute is overall more beneficial than harmful.

Vitalik pointed out that as algorithms continue to improve, AI capabilities that would require larger-scale data centers to achieve in the next two years may be achievable through existing data centers in about four years or even less. Therefore, even if the anti-data-center movement achieves significant results, it may not necessarily cause catastrophic consequences.

He further stated that the world needs more time to address cybersecurity threats that have already emerged and biosafety and other risks that may arise in the future. At the same time, scientific research is also constrained by many factors beyond AI, so slowing AI development by one year may cause an actual delay to scientific progress that is far less than one year. He emphasized that this judgment does not depend on the controversial AI doomsday risk theory.