WuBlockchain
Vitalik Buterin says he disagrees with the strong negative reaction to the anti-data-center movement, arguing that slowing the hyperscale expansion of AI compute may do more good than harm.
His case rests on algorithmic efficiency: capabilities that would otherwise require much larger data centers in two years could be achieved within roughly four years on existing infrastructure. A slower buildout would also give society more time to work on cybersecurity, biosecurity and other risks.
He adds that scientific progress is constrained by factors beyond AI, so a one-year slowdown in AI compute would not necessarily translate into an equal delay in research.
Frequently asked questions
-
Why does Vitalik Buterin support slowing AI data center expansion?
He argues that gains in algorithmic efficiency could let AI achieve in roughly four years capabilities that would otherwise require much larger data centers in two years, buying society time to address safety risks.
-
What risks does Vitalik say a compute slowdown would help address?
He cites cybersecurity and biosecurity, along with other AI-related risks, as areas where extra time from a slower buildout would be valuable.
-
Does Vitalik think slowing AI compute would delay scientific research?
No. He argues scientific progress is constrained by factors beyond AI, so a one-year slowdown in AI compute would not necessarily translate into an equal delay in research.
-
What is the anti-data-center movement Vitalik is responding to?
It is a pushback against hyperscale expansion of AI compute infrastructure. Vitalik says he disagrees with the strong negative reaction the movement has received.
-
How much longer could algorithmic efficiency stretch AI capability timelines?
According to Vitalik, capabilities expected to require larger data centers in two years could instead be achieved within roughly four years using existing infrastructure.
CryptoSlate
Crypto News
CoinTelegraph
WatcherGuru
CoinDesk