Local AI Can Protect Privacy Without Sacrificing Speed, Says Ethereum Co-Founder

Ethereum co-founder Vitalik Buterin claims that local AI can balance speed and privacy in wallet software. In a recent post, he mentioned that laptop AI is approaching a practical turning point, citing improvements in Qwen 3.8 Flash and llama.cpp as examples.

Buterin explained that a local model can handle a large share of tasks on his Strix Halo laptop without compromising speed or losing responsiveness. He demonstrated this by sharing benchmark images showing input-processing rates ranging from 109.82 to 373.22 tokens per second, and output generation ranging from 18.42 to 33.37 tokens per second.

However, he emphasized that model judgment, resistance to malicious instructions, and transaction authorization remain unresolved issues. Nevertheless, local inference can improve privacy while keeping the power to move funds behind separate, enforceable controls.