Buterin Tests Private AI Setup for Personalized Health Advice

Vitalik Buterin, the co-founder of Ethereum, has been experimenting with a private AI setup that generates personalized diet and exercise recommendations while limiting personal information sent to remote models.

The setup uses a local model, Qwen3.8-Flash-Next, zkAPI, and Tor to handle personal data locally and protect user identity. Buterin noted that Tor latency remains a problem, with request-by-request unlinking being inefficient in current tests.

The local model is intended to reduce content exposure, and it instructs the model when to use a remote system and how to construct a request that contains less identifying information. The remote model receives only the portion selected for a particular task, while personal health and travel records remain available to the local system.

Buterin's experiment uses three layers of protection: the content of requests, payment information, and internet traffic. He believes that hiding payment information alone does not prevent an AI provider from learning details through prompt content or network metadata.

The setup has produced diet and exercise recommendations, and information returned by frontier models has improved the results. However, Buterin acknowledged that the request-writing rules in his current experiment still need improvement, as removing more personal context can reduce the usefulness of remote models.