Bankless
AI, Surveillance, and the Fight for Digital Sovereignty | Near's Illia Polosukhin
- AI export controls & internet precedent: The US government's ban on Anthropic's Claude 3.5 Fable sets a dangerous precedent for restricting internet services by country, potentially fracturing global access to critical infrastructure and technology.
- Nationalization risk for AI labs: As governments recognize AI's power (comparable to nuclear weapons), they will likely move to nationalize leading AI companies to maintain control, forcing KYC requirements and full data surveillance on users.
- Decentralized AI as the countermeasure: Near is building a sovereign AI stack—confidential inference, Ironclaw agents, and private smart contracts—allowing users to access powerful models without KYC, data leakage, or government surveillance.
- User-owned alignment vs. corporate alignment: General alignment is a myth; real alignment means systems work exclusively for the individual user or specific company, not broad society. Blockchain tokenization solves OpenAI's capital formation problem and ensures user-company alignment.
- Agent marketplace & confidential compute: Specialized agents can be hired on-chain with verifiable privacy guarantees, enabling businesses to delegate sensitive work (CRM access, email, financial data) without data breach risk.
- Compute as the real bottleneck: Hyperscalers control most GPU supply through 2026, but cost arbitrage and model routing make open-weight alternatives economically viable for most enterprises, especially as Claude costs spiral.