We are gonna have genuinely useful long running agents in 2026. Somehow, this feels both entirely predictable, yet scarily futuristic

Jerry Liu predicts that genuinely useful long-running AI agents will become a reality by 2026, viewing this development as simultaneously inevitable and remarkable. The post reflects on how advances in AI capabilities will enable agents that can operate autonomously over extended periods while maintaining practical utility. This represents a significant milestone in AI development, moving beyond current limitations of context windows and task execution capabilities.

Source: Jerry Liu (X) | 2026-05-02