Discusses the benefits of a Jev-powered assistant anticipating low-risk actions and the need for confirmation in high-risk scenarios like messages, payments, or deleted files, prompting reflection on AI safety boundaries.
A Mac assistant powered by Jev can open an app before the user finishes speaking. That feels natural for low-risk actions, but dangerous for messages, payments or deleted files. Should AI anticipate harmless commands but require confirmation for anything irreversible?
Gregor Zunic showcases an open-source browser agent combining Browser Use and Jev, finding flights in 7 seconds for $0.0039, using a new action space per step, DOM state space, and small LLM fallback for typing.
Scott Williams predicts custom Jev-style models (parallel constrained decoding) will make existing company agent systems more token efficient, creating millionaires.
The user praises the JEV model for its extremely fast reproduction of each movement and command, believing bots will become smarter as a result.
