Gemini task automation is slow, clunky, and super impressive

Editor's take

Gemini's nascent ability to autonomously navigate and interact with mobile applications marks a significant, albeit early, step in on-device AI agents.

This development is crucial as it represents a tangible shift from AI as a passive information provider to an active participant in user workflows, potentially impacting how millions interact with their smartphones for routine tasks like ordering food or booking rides. The initial limitations to a few specific services highlight the current challenges in achieving broad app compatibility and robust error handling, a familiar hurdle for AI automation.

Future progress will hinge on Gemini's ability to expand its repertoire of controllable applications and improve its reliability in executing multi-step processes without user intervention. Observing Google's speed in patching the current clunkiness and the success of competitors like OpenAI's ChatGPT in developing similar agent capabilities will be key indicators of this technology's trajectory.