Woofun AI reports that Zhipu founder Tang Jie unveiled the "Touch High" initiative via an internal letter, reaffirming a commitment to AGI research over immediate commercial gains. The strategy targets four critical challenges: Long Horizon Task, Autonomous Agent System, Fully Self Training, and Extreme Safety Governance.

Zhipu intends to deploy tens of billions in resources toward mechanistic interpretability. This effort aims to decode the neural logic behind model decisions, facilitating a shift from opaque black-box systems to transparent, interpretable architectures.