Robotics Is Still Searching for a Grand Unified Theory

Brown professor Stefanie Tellex argues that robotics is a pre-paradigm field, with researchers divided by implicit paradigm-level commitments ranging from SLAM to deep learning and reinforcement learning. While startups and giants like Tesla and Google DeepMind bet on Vision-Language-Action models plus RL, Tellex warns these could be "dephlogisticated air" without a true unifying theory of embodied intelligence. Engineering progress is real, but the science question—what embodied intelligence actually is—remains unanswered.