The goal for computer use agents has shifted beyond mimicking human actions to exceeding them in speed. The primary bottleneck is no longer the AI's reasoning but the real-world latency of the software it operates, like website loading times. This changes how developers must think about agent performance optimization.

Unlike cloud-reliant AI, Figure's humanoids perform all computations onboard. This is a critical architectural choice to enable high-frequency (200Hz+) control loops for balance and manipulation, ensuring the robot remains fully functional and responsive without depending on Wi-Fi or 5G connectivity.

While letting a robot 'think' longer improves decision accuracy in lab tests, this added latency poses a significant risk in the real world. If the environment changes during the robot's reasoning period, its final decision may be outdated and dangerous, questioning its practical deployability.

A core challenge in physical AI is the tension between large, powerful models (offboard, in a data center) and the need for low-latency models (onboard, on the machine). The key is using techniques like distillation to create smaller derivatives that run in milliseconds for safety-critical decisions.

Language models execute identically regardless of device, but robotics software is constrained by the physical robot body. Keerthana Gopalakrishnan places current robotics in the 'GPT-2 era' because generic robotic brains cannot reliably control arbitrary, unseen robot embodiments out of the box, nor can they reliably generalize from few-shot demonstrations across hard, diverse tasks.

Separating robotic autonomy into a high-level reasoning model (like Gemini Robotics ER) and a low-level vision-language-action (VLA) policy causes practical bottlenecks. Beyond the communication and processing latency of reasoning models, sequential multi-step tasks suffer from compounded errors where the success rates of both models multiply across every handoff, creating compounding failure points during execution.

Waymo uses a foundation model to create specialized, high-capacity "teacher" models (Driver, Simulator, Critic) offline. These teachers then distill their knowledge into smaller, efficient "student" models that can run in real-time on the vehicle, balancing massive computational power with on-device constraints.

A 2-second delay is acceptable for a single user prompt. However, in an agentic system where 20 agents communicate sequentially, that delay compounds to 40 seconds, rendering the application unusable. This shift necessitates infrastructure with sub-second response times, driving hardware deployment to urban centers.

Cloud AI can batch user requests for efficiency. Physical AI devices like robots operate in real-time on a single stream of data ("batch size one"). This fundamental difference necessitates different, more efficient model architectures that don't rely on aggregation for performance.