On the Limits of Direct LLM Control for Humanoids
philfung · x · 2026-07-15
A recent robotics study from Anthropic has sparked discussion. The researchers point out that directly using Large Language Models (LLMs) to control robots is not the optimal solution.
Core takeaways include:
- Lack of speed and physical common sense: LLM response times are about two orders of magnitude slower than real-time control requirements. Furthermore, they lack a deep understanding of embodiment physics, making it difficult to directly compute continuous motor commands for multi-degree-of-freedom robots.
- The correct architectural approach: Robot AI should rely on narrow AI models specifically trained for different control layers to execute low-level operations. LLMs are better suited to act as the top-level "brain," handling human interaction and high-level task planning, rather than directly taking over low-level motor control.
Related event: Anthropic Research Suggests LLMs Are Not Ideal for Direct Robot Control(3 posts)→
More from Embodied
- StackChan robot performs an AI-generated skit, orchestrated by Hermes Agent via JSON script — alexcovo_eth · 2026-10-27
- DONIC hits 86.79% on bimanual dexterous tracking, crushing prior SOTA's 9.43% — liuziwei7 · 2026-10-27
- Bedrock Robotics raises $600M as its fully autonomous excavators do paid work on live job sites — Sethwinterroth · 2026-10-27
- AW-0 hits 80.8% on RoboTwin tasks, beating strongest baseline by 16 points — trevordarrell · 2026-10-27
- AW-0's code stream fixes world model prediction: picks mango where baseline grabs apple — trevordarrell · 2026-10-27
- AW-0 runs at 230ms on a single RTX 5090, 2.4x faster than autoregressive code planning — trevordarrell · 2026-10-27