Show-Harness: A VLM Agent Alone Can Drive Robots Via Discrete Semantic Actions
showlab · hf · 2026-09-10
ShowLab's Show-Harness links vision-language models to robot control via discrete semantic actions interpreted by embodiment-specific modules, enabling zero-shot deployment and efficient fine-tuned use across robots and GUIs, lowering the barrier for VLM-driven manipulation.
More from Embodied
- Cockroach brain plays Fruit Ninja: tracks, targets and slices fruit mid-air in real time — max_paperclips · 2026-09-10
- GPT-6 Turns Two Raw Videos Into Real Robot Dexterous Rollouts, From Rubik's Cube to Egocentric Hands — TheMoonMidas · 2026-09-10
- Claude drives Asimov humanoid robot via new Python SDK in first demo — freelerobot · 2026-09-10
- First iMessages sent from non-Apple glasses, built on iOS Shortcuts automation — AnhPhuNguyen1 · 2026-09-10
- Apple's $1,999 folding iPhone: A20 Pro chip with second neural engine drives the price — BenBajarin · 2026-09-10
- $2,000 AI phone criticized for weaker camera and no Face ID — firstadopter · 2026-09-10