Kevin Lin on X: "🌟Check out our newest work Show-Harness 🕹️Towards Robot-Use for VLM, no calibration or extra VLA training needed. your VLM itself can play robot. 🦾 https://t.co/kn1Dg4Ygf8 https://t.co/s2sSc4beKi"

🌟Check out our newest work Show-Harness 🕹️Towards Robot-Use for VLM, no calibration or extra VLA training needed. your VLM itself can play robot. 🦾 arxiv.org/abs/2609.10522 showlab.github.io/Show-Harness/
[1/8] 🦾 Could a strong robot policy already be hiding inside your VLM, not just GPT Astra, but also Gemini, Claude, even light-weight Qwen and InternVL? 🕹️ We are thrilled to introduce Show-Harness, an Embodied Harness that lets VLMs play real robots—like Claude Code for
00:00