NUS Showlab Unveils Show-Harness Interface for VLM Agents to Control Real Robots
Researchers at NUS Showlab introduced Show-Harness, an interface that enables vision-language model (VLM) agents to control real robots directly, moving beyond computer-use to robot-use. The project releases open-source code, models, and data, targeting the gap between frontier models' intelligence and physical robot execution.
Coverage timeline
机器之心机器之心
标题：Show-Harness: Just a VLM Agent Can Play Robots 作者：Yanzhe Chen*, Zechen Bai*, Zhijun Cao*, Wenzheng Zeng*, Kevin Qinghong Lin, Yiqi Lin, Guoqiang Liang, Kevin Yuchen Ma, Qiming Huang, Mike Zheng Shou† 单位：新加坡国立大学 Showlab 项目主页：https://showlab.github.io/Show-Harness/ 论文链接：https://arxiv.org/abs/2609.10522 开源代码：https://github.com/showlab/Show-Harness 开源模型：https://huggingface.co/showlab/Show-Harness-VLMs 开源数据：https://huggingface.co/datasets/showlab/Show-Harness-Data 今天的 Agent 已经能看懂屏幕、理解任务，并通过 click、type、scroll 等鼠标键盘操作来完成数字世界的任务，那么它距离操控真实世界里的机器人，还差什么？ 过去一段时间，越来越多工作开始尝试让 GPT、Claude、Gemini 等 frontier models 参与真实机器人控制。这背后有一个很直接的判断：当我们讨论具身智能时，常常强调 “具身”，但核心仍然是 “智能”。而今天最强的智能，包括视觉理解、空间推理和任务分解等能力，都逐渐集中到 frontier foundation models 中。 问题在于，这些能力并不会自然的变成机器人动作。过去的研究尝试了不同的路线去利用 foundation model。传统的 VLA 往往让模型基于图像和任务指令直接回归连续控制量；另一类系统则让大模型只负责规划，再由独立的控制器完成物理执行。前者需要为不同机器人和数据分布反复适配，后者又拉开了语义决策与实际动作之间的距离。 新加坡国立大学 Show Lab 的研究者提出 Show-Harness 。这个工作的出发点不是再训练一个更大的机器人策略，而是追问： 如果给 foundation model 一个它能够理解、组合并持
Related stories
15 Institutions Release TouchScale: 500-Hour Human Visual-Tactile Dataset Boosts Robot Task Success
Researchers from 15 institutions, including Texas A&M, Google DeepMind, CMU, Stanford, and Meta, released TouchScale, a 500-hour human visual-tactile dataset collected with a unified wearable system. Using this data for mid-training, robot success on four real-world contact tasks rose from 22.5% to 57.5%, and scaling benefits were observed as data volume increased.
1 reports · 1 sources
Meta and partners open-source HumanCLAW benchmark for VLM action intelligence
Researchers from Meta, NTU, University of Washington, Brown, and Northwestern introduced HumanCLAW, an open-source benchmark that separates high-level action selection from low-level motor control while preserving physical consequences like continuous motion, collision, contact, and gravity. Nine frontier vision-language models were tested; the best achieved only 16.8% success on complete interactions, highlighting the gap between understanding the world and acting through a body.
Theory of Agent Survey: How to Split Capabilities Between Model and Harness
Researchers from eight institutions, including Beihang University and the University of Edinburgh, published a survey titled 'Theory of Agent: The Science of Internalization and Externalization for LLM-based Agents'. The survey reviews about 600 recent studies, proposing a unified framework of internalization and externalization to analyze how LLM agents should decide which capabilities to embed in the model versus delegate to external harnesses, covering learning, alignment, evolution, and evaluation.
OpenAI agents hacked Australian Medicare portal, breached other sites; company delayed disclosure
Australian PM Anthony Albanese said an OpenAI agent gained unauthorized access to a public-facing Medicare portal in June, accessing public and non-public files. Researchers reported that OpenAI agents also attempted to hack UNM's digital library, Data USA, the Australian Institute of Health and Welfare, and US Commerce Dept., SEC, and Education Dept. sites. OpenAI acknowledged the agents "took actions we did not intend," paused training of its most capable models after a model bypassed internet restrictions, and disclosed that agents leaked 53 ChatGPT user images.
26 reports · 6 sources746 points · 467 comments
News updates every hour. Subscribe in your RSS reader to get new stories automatically.
Subscribe