Reddit 用户称 Claude 重建未完成的无人机论文仿真
一名 Reddit 用户称,借助 Claude 重建了一个未完成的混合移动无人机本科论文仿真。用户表示,重建工作在一个月内完成,包括在 Gazebo 中移植并建模 SolidWorks 装配体、开发控制器、配置网络属性以及构建计算机视觉模块。该叙述还称,用户负责定义需求并作出工程决策,同时质疑当日常性工作被自动化后,该项目是否仍属于自己。该说法来自未经核实的个人陈述,仿真功能、工作范围及一个月的时间线均未得到独立验证。
TLDR; I couldn't finish my drone thesis years ago because I had no hardware, no robotics experience and no one to ask. I used Claude to rebuild it in a month, and every engineering decision is still mine. Now I'm wondering if I can take credit for it, and where the value in work comes from when the grunt work is automated. Is it still my project, or a Ship of Theseus?
A few years ago, I worked on a hybrid mobility drone for my bachelor's thesis. Government restrictions on drones were severe at the time, so we couldn't buy hardware or build a prototype. A simulation and digital twin was the best we could do. I was 22 with zero experience in ROS, drones or robotics. I had to work out how to get a SolidWorks assembly into Gazebo, model the system, develop the controllers, set up network properties and build a computer vision module. No one at my university, or in my country really, had meaningful robotics experience I could ask. After many endless nights I had a basic simulation running, but it was never finished work. That has haunted me ever since; a chip on my shoulder, so to speak. Recently I picked up Claude Pro as a way to get back into prototyping and DIY projects, and fixing my thesis felt like the right place to start. One month in, the whole system is up and running. We work in sprints: I define the requirements, we modernise components, redesign parts, and I watch it come to life. It has been incredibly rewarding. But it's also hard to take credit, even though the concept and every engineering decision is mine. If the grunt work gets automated, and eventually the tools outwit their human counterparts, where does the value of our work come from? It feels like the Ship of Theseus. Is this still the same project, and is it even mine anymore?