北京时间 10 月 21 日,英伟达创始人兼 CEO 黄仁勋在 GTC 柏林大会上表示,AI 的发展将增加软件工具的使用者,工具公司也将向 AI 公司和智能体公司转型。
黄仁勋在演讲中提到,有人认为 AI 的到来意味着软件和工具公司的终结,他对此持相反看法。他认为,随着 AI 智能体承担更多工作,过去开发的软件工具将迎来更多使用者。
支撑这一判断的原因是,AI 要完成实际工作,仍然需要调用专业工具。黄仁勋以科学家和工程师的工作为例说明,即使具备很强的智能,也需要实验室、模拟器和求解工具。AI 智能体同样需要借助工具开展实验、仿真和数据处理。
他表示,编程是智能体较早落地的领域,编译器或 Python 环境就能兼容其中一些任务。随着智能体进入更多行业,需要调用的工具也将更加多元,涵盖流体力学、量子化学和工业外观软件。
大会播放的一段演示展示了这种工作方式。在达索平台相关演示中,AI 智能体依据工程要求,将大众高尔夫的车门外观调整用于途观,并调用外观与仿真工具测试思路。仿真发现振动难题后,外观智能体继续修改,最终由工程师审核并挑选思路。
按照黄仁勋的判断,工具公司将来将同时给出软件,以及经过专门优化、擅长使用这些软件的智能体。
这一判断也是英伟达拓展业务的关键依据。黄仁勋在演讲中反复介绍其模型、计算平台和软件库,希望为智能体调用专业工具给出计算兼容。现场演示展现了这种工艺路径,但并未给出其大规模部署后的成本与效率数据。
相关阅读:
本文来源:https://www.ithome.com/1/016/125.htm
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。
On October 21, Beijing time, Huang Jenxun, founder and CEO of Nvidia, said at the GTC Berlin Conference that the development of AI will increase the use of software tools, and tool companies will also transform into AI companies and agent companies.
In his speech, Huang Jenxun mentioned that some people think that the arrival of AI means the end of software and tool companies, and he believes the opposite. He believes that as AI agents take on more work, the software tools developed in the past will usher in more users.
The reason for this judgment is that the AI still needs to call professional tools to complete the actual work. Taking the work of scientists and engineers as an example, Huang Jenxun needs laboratories, simulators and solving tools even with strong intelligence. AI agents also need tools for experimentation, simulation, and data processing.
He said that programming was an area where agents landed earlier, and compilers or Python environments were compatible with some of these tasks. As agents enter more industries, the tools that need to be called will also be more diverse, covering fluid mechanics, quantum chemistry, and industrial appearance software.
A demonstration played at the conference demonstrates how this works. In the relevant demonstration of the Dassault Systèmes platform, the AI agent adjusted the appearance of the door of the Volkswagen Golf to the tip according to the engineering requirements, and called the appearance and simulation tools to test the ideas. After the simulation finds the vibration problem, the appearance agent continues to modify, and finally the engineer reviews and selects the idea.
According to Huang Renxun's judgment, tool companies will give software at the same time in the future, as well as agents who are specially optimized and good at using these software.
This judgment is also a key basis for Nvidia to expand its business. In his speech, Huang Jenxun repeatedly introduced his model, computing platform and software library, hoping to give compute compatibility for agents to call professional tools. The live demo shows this process path, but does not give cost and efficiency data after large-scale deployment.
This article is from: https://www.ithome.com/1/016/125.htm For copyright issues, please write to: 13951431913#139.com or submit a review on the right.