在阿联酋首都举办的“万物皆AI”活动汇集了来自60多个国家的科技公司、学者和投资者。
The Ai Everythng event in the capital of the UAE brought together technology companies, academics, and investors from more than 60 countries.
阿布扎比国家展览中心主办了此次活动，超过400家人工智能（AI）基础设施提供商和初创企业参与其中。
The Abu Dhabi National Exhibition Center hosted the event featuring more than 400 artificial intelligence (AI) infrastructure providers and startups.
亚马逊、英特尔、微软以及许多其他公司在活动期间向200多名投资者展示了他们的AI解决方案。
Amazon, Intel, Microsoft, and many other companies showcased their AI solutions to more than 200 investors during the organization.
波士顿动力公司的联合创始人兼机器人与人工智能研究所执行主任马克·雷伯特在活动期间接受阿纳多卢通讯社采访。
Marc Raibert, co-founder of Boston Dynamics and executive director of the Robotics and AI Institute, told Anadolu at the event.
雷伯特表示：“你知道，现在是投身AI领域，尤其是机器人物理AI方向的绝佳时机，而这正是我的研究领域。”他指出，该行业近期取得了重大进展。
"You know, it's a great time to be in the AI and especially the physical AI part of robotics, which is what I do," Raibert said. He noted the sector experienced significant developments recently.
雷伯特表示：“我认为今年以及过去这一年发生了太多事情，投入的资金也非常巨大。”
"And I think there's just so much going on this year, the last year, there's been huge investment," Raibert stated.
展望未来，他补充道：“看到所有即将实现的目标，将会令人惊叹。”雷伯特解释说：“在机器人与人工智能研究所，我们开发的是物理系统和移动系统，这些系统有些与众不同。”
Looking ahead to future years, he added, "It's going to be amazing to see all the things to get it accomplished." "At the Robotics and AI Institute, we develop physical systems, mobility systems that are a little bit unusual," Raibert explained.
他强调：“我们开发手部灵巧性系统，让机器人能够用双手完成更多任务。”
"We develop manual dexterity systems that will make robots able to do more things with their hands," he highlighted.
但雷伯特指出：“其中一个理念不是专注于即时产品，而是致力于长期解决方案，这些方案将使下一代机器人比今天的能力更强。”
"But one of the ideas is not to work on immediate products, but to work on the longer range solutions that are going to make the next generation of robots much more capable than how they are today," Raibert noted.
在题为“物理AI的崛起”的专题讨论中，他提到，AI语言模型可以轻松访问海量互联网数据，但为物理机器人收集数据却面临巨大挑战。
During his speech at a panel titled The Rise of Physical AI, he mentioned AI language models easily access massive internet data, but gathering data for physical robots poses significant difficulties.
他解释说，在名为“观察、理解、执行”的项目中，他们利用大型语言模型来教导机器人该做什么。
He explained they utilize large language models to teach robots what to do in their project named Watch, Understand, Do.
雷伯特表示，将人类的学习过程应用于工厂中的机器人，会引发一个难题，开发者必须从零开始构建技能，并专注于感知能力。他将开发人形机器人比作攀登机器人领域的珠穆朗玛峰。
Applying human learning processes in factories to robots creates a difficult problem requiring developers to build skills from the ground up and focus on perception, according to Raibert. He described developing humanoid robots as climbing the Mount Everest of robotics.
Raibert表示，人类在手动灵活性和感知世界方面仍然独一无二。
Raibert stated humans remain unique in manual dexterity and perceiving the world.
他补充说，机器人技术每天都在进步，但目前尚未达到能够真正与人类竞争的水平。
He added robot technology advances every day but currently lacks the level to truly compete with humans.
谈及特斯拉首席执行官埃隆·马斯克希望在五年内生产数十亿台人形机器人的目标时，Raibert指出该领域存在过于乐观的预期。
Touching upon Tesla CEO Elon Musk's goal to produce billions of humanoid robots in five years, Raibert pointed out overly optimistic expectations in the field.
他对马斯克在SpaceX、Starlink和特斯拉方面取得的成就表示钦佩。
He expressed admiration for Musk's achievements with SpaceX, Starlink, and Tesla.
Raibert表示，马斯克可能无法在五年内实现所宣称的机器人产量，但他已成功成为推动公众关注该领域的一股重要力量。
Raibert stated Musk will probably not produce the stated number of robots in five years but successfully became a fundamental force drawing public interest to the sector.
超过180名专家在活动的多个专题讨论中评估了人工智能的最新进展。
More than 180 experts evaluated the latest developments in AI during various panels at the event.