内容来源：https://news.microsoft.com/signal/?p=724

内容总结：

微软首席技术官凯文·斯科特近日在旧金山与创业者交流时指出，当前是早期创业者把握人工智能机遇的“黄金窗口期”。他认为，如今实验成本更低、工具更强大，为快速试错的创业者留下了广阔空间。斯科特结合自身学术、创业及大型工程领导经验，分享了五点关键建议：

一、关注真实趋势，重视高频反馈。创业者应辨别技术、经济或行为中“不可忽视的变革力量”，并基于实际使用数据持续验证方向，避免被市场噪音干扰。

二、善用“能力盈余”，直面创业艰辛。现有AI系统蕴藏远超日常应用的能力，创业者需通过扎实甚至“枯燥”的产品构建与工程实践，将其转化为实际价值。

三、开放与封闭模型不是对立阵营，而是工具箱。应根据具体问题灵活选择技术方案，以务实态度组合使用开源与专有工具。

四、把握低成本实验窗口。借助普及化的工具与智能体，以往需大量资源投入的试错如今可快速进行。创业者应敢于广泛探索，将实验视为持续过程而非前期步骤。

五、聚焦赋能人类与创造真实价值。技术意义在于拓展人的能力，产品应致力于切实改善人们生活与工作。在AI快速演进中，以“服务他人”为锚点的思考将成为关键差异点。

斯科特强调，当前阶段充满“前所未有的机遇”，鼓励创业者以无畏精神运用技术为世界带来积极影响。

中文翻译：

– 本文阅读时间约4分钟。

初创企业如何把握当下AI机遇：微软首席技术官凯文·斯科特的五点洞见

作者

对于早期创业者而言，此刻正迎来前所未有的机遇窗口。

微软首席技术官凯文·斯科特上周在旧金山南园公地（South Park Commons）与创业者和技术专家们交流时如此表示。他指出，代码助手与大上下文模型等工具如今能力显著提升，试错成本却降至历史低点。对于敢于快速验证想法的人们而言，众多领域仍存在广阔的开拓空间。

在与南园公地联合创始人阿迪亚·阿加瓦尔的炉边对话中，斯科特谈到，当前创业者正面临一个难得的机遇窗口——关键在于屏蔽外界杂音、把握真实信号，并通过反馈保持务实态度。

拥有横跨学术界、初创企业及大型工程领导经验的斯科特，呼吁在场的创业者们积极拥抱动态AI格局中蕴藏的机遇。

"我深知创造事物的艰辛，这需要持续打磨。但请务必认清此刻的独特性，"他说，"我相信在座各位都有潜力对世界产生巨大影响，只要你们能无畏地运用这项技术造福他人。"

这场对话揭示了创业者充分把握当前机遇的五条路径：

1. 聚焦现实，持续获取反馈

斯科特认为，无论创业者意志多么坚定或乐观，客观现实始终是必须面对的构建基础。他强调，必须清醒区分主观愿望与正在发生的实际变革。

成功的构建者能识别那些不可忽视的力量——无论是技术、经济还是行为层面的——并使自身工作与之同频共振。

但他同时指出，这些判断需要持续接受压力测试。在当前AI领域，必须尽早且频繁地验证想法，优先关注高质量信号而非市场杂音。他建议创业者关注那些来自实际使用和持续互动的、更细微却更可靠的指标。

2. 善用"能力盈余"，但需做好苦练准备

斯科特指出，当前AI系统蕴含的能力远超大多数应用的开发程度。他将这种"能力盈余"视为重要机遇，但强调需要创业者投入繁琐工作来释放这些潜力。

"要榨取这些系统的全部能力，有些工作就像不起眼的管道工程，或是枯燥的产品构建，"斯科特坦言，"但既然选择创业，这就是常态。更多时候需要的是持续打磨。"

3. 开源与闭源并非对立，而是工具箱的选项

尽管开源与专有AI模型常被置于对立框架讨论，斯科特提出了更务实的视角。

他认为不同工具适用于不同场景，最高效的解决方案往往需要根据具体问题组合多种方法。这种选择应基于实际需求而非意识形态，关键在于为特定任务选择合适的构建模块。

"将二者视为非此即彼的选项，本身就是一种认知误区。"他总结道。

4. 拥抱实验精神，把握极低的试错成本

斯科特特别强调了当前实验验证的廉价与高效。曾经需要专业团队或大量计算资源的构想，如今借助普及的工具和代码助手即可快速验证。他指出，这种环境更青睐那些广泛探索、追踪意外发现，并将实验视为持续构建过程而非前期步骤的创业者。

"放手实验，大胆尝试，"他鼓励道，"我衷心建议大家不要因畏惧失败而束手束脚。"

5. 始终聚焦赋能人类与创造真实价值

斯科特强调，只有当技术能拓展人类能力边界时，它才真正具有意义。即便AI系统飞速演进，核心问题始终不变：这项成果是否切实改善了人们的生活或工作？

他指出，关键在于选择值得解决的问题，深入理解背后的人文情境，并致力于产生真正助益他人的成果，而非单纯追逐新奇。在AI既能放大影响力也可能加剧干扰的时代，将产品决策锚定在"帮助他人"这一原则上，既能成为稳定基石，也能形成独特优势。

"我们运用这些工具时，究竟能为同胞提供何种服务与价值？"他反问道，"某些领域必将发生剧变，但我认为所有变化同时也为人们带来了一系列绝佳机遇，因为这里存在着多方共赢的赋能机制。"

观看斯科特与阿加瓦尔的完整对话视频请点击此处。

英文来源：

– The estimated reading time is 4 min.
Startups can do more right now with AI: 5 insights from Microsoft CTO Kevin Scott
Author
This is an unusually promising moment for early-stage builders.
That’s what Microsoft Chief Technology Officer Kevin Scott told a group of entrepreneurs and technologists last week at South Park Commons in San Francisco. Tools such as coding agents and large-context models are more capable, he said, and the cost of experimentation is lower than ever. Entire categories are still wide open for people who are willing to test ideas quickly.
In a fireside chat with South Park Commons’ Aditya Agarwal, he said there’s currently a rare window for founders who can ignore the noise, follow the real signals and stay grounded through feedback.
Scott, whose career has spanned academia, startups and large-scale engineering leadership, urged the audience of founders to embrace the opportunity embedded within the dynamic AI landscape.
“I know building things is hard. It’s a grind. Just don’t lose sight of how special this moment is,” he said. “I think everybody in this room has the potential to make a massive impact on the world, by just being fearless about how you put this technology at work for other people.”
The discussion highlighted five ways founders can make the most of this moment.

1. Focus on what’s real, and seek frequent feedback
   For Scott, tangible realities define the landscape founders must build within, no matter their willpower or optimism. So it’s crucial to recognize the difference between what someone wishes would happen and the shifts that are already underway, he said.
   Successful builders, he said, spot the forces that can’t be ignored — technological, economic or behavioral — and align their work with that trajectory.
   But those assumptions have to be pressure-tested constantly, he added, emphasizing the need to validate ideas early and often and to prioritize high-quality signals over noise in the current AI landscape. He encouraged founders to identify the smaller, more reliable indicators that come from actual usage and sustained engagement.
2. Take advantage of the ‘capability overhang’ — but be ready for the grind
   Today’s AI systems already contain far more capability than most applications draw upon, Scott said. He highlighted this “capability overhang” as a major opportunity for founders willing to do the messy work required to unlock it.
   “Some of the things that you need to do to squeeze the capability out of these systems is just ugly-looking plumbing stuff, or grungy product building,” Scott said. “But you’re in a startup, that’s kind of your life. It’s more about the grind.”
3. Open vs. closed isn’t a battle — it’s a toolbox
   Debates often frame open-source and proprietary AI models as opposing camps, but Scott suggested a more pragmatic view.
   Different tools serve different purposes, he said, and the most effective solutions will likely combine approaches depending on the problem at hand. The choice is practical, not ideological, he said, and what matters is selecting the right building blocks for the right job.
   “The category error here is thinking that it’s got to be either/or,” he said.
4. Embrace experimentation and the unusually low cost of trying things
   Scott underscored how cheap and fast experimentation has become. Ideas that once required specialized teams or significant computing resources can now be tested quickly with widely available tools and coding agents. This environment rewards founders who explore widely and follow unexpected findings, he said, and who treat experimentation as an ongoing part of building, not a preliminary step.
   “Do the experiments. Try things,” he said. “I would really encourage folks to not be precious about the possibility of failure.”
5. Keep the focus on empowering people and building things of real value
   Technology only matters when it expands what people can do, Scott said. Even as AI systems improve rapidly, he said the core question stays constant: Does the work make someone’s life or job meaningfully better?
   Scott pointed to the importance of choosing problems worth solving, understanding the human context around them and aiming for outcomes that genuinely help others rather than chasing novelty for its own sake. In a moment when AI can amplify both impact and distraction, he said, grounding product decisions in helping people becomes an anchor — and a differentiator.
   “What am I doing with these tools that is of service and of value to my fellow human beings?” he said. “There are certain things for sure that are going to change in pretty dramatic ways, but I think everything that’s changing is also presenting a set of super interesting opportunities for people, because there is a positive-sum empowerment mechanism that’s at work here.”
   Check out the full conversation between Scott and Agarwal: