Newsletter 7 / Can AI truly boost your productivity?

I didn't send my BearwithAI newsletter last week because I was swamped with my day job. We're close to delivering a new feature, so I had to finalize the design and deliver it to the dev team, and properly explain the design decisions to stakeholders. In short, I was too busy to send out the newsletter.

Interestingly, not only did I lack the time to share AI tips, but I also didn't use AI to support my work during this crunch time. This struck me as odd, considering I once believed AI could be a major productivity booster. If that's true, why didn't I use it?

To answer that, let's first define productivity. For me, productivity means accomplishing similar tasks in less time to produce more outputs. While AI, specifically generative AI, can create content quickly, it often can't solve complex problems and sometimes ends up taking more time overall. So, in general, AI for productivity can be an illusion.

For example, over the past few weeks, I've been iterating high-fidelity designs, gathering feedback, finalizing designs, documenting them, and communicating with the team. I also worked closely with developers to ensure everything was accurate. This required precise actions where AI couldn't really help. The only AI support I used was Grammarly and ChatGPT to quickly check messages I was sharing with my team.

Reflecting on this, there are some steps AI might handle, like documenting design decisions or checking design specs. However, this requires investing time to explore and integrate AI into my workflow, which might take more time than doing it myself. Plus, current tools aren't quite there yet for tasks like reviewing and correcting design specs from a prompt.

So, what's the takeaway? First, AI isn't a silver bullet. It can improve productivity for specific tasks, but it isn't a game changer for designers. Critical steps like team communication and articulation are still human domains. Second, invest time to explore AI options when you're not busy. Check if AI can be integrated into your workflow and gather real data. Then, when you're in busy mode, implement the specific steps you've tested to ease some of the workload.

For me, I'll explore using AI for design documentation. That way, I won't miss another newsletter just because I'm busy writing design docs after work.

上周我没有发送 BearwithAI 新闻简报，因为我被本职工作忙得不可开交。我们快要交付一个新功能了，所以我必须完成设计并交给开发团队，并向利益相关者详细解释设计决策。简而言之，我太忙了，没时间发送简报。

有趣的是，我不仅没有时间分享 AI 小贴士，而且在这个紧张的时期，我也没有使用 AI 来支持我的工作。这让我觉得很奇怪，因为我曾经认为 AI 可以大大提高生产力。如果这是真的，为什么我没有使用它呢？

生产力的提升是不是像 ChatGPT 有时产生的虚幻一样，只是一种错觉？

为了回答这个问题，我们首先需要定义生产力。对我来说，生产力是指在较短的时间内完成类似的任务以产出更多的成果。虽然 AI，特别是生成式 AI，可以快速创建内容，但它往往无法解决复杂问题，有时总体上反而会花费更多时间。所以，如果不具体场景的话，用AI 提升生产力可能是一种错觉。

举个例子，在过去的几周里，我一直在反复修改高保真设计，收集反馈，最终确定设计，记录它们，并与团队沟通。我还与开发人员密切合作，以确保一切准确无误。这需要精确的操作，而 AI 并不能真正帮上忙。我唯一使用的 AI 工具是 Grammarly 和 ChatGPT 来快速检查我与团队分享的信息。

回顾这段经历，有些步骤 AI 可能能够处理，比如记录设计决策或检查设计规格。然而，这需要花时间来探索并将 AI 集成到我的工作流程中，这可能比我自己做还要花更多时间。而且，当前的工具还不足以处理从提示词中审查和纠正设计规格这样的任务。

那么，结论是什么呢？首先，AI 不是灵丹妙药。它可以提高特定任务的生产力，但对设计师来说并不是一个颠覆性的改变。像团队沟通和表达这样的关键步骤仍然是人类的领域。其次，当你不忙时，花时间探索 AI 选项。检查 AI 是否可以集成到你的工作流程中并收集真实数据。然后，当你忙碌时，实施你已经测试过的具体步骤，以减轻部分工作负担。

对我来说，我将探索使用 AI 来记录设计文档。这样，我就不会因为忙于下班后写设计文档而错过另一封简报了。