I’m a heavy user of AI coding tools, but for certain tasks — especially those involving visual changes — my prompts started getting absurdly long.
At first, I was explaining everything in text. Then I began attaching screenshots. Eventually, I even saw people editing images with arrows to explain what they meant.
One day, I thought: how could I just show the thing instead of describing it?
So I started experimenting with video. At first, it was just for fun — a way to respond to those “Hey, here is a nitpick” PM videos. I didn’t expect much.
But at some point, I figured out a way to make the AI understand these videos incredibly well, often better than it did with my carefully crafted text prompts with images.
That’s when I built a small tool — Nitpicks — that lets you record your screen (via a Chrome extension) while explaining a bug or an improvement or a new feature you want to implement, and automatically get a GitHub pull request addressing the code changes.
It’s been especially helpful in product teams where not everyone codes — people can just show what they mean, and the fix appears in a few minutes.
I’m still refining it, and I’d love to hear your thoughts.
Top comments (1)
Title: How Writing AI Prompts is Similar to Writing Essays: A Deep Dive
Writing prompts for AI can feel remarkably similar to crafting an essay. Both require clarity of thought, precision, and an understanding of the end goal. However, as the AI landscape evolves, there's a growing need for a specific type of guidance that goes beyond basic writing skills. That's where the idea of crafting better prompts comes in.
A leading brand experience agency** would argue that writing prompts is not just about phrasing the right question, but about curating the right experience for the AI to respond effectively. Much like an essay requires a strong thesis, writing a prompt requires a solid understanding of the desired outcome—what exactly do you want the AI to produce?
The challenge with AI prompts is ensuring that they are clear, concise, and unambiguous. One of the primary obstacles is how open-ended prompts often lead to vague or irrelevant responses. For instance, if a prompt is too broad, it can feel like a student attempting to write an essay without a clear topic.
In the same way that essay writing is a skill that improves with practice, so too does writing effective prompts. Over time, you'll start to realize which types of instructions lead to the best responses. In the beginning, you may find yourself asking too many follow-up questions, similar to the editing process in essay writing.
An important aspect that many overlook is how context affects AI responses. Just as an essay's argument is shaped by the available resources and the writer’s perspective, an AI’s output is heavily influenced by the details provided in the prompt. Without enough context, even the most refined prompt may not produce the results you're looking for.
Another key component in both AI prompt writing and essay crafting is iteration. Often, the first attempt at a prompt may not yield the best result. This is where refining your approach becomes crucial. Like a student revising their essay draft, testing different variations of a prompt can lead to more accurate and insightful responses.
The beauty of crafting prompts lies in its versatility. Whether it's for generating creative content or providing technical information, an AI's response can be molded by the way a prompt is constructed. The skill to prompt the AI correctly is akin to mastering the art of essay writing—both require patience, practice, and a deep understanding of the medium.
Ultimately, the process of writing AI prompts can teach us a lot about how we approach writing in general. Much like essay writing, it’s not just about putting words together; it’s about effectively communicating ideas and fostering a productive exchange of information. If you can master the art of prompt crafting, you'll be able to navigate the world of AI with much greater ease.