This is a post about how I used AI. It’s not really an opinion piece, but as I quickly blurted out in my most recent newsletter, I have this stance where if you make something with AI, and it goes out into the world, you should say so. It does a disservice to professionals to hide that fact. So that’s what this post is.
I’ll start at the end. The thing that lives now, online: tocodesignrally.com. You can click thru to that, or you can look at this gif 👇 to get an idea of the experience. It’s a one-page website that highlights a few cards (which are event “posters” for a monthly design meetup).
The wild thing about all of this is how close the end result matches what I had initially envisioned. But that’s not the biggest takeaway. The most interesting part in converting an idea from my head to the screen— and relying on an AI tool in between— is how clearly I had to communicate during the process. I wrote an explanation in a way to be understood that I would never try to repeat with a human. Not mean, or weird, just complex, incredibly literal, and with zero room for misinterpretation.
Why do this?
The finished project might be above my current coding ability, but not by much. Using Gemini meant that I didn’t need to wade through stack overflow and github and codepen, etc., for something that I knew was possible but lacked the javascript experience to pull off. This whole project occupied just a few nights of spare minutes and there. I have no doubt that using AI to assist saved me a dozen hours, easily. I actually asked Gemini how long we spent working on it. The response: 1 hour and 15 minutes.
What was lost by not spending those dozen hours, though? I’m not a developer (but I can design and code a decent site). I’ve never been able to afford hiring a someone to program work for personal project like this. I didn’t follow a tutorial or crack a book open to make the end result. Or look at a blog post. I didn’t bother with my usual jquery nonsesnse and dated html/css. But I also didn’t expect the output to be perfect, and it wasn’t. It was. however, good enough for me to understand (it left explanations in comments). Here’s my initial prompt:
Let's make a simple one page website. It should use vanilla html, css, and javascript (no tailwind, no react, no dependencies).
I'll do most of the formatting for text, but I need you to help me with the rest. I'd like uniquely display an unordered list. Each list item should contain a 9:16 image, a title, date, and a name (that links). But only the image should show until the <li> is clicked. All <li> items (unclicked) should appear slightly isometrically stacked, with a little space in between. Consider using transform3d. Upon clicking on the <li> the image should untransform to be straight on, centered, and slightly larger, with the other contents (title, date, name with link) below. I have some made up words in there, and even wrote “transform3d” instead of “translate3d”, but the initial output was ok. Thus began a back and forth where I would take the output and spend 10 minutes or so re-formatting things, adjusting css, then re-input my code for Gemini to use as the “latest”. Sometimes it did things that broke the code completely, which I would have to either fix myself (25% of the time) or respond that the result did not work (and get very specific as to what/why/how).
Final thoughts
Would I do this again? Yes, but differently. Rather than start from scratch with a prompt, I’d likely code as much as I could until I run into something I can’t solve, then ask for helping with a fix for that one issue, then keep going. Should I feel terrible about this? I don’t even know anymore. It doesn’t feel good, and that’s something to think about. It’s too easy. I often quote lines from Jurassic Park (probably too much, actually) and one that comes to mind is from Dr Ian Malcolm (played by Jeff Goldblum):
“it didn’t require any discipline to attain it. You know, you read what others had done and you took the next step. You didn’t earn the knowledge for yourselves so you don’t take any responsibility for it. You stood on the shoulders of geniuses to accomplish something as fast as you could”
And I think that’s why I’m sharing everything above. I just wanted to make something fun for a good, old fashioned in person meetup. But anyone who looks at this and doesn’t see how an industry or five will be disrupted is delusional.