Case of a designer being changed by ChatGPT
It’s 2024, and using a computer currently feels like stepping into the special results of the motion pictures we watched as kids. We were promised to be able to talk with our computer systems, zoom-in right into low-res pictures and see all the information in one click etc. I bear in mind checking out my C- 64 wishing someday computers will certainly be effective enough to in fact do this type of magic, while twisting the screwdriver in a cassette player in order to load the Great Giana Siblings from a ferro-magnetic bootleg tape. Fast-forward 30 something years:
To be straightforward, talking with ChatGPT as my clients once did when I was a graphic developer yields remarkably excellent results. They’re not perfect at first, as this example reveals:
My very first response was “OK this switch actually IS uber-clickable but not actually useful of what I was servicing”. I just needed a switch that “looks better” i.e. has the very same form, proportions, shades, font styles … yes sounds like useful context to start with. However it does not have to be. I made certain we can still arrive in a really couple of motivates. What worked was including the adhering to information:
– it is a possession on a webpage it needs to be approximately the exact same size form and bgr is white,
– see to it that the form of the button is rectangle-shaped, vast and short sufficient to put “REGISTER” in one line
Utilizing it, ChatGPT redacted the timely:
A realistic-looking rectangular button with a smooth gradient from light orange to crimson, and the message ‘SUBSCRIBE’ in vibrant, white, sans-serif font style throughout one line. The switch has a subtle, sensible appearance recommending a semi-gloss surface and is designed with minor shadowing to show up clickable against a white page background. The form is a wide rectangular shape, ensuring the message fits on one line without breaks, dimension 1792 x 256
and used DALL-E to produce:
All this took place about in the amount of time that 10 years ago I would certainly make use of to start my computer, open Blender and erase the default cube. However is this possession any type of helpful for prototyping? Virtually.
Next off, I required to cut it out from the history and separate the darkness. And for this I went super old-school and pre-owned Adobe Photoshop. Some 9 months ago they have presented the “Things Selection Device”– it should take care of an easy form similar to this, right?
It is just a rectangular shape with rounded corners and Adobe constantly seemed to me like a specialist on mathematics behind shapes like these however I assume, judging by the outcomes, these are still two different modern technologies:D
So yes, there is still some area for manual labor to be done. You can specify a vector shape, use it as a mask, invert it, choose the shade variety of the darkness and cut it out. In this manner you will certainly have 2 elements and a transparent history. I would claim it suffices to utilize it on a harsh mockup. There are also some common AI-hallucinated artefacts like forms that make extremely little feeling from a typographic point of view. But once again it’s absolutely nothing that a great old stamp and brush will not repair:
Now this ultimately appears like a task for a human designer! One could invest at least a couple of hours making the shape appearance clean. Repairing the lights, shadows, geometry and gradients up until it looks excellent. However do we truly need it? Depending upon the context we could hold a series of conferences to review what to do with it or validate what works better through abdominal muscle examinations, but considering that we’re a small indie workshop, at this stage all this activities are out of scope.
… and it still looks better than just a level shape performed in Miro.
Wait what? Using Miro for UX?
Several of you may react allergic to creating mock-up screens in Miro as opposed to preparing them in Figma, however when the layout is relatively simple, there are not many elements and all you need is some standard format– I would certainly suggest that Miro + Photoshop + standard HTML and CSS is enough. And if you want to have responsive style or parallax impacts, I think you can just ask ChatGPT to compose some custom.js Would I wager my last dime on it? Definitely not. Would I compose an another write-up if it looks respectable enough? Absolutely!
(EDIT: see https://medium.com/@mdzikowski/create-responsive-websites-with-chatgpt-miro-and-canva-in-minutes- 631 ae 9452 c 83 for the follow-up)
Verdict
So right here’s the thing: even as AIs like ChatGPT and DALL-E make layout faster, we’re the ones that make certain that our design does not simply look good, it has to do with making it feel right. I don’t think we’re being replaced– we’re developing. As designers, our duty is to mix AI performance with our creative instincts. We’re adding heart to the silicon. And that’s something AI can’t replicate.