me, myself & a(i)

Years ago, when I was an engineer, I felt like I wanted more control over creating experiences. Simply developing them wasn’t enough, so I pivoted into design. And I didn’t just do a bootcamp or an online certification. I went to the US to do a master’s in HCI. Call me type A or old school, but I wanted a formal education to take on this career switch. After being a textbook nerd all my life, I was ready to do some serious learning- one that had very little to do with exams and tests.

And so began my completely co-dependent, slightly unhealthy but absolutely fulfilling journey with design. It wasn’t just about getting straight As. It was about creating something that worked seamlessly and solved a real problem, and if I was lucky, I got to see it get built. Prototyping an idea felt like being cooped up in a science lab inside my brain and seeing what came out. Not only did I fall in love with design, but I also fell in love with the process of actually designing screens in Figma on my own. Figuring out a new shortcut, improving the workflow, or executing a cool animation felt like a real flex.

Cut to today, a world rife with products, designs, and all sorts of imagery built completely using AI. Anybody can be a designer when you’ve got a Claude subscription or a multitude of tools available at your disposal. It’s become so easy, and cheap, to build stuff that practically everyone thinks they’re sitting on a gold mine of a SaaS product. But how much of what’s being built actually makes sense? Forget being good or useful.

Designers are bragging about how they’ve stopped using Figma altogether. They’ve started building tools, designing on Claude, and shipping websites without ever making a wireframe. While there’s nothing wrong with directly building things, there’s always a price to pay. Doing things the old-school way allowed us to have complete control over everything. You knew the entire idea inside and out, knew the constraints, trade-offs, and reasoning behind every design decision being made. But when you build your ideas directly, you lose some of the control and confidence in the outcome to the sheer speed with which you can ship. You can obviously spend more time and prompt your way to restore that control and confidence, but how many are actually doing that before bragging on LinkedIn with a post that’s also written by AI?

As a product designer, my job today includes a lot of prompting, decoding AI responses, and screaming at a wall when it can’t understand my vision. My darkest horror story is running out of session credits. The turnaround time to deliver every task has been cut down drastically. Otherwise, the expectations within the same time period have gone up. This has led to a feeling of fatigue and a lack of interest and connection with the work. The most draining task, taking up most of my mental bandwidth, is decoding AI responses to determine what’s useful, what’s complete slop, and what’s a level of detail that’s completely unwarranted.

It’s hard to comment on where the industry is headed, but it looks bleak at best. Everyone is now trying to play catch up with someone else’s AI responses/outputs. But no matter what prompt you give, it’s never gonna give you the exact same result because that’s how LLMs work. And isn’t that completely laughable. We’ve stopped getting inspired from humans but rather just narrowed down to getting our AI agents to work a certain way (like that one person’s LinkedIn post).