Designy Digest / 002

The Daily Sprint, Episode 6

Give it a listen and find out whether design is dead. Spoiler alert: you know...

Design is Dead Final, Final

Design is dead (again) and AI does all the heavy thinking for you—is that true? Yet, there is a fundamental framework that applies to any design process regardless the tool.

DesignyDarrell Estabrook

Something’s up on LinkedIn

This week, I noticed something odd about certain LinkedIn posts. I suspected they were written by AI, but now I know. How?

I stumbled upon a service called GPTZero.

When you analyze text, it creates a report with the percentage chance the text was written by AI. The web app goes into extreme detail and lets you investigate sentence by sentence and why it was classified as human or AI.

There's a Chrome link which makes scanning LinkedIn posts, or any webpage content, very easy.

GPTZero assessed this LinkedIn post with a 100% chance of being AI-generated. It sounds great, but by definition it’s not an original thought from the author. That is, even if we assume this was content based only on a library of the author’s own work, the output is contrived in such a way as to be detectable as being AI-generated.

I went to the web app and played around with running my own text through it and it comes back as 100% human. Even more, running text I generated with AI, even though it sounded great, came back as 100% AI. Editing that text came back with mixed percentages.

If you're trying to figure out if a post is AI only because it uses em-dashes that's not going to cut it—believe me, I love em-dashes because they're so typographically interesting.

Authentic writing is a cornerstone of building genuine relationships. So, it's a shame to post something one didn't actually write. You can be sure I'm writing this and all the content at Designy and for The Daily Sprint podcast

Even so, here’s the caution—this is an AI analysis and AI can make mistakes! Even so, it’s fascinating to see there are so many uses of AI embedded in our everyday communication, but it doesn’t need to be.

I’m including a referral link to try GPTZero. I'm happy for you to try it with or without the referral because it can come in handy during everyday browsing.

Oh well, the old product design process doesn’t work now

Our entire process was designed around one fear: Building the wrong thing is expensive. PRDs existed to make sure we build the right thing. Design existed to make sure we build it right. Reviews… | Matt Przegietka | 13 comments

Our entire process was designed around one fear: Building the wrong thing is expensive. PRDs existed to make sure we build the right thing. Design existed to make sure we build it right. Reviews, approvals, sign-offs,... all risk reduction layers because building was slow and costly. That made sense when a prototype took a dev team 3 weeks. It doesn’t when it takes 30 minutes. When the cost of building drops to near zero, the entire process built around avoiding building... breaks. PRDs are already becoming a bottleneck. Design is next. The role is flipping. From gatekeeping before the build to shaping after it. Instead of: Think → Specify → Design → Build It’s becoming: Build → Evaluate → Refine → Document We used to need 3 layers of protection before writing code. Now we need taste, speed, and judgment after. The designers and PMs who thrive will be the ones who stopped doing their jobs in the wrong order. ✌️ What do you think about this? How will our work evolve? Let’s discuss. | 13 comments on LinkedIn

LinkedInMatt Przegietka

This particular post came out as a mixed AI generated by GTPZero, but I was more interested in the graphic showing the Old process and the New process we’re encouraged to adopt.

I go in depth on this on The Daily Sprint episode 6, Design is Dead Final, Final

Interesting, but this is simply rearranging the deck chairs.

“Thinking,” “Specifying,” and “Designing” are more relevant than ever and will be differentiator skills in the world of AI. Well, they've already been differentiators for decades of product design. Actually, for millennia of human history. It’s how we’re made.

AI is short-circuiting the concept of design if it simply starts with building.

Defining purpose and discovering context is critical if any product will be designed to fulfill an outcome.

You can read more about what I mean by reading Know purpose, know design, and listening to the podcast on Making messy design clear.

Do you say, “I’m sorry” in your design reviews?

There’s a phase in almost every design career where you start apologizing for the work. You overexplain. You justify spacing, padding, component decisions. You walk people through every pixel like… | Tyler White

There’s a phase in almost every design career where you start apologizing for the work. You overexplain. You justify spacing, padding, component decisions. You walk people through every pixel like you’re asking for permission to exist in the room. I did that early on. A lot. I thought being collaborative meant making everyone comfortable. So I led with craft. With design language. With the how instead of the why. And every time I did that, I made design feel optional. What finally clicked for me was watching other disciplines. Engineering never apologizes for constraints. QA never explains why tests are necessary. Their process is assumed to be required to do the job properly. Design is the only function that keeps trying to earn the right to do its work. And when you do that long enough, people start treating design like a nice-to-have. Something you include if there’s time. Something you skip when deadlines get tight. The shift wasn’t about being louder or more assertive. It was about changing the frame. Stopping the pixel tour. Stopping the jargon. And talking about consequences instead. What breaks if this isn’t tested. What risk we’re taking on by skipping this step. What it costs to fix after launch instead of before. Once design is framed as protection instead of preference, the dynamic changes. You’re no longer defending your process. You’re explaining the tradeoffs the business is about to make. That’s when the apologies stop. Not because you toughened up, but because the work finally sounded like what it is. Necessary.

LinkedInTyler White

This was 100% written by human Tyler White according to GPTZero. He makes a great point about apologizing for your designs. Although his point is to turn the conversation to making the observer think about consequences of the design.

I have a more collaborative approach.

Tyler is correct—you never should defend your designs. However, you should always explain them. What's the difference?

Defending weakens your position as the lead designer. It's admitting you didn't design it well. That makes the design subjective.

Instead, explaining the design reveals the reason for every design decision you made. That's leadership and that's objective.

The trick, though? You have to have a clear understanding of what the intended outcome was for the feature or app you're designing before you begin. If the outcome was vague and you never clarified it, then those design decisions will come across as a miss.

Even so, there's a clear way to get feedback that's productive and actionable.

Thanks for reading. Have a great day designing with a why.