Let's compare xitter.cf/unrootdesign/status/21…

Nowadays people spend more time checking who they talk to than talking. You write someone with a normal question, and instead of an answer you get “are you AI? is this an agent?” The conversation dies right there. Not because anyone was a bot, but because suspicion ate the point of talking. Funny time. Proving you’re human got harder than solving the actual problem

Clients taught me how urgent a website really is. A founder comes in needing it ASAP. The project starts, and then he drags it himself for a year. Feedback waits, content waits, decisions wait. I stopped taking "urgent" at face value. When it's truly critical, the client moves as fast as we do. Now I judge urgency by week 1, not by the kickoff call

Nowadays animation is must

Design for agentry in progress

A client challenged us He showed us a website he'd built himself with AI "Can you do better?" That was a bold question The website was genuinely good But we're the experts. We couldn't let the client do a better job than us Challenge accepted We designed the site in Figma and built the animations and interactive elements with Claude It took more time than usual But the client got more than just static mockups He got a website he could see in motion and try out right away We pulled it off And that win still feels good

We now hand clients animated heroes at the design stage, built with Claude. A year ago our animations lived only in Webflow, at the build stage. Clients approved static mockups and imagined the rest. Now they see their website moving before a single line of production code. Hard to believe how fast the industry turned. This got so accessible with AI that not using it feels wrong

A new client started with our 1-week trial. Week 1 is done, everything delivered, all running smoothly. We'll see where it goes. Love this format more and more. The client risks little, and we show real work instead of promises

Yesterday, you added another feature to your website Now there are 10, and the site still isn't live You're betting on 10 things at once A month of sleepless nights, and you could end up with a product nobody wants Here's what to do instead - pick one problem you want to solve - build a simple version with just one feature - show it to real users - see if they're willing to use it Don't try to impress the market Find out if people actually want your product

your thoughts?

For years I repeated the industry mantra: a designer doesn't draw pictures, he solves business problems. I stopped saying it. To actually solve a business problem you need the business inside out: the product, the math, where clients come from, what they pay for. An outside designer almost never has that access. Me included. So now I put it humbler: we make the website clear and honest, the client runs the business. This split works better for everyone

wrapped up a video for SIntra AI

Our own website waited about a year for a redesign. I kept wanting it perfect. Last month I set myself a hard frame, rebuilt it between client projects and shipped. Designers do the same with portfolios for years. Startups do it with products. For me the deadline did more for quality than the year of waiting

bento animations made with Claude