In partnership with
Hey {{first_name | there}},
A designer I know spent 3 years learning Figma, color theory, and interaction design.
Last month, someone shipped a better landing page than hers using Claude and a SKILL.md file they found on GitHub.
She was not replaced by AI. She was replaced by someone using AI better than her.
That difference matters a lot.
Here is what’s actually happening right now.
A set of open-source Claude Code skills are going viral on GitHub, and collectively they are turning Claude into something that used to take a full design team to produce.
I want to walk you through four of them because once you see what they do, you will immediately understand why UI/UX job postings fell 71% between 2022 and 2023 and the numbers in 2026 are not getting friendlier.
1. Impeccable
Most AI-generated interfaces look the same because AI has no design vocabulary going in. Impeccable fixes this by loading 7 reference files before every prompt runs.
These cover typography, color using OKLCH, spatial design, motion, interaction design, responsive breakpoints, and UX writing. It is basically handing Claude a design school curriculum before it touches a single line of code.
Best for: Anyone who does not know why their AI-generated UI looks generic but cannot pinpoint what is wrong.
2. Taste
This one is interesting. It forces Claude to randomize layouts rather than defaulting to the usual hero-features-CTA structure.
It also builds GSAP scroll animations in automatically and structures pages using the AIDA framework (Attention, Interest, Desire, Action).
The result genuinely looks like something that could win an Awwwards entry rather than a $5 Fiverr template.
Best for: Anyone tired of every AI landing page looking identical to the last one.
This one gives Claude searchable design intelligence. It knows UI styles, color palettes, font pairings, landing page patterns, and UX guidelines.
The key is that it tailors all of this to your specific tech stack and product type. Building a SaaS dashboard in Next.js is different from a consumer mobile app, and this skill knows the difference.
Best for: Anyone who wants the output to actually fit their product rather than look like a template applied on top.
Created by Addy Osmani from Google, this is less about aesthetics and more about how senior engineers actually ship interfaces.
It includes structured workflows for testing, code review, performance, CI/CD pipelines, and security. This is the skill that closes the gap between a nice-looking component and something you can actually push to production.
Best for: Anyone who builds or modifies real interfaces and needs the output to survive contact with the real world.
AI Updates You Should Know
Elon Musk lost his lawsuit against OpenAI and Sam Altman after a federal jury ruled in OpenAI’s favor in just two hours. The case was one of the biggest legal battles in AI so far and could influence how future AI companies are structured and governed.
A college graduation ceremony in Arizona turned chaotic after an AI-powered name reading system glitched and skipped multiple students during the ceremony. The crowd started booing, and the college eventually switched back to a human announcer. Another reminder that AI still struggles in real-world, high-pressure moments.
Anthropic acquired Stainless, a developer tools startup used by OpenAI, Google, Runway, and Cloudflare. Stainless automates SDK generation for APIs, meaning Anthropic is now strengthening its developer ecosystem and infrastructure advantage, not just competing on model quality.
Seth Rogen slammed AI-generated writing in a recent interview, calling it “stupid dog sh*t” and saying people who rely on AI for writing “shouldn’t be writers.” He also emphasized that his new animated film used fully hand-drawn animation with no AI involved.
Claude announced a major upgrade to Claude Design, doubling token limits across every plan. People are already using it for larger design workflows, longer contexts, and more complex creative projects. The race between Claude, Codex, and other AI tools is getting more intense every week.
Now here is the uncomfortable part.
An MIT study published this month scanned the brains of 54 people across three groups: those using ChatGPT, those using search engines, and those writing on their own.
The group relying on AI showed up to 55% reduced brain connectivity. The wildest finding? 83% of LLM users could not quote from essays they had just written using AI.
The researchers called this cognitive debt.
Now think about this alongside the design skills above. If you use Impeccable and Taste to build beautiful interfaces without understanding why they work, you are probably getting great output today.
But when the tools change, or when a client asks you to explain your design decisions, or when you need to solve a problem these skills were not trained for, what do you have?
The designers who are genuinely winning right now are not the ones who outsource design thinking to AI. They are the ones using AI as a finishing tool after they have already thought through the problem.
The skills above are powerful. They are also dangerous if used as a replacement for knowing what good design actually is.
Use them. Just do not let them think for you.
Reply and let me know which of the four Claude skills you are going to try first. I’m curious what you’re building.
- Aashish
Smart starts here.
You don't have to read everything — just the right thing. 1440's daily newsletter distills the day's biggest stories from 100+ sources into one quick, 5-minute read. It's the fastest way to stay sharp, sound informed, and actually understand what's happening in the world. Join 4.5 million readers who start their day the smart way.