Articles

A lot was written last week about Alan Dye’s departure from Apple as VP of Human Interface Design to lead design at Meta as Chief Design Officer. I shared a few critical pieces about Liquid Glass in the last few months, and thought this story from John Gruber was quite telling: ​“After I published that post, I got a note from a designer friend who left Apple, in frustration, a few years ago. After watching Jobs’s Aqua introduction for the first time in years, he told me, ​“I’m really struck by Steve directly speaking to ​‘radio buttons’ and ​‘the key window’.” He had the feeling that Dye and his team looked down on interface designers who used terms like Jobs himself once used — in a public keynote, no less. That to Dye’s circle, such terms felt too much like ​“programmer talk”. But the history of Apple (and NeXT) user interface design is the opposite. Designers and programmers used to — and still should — speak the exact same language about such concepts. Steve Jobs certainly did, and something feels profoundly broken about that disconnect under Alan Dye’s leadership.”

A really interesting look at modern CSS patterns that 37signals use in Campfire, Writebook, and Fizzy. No Tailwind, no build process, and lots of cutting-edge CSS with good browser support — custom properties (variables), native nesting, container queries, the :has() selector, CSS Layers for managing specificity, color-mix() for dynamic color manipulation, and clamp(), min(), max() for responsive sizing without media queries. 

Great piece by Laura Escobar on what it’s like to be a founding designer: ​“Ship. Ship like your life depends on it. Because, for the life of the business, it depends on it. And to ship the right bets, you have to step away from just looking at pixels to understand what needs to get built and why.”

Also, this: ​“This doesn’t happen in isolation, which is why influence is the best design tool. Your job isn’t designing features. Your job at a startup is about staying alive. Treading water until a boat floats by, and then figuring out how to build a boat that matches the boat that floats by.”

Fast Company spoke with Figma’s head of AI, David Kossnick, about what the company has accomplished so far, where he’s trying to steer it, and why the tech industry needs to move past prompting and create experiences that are ​“more visual, more exploratory.”

Reuters: ​“Design software company Figma was hit with a proposed class action in California federal court on Friday for allegedly misusing its customers’ designs to train artificial intelligence models. The lawsuit said the company used its customers’ data and intellectual property without permission to train its generative AI tools, which according to the complaint led to San Francisco-based Figma’s ​“sky-high valuation” in a $1.2 billion initial public offering earlier this year.”

Loredana Crisan, Chief Design Officer at Figma, introduces Gemini 3 Pro as a new experimental model in Figma Make.

Connect external tools to Make to pull in PRDs, tickets, and product documents, so you can create prototypes with full context. Update your connected docs or create tasks directly from Make to keep everything in sync. Supported connectors: Asana, Atlassian (Confluence, Jira), GitHub, Linear, mon​day​.com, and Notion.

Molly suggests using separate color palettes for marketing and product design, as they have different goals and speed of iteration.

Noah Levin, VP of Design at Figma, shows how creative image prompting with the new Nano Banana Pro model can add serious value to your work across all of Figma’s products. I really liked the practicality of examples in this article, from updating a headshot to match the rest of the team to preparing a dark version of the illustration.

Max Woolf ran experiments with image generation models and was really impressed with Nano Banana 2.5 — note that this blog post was published before Google introduced the new Pro version. ​“After running Nano Banana through its paces with my comically complex prompts, I can confirm that thanks to Nano Banana’s robust text encoder, it has such extremely strong prompt adherence that Google has understated how well it works.”

Yann-Edern Gillet adopted his original talk at Granola’s Design Engineering Night into an article: ​“The real Rosetta Stone didn’t solve languages, it overlapped them. Same meaning, carved three times, so people could decode one through the other. Design engineering is the same: we’re constantly trying to express intent twice — once visually, once in code, without losing the meaning in between. When the languages don’t match, the wall gets thick. When they overlap, the wall becomes translucent. And when the overlap is high enough… the wall disappears. It becomes glass. I think that overlap, that shared percentage of language, is the real definition of velocity.”

A new survey, based on ten in-depth interviews with 340 people, reveals that product designers are facing high stress, burnout, and slow career progression, largely due to chaotic organization, poor management, and unpredictable communication within the tech industry.

“For years, India’s design and developer community has pushed the boundaries of what’s possible in Figma. Now, with our new hub in Bengaluru, we are excited to meet these builders where they are—and learn firsthand how they’re reimagining what’s next and shaping the future of product development in India.”

Luis explores a few approaches for handling components used in multiple design files but not yet ready for the main library. Delaying the creation of components until later doesn’t scale. Using naming conventions and canvas organization to separate local components doesn’t help with bringing them to another file. The best solution is to create Staging Libraries, ​“where your specific team, feature, squad, whatever can create what is in effect an extension of the system for your specific piece of work.” When the component needs to be used across multiple files, it can be pushed from the local level to the Staging library, and once it’s ready to become a part of the design system, it can be moved to the Global library.

A smart technique for debugging design system usage that will nicely complement the new Check designs linter: ​“I wanted to see which parts of my designs were using tokens and components from my design system, just by looking at the canvas. The solution: an additional variable mode paired with an outline component. Toggle it on, and everything using the system lights up: tokens in bright cyan, components with dashed outlines. Everything else stays unchanged.”

“Code Connect UI lets you map design components in your Figma libraries to the corresponding code components in your repository. These mappings enhance the Figma MCP server by giving AI agents direct references to your code, enabling more accurate implementation guidance.”

A summary of everything Figma announced at Schema to help teams design for the AI era. Extended collections are a new way to manage multi-brand design systems, where authors can release a simple whitelabeled version of their design system that designers across the company can extend with their own themes, publish, and reuse. Slots let you add your own layers within instances and easily specify which instances a slot accepts, allowing for both increased usability and compliance with your design system. Check designs linter matches your raw values with their corresponding variables. Finally, the team completed a massive rewrite of the architecture for massive performance gains.

In addition to new design features, Figma has been working hard to bring context from your codebase into your design system. With the new Code Connect UI, users can connect Figma directly to their GitHub repositories and use the new AI suggestions feature to quickly find the right code file to map to Figma components — no coding necessary. The MCP server is out of beta and generally available — now you can add guidelines for how AI models should adhere to your design system. Make kits let you generate React code components and CSS files for your styles and variables, then package those outputs for use in Figma Make. Additionally, Figma announced NPM package imports, native importing and exporting of variables, simplifying authoring experience for collection, and increased variable modes.

Earlier this year, Grammarly acquired collaborative workspace Coda and email app Superhuman, with the CEO from Coda stepping into the role of CEO for Grammarly and the entire company later changing its name to Superhuman. Smith & Diction developed an interactive identity with motion design at its core, while the brand architecture of the new company was evolving every day. The new icon system by Helena Zhang and custom pattern generators made in Figma Make show an incredible attention to detail from this team.

Dylan Field on the newest addition to Figma’s product line: ​“Figma has acquired Weavy, a platform that brings generative AI and professional editing tools into the open canvas. As Figma Weave, the company will help build out image, video, animation, motion design, and VFX media generation and editing capability on the Figma platform.”

In A Match Made in Heaven, Weavy’s early investor, Ben Blumenrose from Designer Fund, shared three key features of their product approach that make for a very powerful tool — being model agnostic, exposing process, and working as an aggregator.

Dave Martin, Security Engineer at Figma, shares his experience building Response Sampling, a system designed to detect potential sensitive data leaks in real time. ​“By providing ongoing visibility into the data leaving our services, Response Sampling gives our teams the opportunity to investigate and address issues quickly, reducing the risk of exposure and improving our confidence in how data is handled.”