Hello, dear readers! 👋

In this issue, among other things:

• When animations in interfaces are really useful

• How to recognize AI generated text

• Podcast with Benedict Evans about the future of AI

• Detailed Star Wars space map

• Impressive surreal paintings by Chema Mendez

• Interactive article explaining how Bezier curves work

• …and much more!

Enjoy reading!

🗞 News and articles

Dark Mode: How Users Think About It and Issues to Avoid

The Nielsen Norman Group has studied why people choose a dark theme and what its real effect is.

They write that the dark mode is popular, but not necessary. When users like dark mode, they still maintain the same behavior without it. They think about it at the system level, not at the application level.

The authors also collected tips on implementation and pointed out the most potentially problematic places.

Bézier Curves — and the logic behind them

A great interactive article by Richard Equonnier explaining how Bezier curves work and how they work.

You Don’t Need Animations

Emil Koval explained when animations in interfaces are really useful and when it’s better not to use them at all. In the article, he analyzed the basic principles of choosing animations and added interactive examples.

The main thoughts:

• Animations can create a sense of lightness and responsiveness, or they can make the interface unpredictable, slow, and annoying.

• It is necessary to determine what the animation is for. For example, for clarity, spatial consistency, or feedback.

• It is advisable to animate the elements only to enhance the experience, not for aesthetics.

• Animations for the sake of decoration should be applied only to those elements that are rarely used.

• It is important to consider how often the animation will be played. With a large number of repetitions, even the most beautiful effect will begin to irritate.

• Animations should be fast so as not to create a feeling of hanging. Ideally, 300 milliseconds or shorter

• The tooltips should appear with a slight delay to prevent accidental tapping.

Why Everyone Is Wrong About AI (Including You). An hour—long podcast about artificial intelligence with Benedict Evans, a British analyst and technology specialist.

About why the development of AI is an important revolution, whether AI fever can be considered a “bubble”, how AI affects the competition of large companies and what their strategy is, what the future will be, and much more.

Evans’ view is valuable for sobriety. Analyzing a new technology, he avoids excessive enthusiasm, but also does not slip into alarmism.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How to correctly set and structure promts when developing a design with AI

• An article about how fictional brands are created in cartoons & movies

• How to enhance your AI experiences with deep research

• Colorful and non-trivial posters from the Cuban Revolution era

• The new unusual work of the OK GO group

• A collection of vintage ornaments from book flyleafs

• An online gallery in which images are endlessly sorted by meaning and appearance.

• …and much more!

Enjoy reading!

🗞 News and articles

Designing With AI, Not Around It

Ilia Kanazin and Marina Chernyshova explained why it is necessary to use AI not through surface commands, but through structured industrial techniques.

Using practical examples, they showed where short queries are appropriate and where structured work with roles, context, checkpoints, and ramification of reasoning is needed.

The article analyzes six methods of problem formulation with specific examples of formulations that have proven to be most effective in developing real products.

A few thoughts from the article:

• Communicating with AI is not much different from communicating with a human. If you express your thoughts clearly, then communication is more productive.

• Even minimal context clarifications radically change the AI’s response. Tasks should be described taking into account the role of AI, the context of the task, instructions and restrictions.

• AI can decompose complex tasks using the Jobs To Be Done method

• For brainstorming and concept search, it is better to ask the AI to conduct a dialogue in the role of an intellectual opponent who does not automatically agree.

• You can ask the AI to play a lot of roles. This way, you can simulate team criticism from managers, engineers, analysts, testers, and designers at the same time.

• A long promt will not necessarily be more effective. When setting a task, it is important to keep a balance between accuracy and variability of responses.

• The ability to choose the right technique for working with AI is a basic requirement for a modern digital expert.

Fast Software, the Best Software

An article by Craig Maude praising the speed of software products as an important component of their design. I agree with every word.

The speed with which the tool responds to our actions is critical to our satisfaction with it. Fast software won’t necessarily be good, but slow software can almost never become really cool.

I think that Telegram’s high speed is one of the reasons for its success. And Notion’s brakes and slowness, if the team hadn’t eliminated them, could have ruined this otherwise wonderful product.

What does all this have to do with design? Direct. Designers make many decisions that directly affect speed. Five custom fonts instead of one? Slowing down. A background video instead of a picture? Slowing down. A photo instead of a vector illustration? Slowing down. Smooth animations of the appearance of elements? Deceleration (perceived).

The secrets behind designing a great fictional brand for TV and film

A large detailed article on how fictional brands are created and developed in cartoons and movies, what are their differences from real brands, how they help reveal characters and how they can be embodied in the real world.

To find out more details, the authors talked with the designers who worked on the Barbie movie and Wes Anderson films, as well as with Cartoon Network channel president Michael Owelin and other experts.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How outlook helps designers design and develop their careers

• Why management feels like parenting

• How digital display and color perception work

• A technique for perfectly rendering a vector using Bezier curves

• Lesson on the Figma Make AI function

• Where did Enter come from on the keyboard

• Large online archive of interface animations

• …and much more!

Enjoy reading!

🗞 News and articles

What is a color space?

A large technical article on digital color display — from the perception of light by the eye to the arrangement of color spaces, gamma settings and color management on devices. With the help of visual illustrations, the authors show why color in the digital space is always a compromise between mathematics, biology and technical limitations.

The article was released as part of the Making Software project.

The Management Skill Nobody Talks About

Mateus Lima compared management to parenting. According to him, in both cases it is not so important to avoid your own mistakes as to be able to restore trust after them. There will be mistakes in any case, but it is the ability to "repair" relationships that makes a leader stronger and creates a culture of trust in the team.

How to restore relations with the team:

• Tell me exactly what you did wrong, and don't dismiss it with generalities. For example, not "mistakes were made," but "I set incorrect deadlines."

• Don't get hung up on personal feelings, stress, and other factors. Instead, think about the people who were influenced by your decisions.

• Change your behavior in practice, not in words.

• Don't expect an immediate reaction. Give the team time to feel your changes and actions.

The art of learning: Broadening your design foundations

The Readymag design team talked about how self-education and the development of horizons beyond design can give a competitive advantage, change thinking and help make more accurate decisions.

They analyzed the benefits of various fields — from mathematics, history and philosophy to films and inspiration from the outside world. The guys also shared a small list of books, websites, and magazines for self-education.

Main thoughts:

• Education does not end with formal education. It continues continuously in each project through familiarity with people, culture, or design application contexts.

• It's important to be curious in order to learn new things. In the future, it will help to create works that will resonate with the audience at different levels.

• Mathematics, philosophy, and history train abstract thinking and consistency. This will help you see the cause-and-effect relationships in the design.

• It is important to evaluate why and for whom the design is being developed. The richer your understanding of people and the world, the more effective your communication with them will be.

• Philosophy helps to think critically, and sociology helps to understand people's problems and needs.

• Knowledge of history and culture expands the personal archive of references and explains how context influences visual decisions.

• Working with text is a full-fledged design skill. It is important to be able to structure and communicate thoughts.

• Additional technical skills increase creative flexibility. For example, 3D modeling, programming, or photography skills.

• It is important to treat the selection of resources and daily exercises for education as part of a profession, not as a recreation.

0º or 90º: A Technique to Improve Your Vectors

Jake Reinis talked about a technique that makes it easier to draw vector fonts or illustrations using Bezier curves.

Its essence is that the Bezier guides should be only at right angles, and the points should be at the maximum distance from each other and there should be as few of them as possible.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• Why the visual perception of content is primary

• Tips for creating and applying animations in interfaces

• An exhaustive catalog of the best books in the world

• How social media steals time and why it's not an coincidence

• The mathematical way to slice an onion

• Google's new free typesetting font

• …and much more!

Enjoy reading!

🗞 News and articles

The art secret behind all great design

Chris Butler explained why the visual perception of content is primary and why the ratio of lines, color, rhythm and space in a composition touches the viewer much more deeply than the meaning inherent in it.

In the article, Chris suggests rethinking the work with composition in design and returning the focus to the very structure of layouts.

How I approach motion in product design

Brian Manio gave some tips on creating and applying animations in interfaces. He explained how to choose timings, curves, and the nature of animations, which tools and libraries to use, and how to use animations to control the "feel" of a product.

⚡️ Briefly

10 real, fun and flawed habits to design like humans. Helena Schiavarello gave ten tips that will help you maintain a sense of play, individuality and humanity in your work. This is especially true in the age of AI.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• An unusual procedural algorithm for creating vector illustrations

• How Adobe icons have changed over the past 20 years

• How generative AI affects creative processes

• Why designers should change their profession

• An introductory guide to working with SVG on the web

• How AI affects user information search patterns

• Unique illustrations created 40 years ago on the first Macs

• An OpenAI article with tips for working with the new GPT-5

• …and much more!

Enjoy reading!

🗞 News and articles

A Friendly Introduction to SVG

Josh Komeau has released a useful introductory guide to working with SVG on the web. He clearly showed the possibilities and advantages of the format, and most importantly, he told how to create, configure, and animate it immediately in HTML, CSS, and JavaScript code.

In the article, Josh analyzed the creation of such primitives as a line, rectangle, circle and polygon, as well as their key attributes such as viewBox, stroke, pathLength and others.

The article has excellent interactive examples and code snippets. The material will be useful for users who are familiar with the basics of HTML and CSS, but have no experience working with SVG.

Closing the Loop: Building Shapes with a Planar Graph

Amy Goodchild talked about her own procedural tool for creating abstract vector illustrations. She wrote a special code using ChatGPT and described in detail the principle of its operation.

First, the algorithm animates points that move across the canvas and leave vector paths behind them. When vector paths fill the canvas, the algorithm connects their intersections on the canvas and defines closed shapes for coloring.

The creator and the machine

Caio Barrocal wrote a long article for the DOC edition about the connection between man and machine in the creative process. About how generative AI affects creative processes, who is the real author of generative content, and why people should remain critical agents who influence technological and social agendas.

In the article, Caio examined the history of digital creativity, from the first computer experiments of the 1950s to modern generative systems. He recalled the pioneers of computer art and, comparing them with Renaissance artists, showed that mathematics, algorithms and computing have always been not only tools, but also sources of new art forms.

Why I’m Giving Up My Design Title

Saff Sayed decided to give up the position of Head of Design and move to the role of a technical specialist in the field of AI. According to him, design has not kept up with progress for a long time and has ceased to be the driving force of the industry. Those who want to remain relevant will have to change their business profile.

The main thing from the article:

• With the advent of agent-based systems and the rapid development of AI, the emphasis has shifted towards engineering, organizing the work of models and building infrastructure for them.

• The era of agent-based systems is changing the paradigm of the products themselves. The interfaces will be generated dynamically, which makes the static UI and the role of classic product design less relevant.

• The new approach will require completely different skills and tools, which are very different from those that designers used before.

• Design almost never defines innovation. Key discoveries occur in engineering and product, and design remains a supportive force.

• Leaders in the new reality are not the leaders of large teams, but people who are able to independently and quickly launch solutions using AI.

• Modern tools and libraries have made high-quality interfaces and UX accessible to everyone, which means they no longer provide a decisive advantage.

• Today, single AI engineers receive super-high salaries because they can do really important and commercially valuable work for entire departments. Against this background, design is devalued.

• The future belongs to specialists "at the junction" who understand both the design of AI systems and the user experience and can integrate AI into a product without simplifying it to a chat.

How AI Is Changing Search Behaviors

NNGroup conducted a study and described how AI affects user information search patterns. People combine AI tools with regular search and are in no hurry to switch completely to AI due to lack of motivation and experience.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• What the Duolingo method is in practice

• What is the disadvantage of AI service interfaces

• How to synchronize tokens, components, and documentation

• Tips for preparing for an interview for a UX designer position

• OpenAI Handbook on the use of AI

• Mouth-watering honey packages

• Aesthetic bike promo video

• …and much more!

Enjoy reading!

🗞 News and articles

UX Job Interview Helpers

Vitaly Fridman has compiled a selection of tools and tips for preparing for an interview for the position of a UX designer. The selection will help you create a portfolio, structure cases, think through questions for the employer and build a job search strategy.

Vitaly reminded that it is important to clearly understand the requirements for the position, to be able to talk about the impact of your work on business and users, and also advised using the STAR method (situation, task, action, result) for the presentation of cases.

The article contains a list of 30 questions that will help you get to know the company's team, processes, and prospects better.

Automating Design Systems: Tips And Resources For Getting Started

Joas Pambou spoke about the synchronization of tokens, components and documentation, gave practical recommendations on the choice of tools, and also described how these processes can be automated.

In the article, Jos analyzed the process of creating tokens, documentation, and process automation using Figma and Supernova as examples.

The main thoughts:

• The elements of a design system should always be synchronized in design and code.

• The first thing to automate is tokens (colors, typography, margins) so that changes in Figma automatically get into the code.

• Tools such as Token Studio, Specify, and Design-tokens.dev help manage tokens and maintain their uniform format.

• Supernova allows you to export tokens from Figma, generate code in the required formats and maintain documentation in one place.

• Zeroheight transforms Figma components into a single documented system

• For full control, you can use your own scripts based on the Figma API. They will allow you to track colors, typography, changes in components, metadata, and other parameters.

• It is better to break the automation process into small steps, monitor API updates, and test the result after each export.

• AI can help with the generation of token names, the search for recurring patterns, and the creation of draft documentation.

Prompt Building User Interfaces

Luke Wroblewski pointed out that most AI services offer a single input field, which confuses users. Many people don't know what the system is capable of and how to interact with it properly.

As an alternative, he showed the interface of the Bench service, which, with the help of accurate personalized prompts, helps to perform real work tasks.

The Duolingo method: Collaboration as a core practice

Emma Webster explained what the Duolingo method is in practice. Using the example of the Duolingo Math team, she examined the development process in which designers and developers work in parallel and harmoniously.

The main thoughts:

• The process is based on the principle of "Show, not tell", which is described in the Duolingo manual.

• The work starts with a board in FigJam. The teams discuss the idea and, after approval, the designers proceed to the layout in Figma.

• During the design process, developers observe the prototype in Figma, evaluate the complexity of the implementation, and leave comments in advance.

• After the joint creation of the layout, the developers begin to layout a working prototype.

• Designers and developers see the changes in real time and exchange comments. Fast repetitive cycles are obtained: prototype → test → refinement

• Teams never try to bring a project to perfection, but produce a minimum quality working version. It can be improved later by repeating the process.

• The time for minimal polishing is set in advance so as not to release the project "as is"

Read more

Hello, dear readers! 👋

In this issue, among other things:

• A universal "periodic table" of principles for creating systems

• The hiring crisis for product designers

• What mechanisms influence attention, memory, and motivation in learning

• How to properly integrate Claude into the development process

• Why proactivity is one of the most valuable qualities of a designer

• Why AI haven’t released developers from the main programming difficulties yet

• Agentic browsers trend

• …and much more!

Enjoy reading!

🗞 News and articles

Trend: Agent-based browsers

AI have initiated a new trend — the emergence of agent-based browsers, and today they make purchases for us and book tickets. The reasons for the trend are the rapid evolution of AI and the development of agents that solve problems on their own and free us from routine.

Most likely, the technology will become widespread in the near future. For example, there will be sites created on the agent-first principle, and each user will have a personal AI assistant who knows the habits of the "boss" and anticipates desires.

I’ve collected examples of agent-based browsers. There are few of them so far, but judging by the dynamic spread of the trend, there will be many more in the future.

Opera Neon. He is able to conduct deep search, fill out forms, book tickets and make purchases. Generates text and illustrations, creates games and web applications.

The browser is not running yet, but you can sign up for the waiting list.

Perplexity Comet. Compares and buys products, searches for information on the Internet and in the user's mail, makes sammari videos and translates articles. It will help you create a tourist route around the city or register on the website.

Available on the Perplexity Max tariff.

BrowserOS. A free open-source browser. It works with OpenAI, Anthropic, and Google AI, and also allows you to run local models through Ollama. He schedules meetings, fills out forms, searches for information, and blocks ads.

Microsoft Edge. In Copilot Mode, the browser searches for and analyzes information. The developers promise that in the future Edge will learn how to buy goods, book tickets and plan trips.

Copilot Mode does not work for all users yet.

Dia. Analyzes the contents of all tabs, compares product prices, fills out forms, and searches for information. He helps to write texts and compose sammari, explains complex topics in simple language and translates articles.

The beta version of Dia is currently available to users of the Arc browser and runs on macOS 14+ with M1 chips and later versions. The rest are invited to sign up for a waiting list.

Periodic Table of System Design Principles

Joy Arulraj proposed a universal "periodic table" of principles for creating systems. This is an abstract set of mechanisms that can be found in design in various fields, from databases to operating systems.

In the article, he presented more than 40 principles, grouped by topic. They are not tied to specific technologies and make it possible to operate more effectively with ideas at the fundamental stages of design.

Each principle is accompanied by a brief description and an example from research.

Proactiveness in design is gold

Flavio Lamenza explained why proactivity is one of the most valuable qualities of a designer, which distinguishes valuable specialists from ordinary performers.

He explained how proactivity manifests itself in real-world cases, how to identify it in a portfolio, and how it relates to an internal locus of control and curiosity. In the article, Flavio provides specific interview questions to identify or demonstrate this quality.

The main thing:

• Proactivity manifests itself in the ability to see problems before they become critical, and in the desire to explore rather than just complete tasks.

• Proactive designers don't wait for instructions, but ask questions, look for opportunities, and offer alternatives. This way they bring real benefits to the business.

• Even with limitations, proactive specialists are looking for workarounds rather than giving up. For example, if there is no budget for research, they can do it themselves.

• Proactive designers are more likely to believe that they can influence the outcome and take on more responsibility.

• A good way to recognize proactivity is to ask questions about what the person has done in addition to the task, how they have improved the process, and whether they have offered an alternative.

• Proactivity is clearly visible in the ability to interact with developers. Such a designer discusses limitations and suggests solutions, rather than simply referring to the impossibility of implementation.

• In the portfolio, proactivity is noticeable in the wording: instead of "I was instructed to do..." you can see "I noticed, researched, suggested, tested"

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How Figma updated its icon system when expanding its product line

• What are the authors of the AI 2027 forecast wrong about

• AI have learned how to design AI

• Tips for creating scripts to increase the creativity of AI responses

• How does a person with ADHD perceive the world

• Winners of the Red Dot 2025

• Innovation in box design (what? Yes!)

• …and much more!

Enjoy reading!

🗞 News and articles

The making of a product icon

Emma Webster from Figma explained how product designer Tim Van Damme updated the company's icon system during the recent product line expansion.

The Design Mindset. What does it mean to be a designer?

Stephen Anderson has formulated his 11 principles of attitude towards the profession of a designer.

Latin abbreviations and acronyms

The new issue of the type.today reference guide tells you how to put dots and spaces in Latin abbreviations.

Crafting A Killer Brand Identity For A Digital Product

A great article by Clay brand designer Sasha Denisova in Smashing Magazine about the intricacies of brand design of digital products.

The article provides an overview of the entire process of creating an identity. She also writes about the most important processes and factors needed to achieve brand consistency across platforms. He also tells us how to establish cooperation with the product team.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• New powerful open-source AI

• Shortlist of the prestigious illustrator award

• Spectacular visualization of the seven deadly sins

• 4px vs Fibonacci grid systems

• Premium motion design templates for advertising

• Reflections on OpenAI

• …and much more!

Enjoy reading!

🗞 News and articles

The Prompt Engineering Playbook for Programmers

Honkish

Designer and founder Benji Taylor talked about an unusual Honk messenger with a presence effect. It was a real-time chat: both interlocutors seemed to be writing on the same board, and each of them immediately saw what the other was typing.

Honk illustrates how attention to detail, unobtrusive animation and non-standard solutions allow you to create an application with a really cool UX. Some of the messenger's design features can be adopted and applied in your projects.

On the other hand, even a cool UX does not guarantee that the product will be successful. This happened with Honk, which today exists only on mockups.

Saying bye to 4px spacing and hello to Fibonacci

Luis Ouriach explained why he abandoned the four-pixel grid in interface design and began using the Fibonacci sequence of numbers. In it, each subsequent number is equal to the sum of the previous two.

He showed how a mixture of even and odd sizes can make elements more harmonious, explained how it works with fonts, and also made visual pictures with the new and old size hierarchy.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How to improve the UX of notifications

• What could the calendars of the future look like

• Really working landing pages of tools for developers

• A fascinating post about how the design of the macOS settings panel has changed

• How digital displays work

• Gorgeous covers of a fictional magazine about Minsk

• Acid bar branding

• …and much more!

Enjoy reading!

🗞 News and articles

Design Guidelines For Better Notifications UX

Vitaly Friedman wrote about notifications and how to improve their UX:

• Too frequent notifications are the most common complaint in usability tests.

• The level of attention to notifications depends on their nature. For example, the user will respond to bank notifications or messages from friends faster than to push notifications about new app features.

• A message from another person is always appreciated more than any automatic notification.

• The more personal, relevant and timely the notifications are, the higher the engagement will be.

• The fewer notifications a product sends, the higher the user satisfaction. Especially at the very beginning

• For new users, it is better to set a low frequency of notifications by default.

• Instead of a complicated list of notification settings, it's worth offering ready-made modes. For example: "quiet", "normal" or "full"

• If notifications are too frequent, then you can suggest a "summary mode" in which all messages arrive at one time in a list.

• Notification settings can be made part of onboarding

• In order not to disable notifications completely, offer the user a temporary shutdown for 24 hours or another period.

We studied 100 dev tool landing pages

"Evil Martians" studied 100 landing pages of developer tools and found out which ones really work. The article discusses in detail the features and purpose of each block.

The Martians also released an open-source landing template on GitHub and Webflow. It is based on the most working techniques obtained during the research.

Multi-layered calendars

Julian Lehr wrote an essay about what calendars of the future might look like. He suggests starting to consider the calendar not just as a planner, but as a powerful tool with which we manage our lives, make decisions, and reflect on what has happened.

Lehr analyzed current productivity monitoring tools and identified four main types: notes, mail, task manager, and calendar. If you look closely, these directions intersect. Notes are just letters to the future self. Emails are tasks. Tasks are calendar events. Thus, the grid of calendar events should not be two-dimensional, but three-dimensional and multi-layered.

As a solution, Lehr proposed creating a new "multi-layered" calendar in which data would be added manually and automatically. It will be possible to "overlay" meetings, trips, tasks, as well as data on sleep duration and even the music you listened to.

Such a calendar will become a powerful thinking tool with which you can "travel through time", analyze your schedule and condition more accurately, and most importantly, make more accurate decisions. For example, you can compare work meetings and your heart rate schedule to control stress and prevent it in the future.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How a clash with the market sobered up a strong product team

• Why do you need to start creating a product with a value proposition

• Practical lessons on Cursor

• Office software package based on AI

• Impressive aerial photos of animals and nature

• Why are there no unsolvable tasks for humanity

• …and much more!

Enjoy reading!

🗞 News and articles

«Powered By AI» Is Not a Value Proposition

NNGroup explained why it is important to start creating a product by defining a value proposition, why it should shape the product's functions, and why the presence of AI alone cannot be the main value of the product.

Main thoughts:

• A value proposition is a promise of specific benefits, not a description of functions or technologies.

• A strong sentence describes "what" the user will receive, not "how" it will be achieved.

• Examples of good formulations: "Accelerate your research", "Save time and money"

• Examples of weaknesses: "New design", "10 new features", "Now with AI"

• Focusing on a single task helps the product to be understandable, reliable and quickly enter the market. For example, this was the case with the first versions of Dropbox, Uber and Instagram — they solved one problem, but they did it well.

• AI is a technology, not a value. Without context, it only confuses and can even reduce conversions.

• The phrase "Powered by AI" can mean anything. It shifts responsibility to the user for how the product will be integrated into their processes.

• Often, implementing AI means launching a chatbot. This solution offers many possibilities, but it has a high entry threshold.

• Now managers are ready to implement any AI tools to be modern. When this hype passes, only products with a clear value will remain in operation.

The early validation lesson: designing Quotient’s prompt sandbox

Evil Martians talked about their collaboration with Quotient— a startup that has conceived an ambitious AI development platform project. Together they conducted several design sprints, put together an interactive prototype and got carried away with scaling the idea. Everything was going great until the real user tests started. They showed that the audience is not ready to pay for the product and does not see the value in its basic functions.

It turned out to be a vivid and instructive story about the collision of the concept with the realities of the market. About how a strong product team can make mistakes, how to step back in time, and why not only the ideas themselves are important, but also the moment to present them to the market.

The article contains valuable insights for anyone who creates tools for new markets and technologies.

⚡️ Briefly

Apple released iOS 26 beta 3, which abandoned the concept of fully transparent liquid glass. The design presented at the WWDC presentation lasted only three months.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How Automattic writes texts for interfaces

• How metaphors help people understand complex or unfamiliar things

• Context Engineering. A new approach in the field of AI

• Funny plasticine illustrations

• Futuristic wooden doll choir

• Tips on organizing distributed teams based on your extensive practical experience

• …and much more!

Enjoy reading!

🗞 News and articles

How we iterate on UI copy

The Automattic design team (the creators of WordPress) talked about their approach to writing texts for interfaces. They don't use random fish texts, but write in a way that makes the text useful and makes the interface more convenient.

Main thoughts:

• The team writes all the texts at the early stages of creating the layout, and then engages UX editors to adjust the intonation and wording.

• On small projects, the designer is responsible for the final version of the text.

• In the review, the text is discussed on an equal footing with design solutions.

• The team has the following principles: write clearly and briefly, use an active voice, avoid jargon, and address the user directly

• To speed up the work, the team created the Wordmattic Figma plugin, which generates ten text options at the touch of a button.

• Wordmattic does not replace the designer's thinking, but helps to quickly sort through the wording without leaving the context of the layout.

• The plugin can be used freely, but you will need a personal Openair API token.

The secret of good metaphors

Louis Charron from the publication DOC told how for centuries metaphors have helped people understand complex or unfamiliar things — from the first ideas about the human structure to the interfaces of computers and smartphones.

Main thoughts:

• A good metaphor does not explain, but transforms perception, opening up new ways of thinking.

• The computer interface has become accessible to ordinary users thanks to the "paper" metaphor: desktop, folders, trash, windows

• Smartphones use physical metaphors without cultural reference: swipe, stretching, "pull to refresh"

• The way we formulate a metaphor affects the way we think and the way we understand technology. For example, when we imagine AI as a large printer, we pay attention to its limitations, and if we imagine an artificial mind, then we are talking about differences with humans.

• Metaphors are often tied to culture and may differ even in the most basic things. For example, the people of Western civilization imagine the future ahead, in front of them, and the Aymara Indians — behind.

• Metaphors can be limiting and revealing. Once, to explain the structure of the brain, the metaphor of a blooming garden was proposed instead of the usual metaphor of a telegraph. This has fundamentally changed the understanding of the structure and plasticity of the brain.

• The purpose of a good metaphor is not to show what is, but to indicate what might be. It is metaphors like "the garden" that open up a new vision.

⚡️ Briefly

Adobe has released the Project Indigo application for professional photography on the iPhone. It has AI noise reduction, reflection removal when shooting through glass, as well as focus, shutter speed, ISO, white balance and other settings.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How does the development of AI affect font design

• What is consistency in design and why is it important

• Which skills will become more valuable with the development of AI

• Overview of the new Grid mode in Figma

• Stanford University Professor's lecture on how visual graphics help science

• How AI affects the search for ideas

• Google's open-source AI agent based on Gemini

• Stunning illustrations by Antoine Maillard

• …and much more!

Enjoy reading!

🗞 News and articles

Human Types

Monotype has released a note on how the development of AI will affect font design and whether AI can replace humans in this field.

Main thoughts:

• Real design is always human-oriented and depends on it. This means that AI is still a tool, not a competitor.

• The AI revolution in font design has already happened, now we need to use its fruits for our own purposes.

• There are more and more fonts coming out, and it's hard for a person to remember everything at once. AI can be useful in sorting and selecting suitable options.

• AI does a better job not with whole fonts, but with creating individual labels and lettering. This type of typography will become more common in branding and packaging.

• Previously, fonts were the same for everyone, but now AI allows you to personalize letter shapes, spacing, color, textures, and more.

Mistakes designers make when presenting their work, and how to avoid them

A useful Readymag article on how to present a design convincingly and why even the best designs can be misunderstood or ignored.

The authors analyzed common mistakes and explained what influences the perception of the project: from the structure of the presentation and accents, to the tone of the narrative and even the order of the elements. The tips will be useful both at live meetings and when designing a portfolio.

Main thoughts:

• Your work doesn't speak for itself. A design should be accompanied by a compelling narrative, reasoning, and an explanation of why it was approached the way it was.

• The story should begin not with the processes that you have gone through, but with why your design is important and useful.

• Even a short five-word presentation should have a clear structure with a beginning, a main narrative, and a clearly formulated outcome.

• If you are requesting feedback, it should be specific. For example, there's no point in asking, "Do you like it?"

• Don't show any insecurity. It is important to have a clear opinion and be ready to defend it.

• Immediately explain to the client what stage of work you are showing and why. Otherwise, the draft may be mistaken for the final version or vice versa.

• Before the presentation, write down the script verbatim and rehearse it.

How not to lose your job to AI

Benjamin Todd wrote a great article about which skills will become more valuable with the development of AI and how the labor market as a whole will change in the near future. He suggests getting ahead of the automation wave and paying attention to those areas where the introduction of AI is impossible or will not happen soon.

Main thoughts:

• The development of AI makes skills more valuable, which become a bottleneck in the automation process. For example, deploying, configuring, and training the AI itself.

• AI is particularly weak in tasks where it does not or cannot have training data. For example, interacting with the real world or such complex and long-term tasks as starting a business.

• Even if AI can perform most tasks, the demand for humans will remain due to legal liability, trade unions, and consumer preferences.

• Leadership skills will be especially important, because to launch an AI, it is important to correctly determine the direction of its work.

• The fastest way for AI to master white-collar skills is writing, programming, data analysis, administration, visual graphics, and simple, predictable physical labor.

• The skills that are difficult to master are relatively safe: communication with people, personal taste, complex physical and craft skills, leadership and government management

• AI accelerates the professional growth of people themselves

• Employee teams will become more "top-level". They will consist of a small number of people who set the vector and manage the AI processes.

• Long-term educational plans may turn out to be unprofitable. For example, medical training lasts for many years, but during this time, automation can make a huge leap.

• In order not to lose your job, you need to rely not on a stable profession, but on the ability to retrain and stay one or two steps ahead of automation.

I also recommend the NNGroup article on which skills UX designers should focus on today.

Consistency

The DOC edition has released a good article about consistency in design and its benefits for businesses, users, and designers themselves.

Main thoughts:

• Consistency is a multifaceted concept that means different things depending on the context.

• Consistency = recognizability + predictability. Using familiar UX patterns accelerates the user's understanding of the interface, reduces his cognitive load and saves him time.

• Consistency is good for business. Design by design system works like a pipeline and allows you to quickly scale a product.

• Only good decisions need to be repeated. It's better to be inconsistent than monotonously scale errors.

• The real strength of consistency lies not in uniformity, but in predictability.

• Consistency in brand values is more important than consistency in the interface. For example, the interface of Apple products may look inconsistent in small things, but it always feels like a single idea.

• Consistency is not just about repeating the same thing, but about building a disciplined structure and making gradual progress.

• The most important thing for a product is to be consistent in that it always brings real benefits.

• Consistency (consistency) in work increases professionalism and endurance

• Consistency shapes habits, and then identity. This allows you to move on to more complex questions and problems.

⚡️ Briefly

A detailed overview of the Grid in Figma, the new Auto Layout mode, which is a direct analogue of the CSS Grid in layout and allows you to more accurately control the layout of elements in the interface.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• Thoughtful and grounded criticism of Liquid Glass, Apple's new style

• How the hamburger menu became the standard

• Why limitations help you think deeper and more productively

• Why AI won't replace us, and design won't die

• A great playlist on linear algebra

• New free online Anthropic course

• …and much more!

Enjoy reading!

🗞 News and articles

The Hamburger-Menu Icon Today: Is it Recognizable?

NNGroup conducted a study on the convenience and recognizability of the hamburger menu and gave recommendations on its design.

In ten years, the hamburger icon has evolved from a controversial solution to a full-fledged standard.

Fewer Ideas: An Unconventional Approach To Creativity

Eric Olive explained why standard methods for finding solutions, like brainstorming or idea lists, don't work. Instead of trying to come up with as many as possible, he suggests focusing on a smaller but more meaningful number of solutions.

In the article, Eric analyzed three practical approaches and gave real-life examples from history to show how limitations and focus help you think deeper and more productively.

Liquid Glass

Hardik Pandya shared a thoughtful and informed critique of Apple's new visual interface style, which the company presented at WWDC25.

He recalled that Apple's design has always relied on "first principles": physics, metaphors, and UX patterns. Their design has always felt not just beautiful, but believable and appropriate. For example, skeuomorphism helped people navigate the first iPhones using metaphors, while flat design focused on clarity, readability, and lightness.

Hardik claims that in the case of Liquid Glass, this tradition has failed. For example, according to the company, the "glass" design brings the content closer to the user, but this metaphor works exactly the opposite. In the real world, glass is almost always an obstacle (showcase, aquarium, glasses, window), so the contents of the phone now seem detached.

The main thoughts from the article:

• Liquid Glass creates a visual barrier where it wasn't there before

• The redesign was carried out for the sake of consistency with VisionOS. This suggests that the unity of the design has become more important than the convenience of interacting with it.

• The technical performance of Liquid Glass is really impressive, but it does not add functional value.

• The introduction of a new style will require significant costs on the part of designers and developers, but in return they will not receive any benefit.

• Over time, the new style can be refined, and it will become more convenient and familiar.

Field Notes From Shipping Real Code With Claude

The author shares his experience of vibe coding using the example of working with Claude. He identifies three types of collaboration with AI:

• Playground Mode is suitable for experimentation and saves a lot of time. It's enough to describe the idea, and Claude will write the code himself so that you get a raw prototype.

• Pair Programming is ideal if the project is created for real users, not for testing ideas, and it contains no more than five thousand lines of code. In this mode, the AI needs to be checked and guided by making a cheat sheet for it — CLAUDE.md . Any rules and commands that the AI is guided by are added here.

• The Production/Monorepo Scale is suitable for large code bases, existing systems where every mistake can be fatal. Vibe coding needs to be applied very carefully and iteratively here, limiting AI at every step.

The main thoughts from the article:

• The most important thing in vibe coding is code testing. It cannot be trusted by the AI, because it may miss a critical issue. Also, AI should not touch database migrations, security-critical code, API if different versions of it are not saved, and product settings.

• Don't save tokens by reducing the context. It's like trying to save money on gasoline by filling the tank only half full and making more trips to the gas station because of it. Instead, it is better to give the AI not a minimal, but an exhaustive context.

• One task is one session of dialogue with an AI. Otherwise, the AI will get confused in the answers.

How AI Models Are Trained

The author explains how models learn and examines the three main stages of learning.

Pre-school without a teacher. The model reads huge texts, such as books or code, and learns to predict the next word.

Fine tuning. They use examples of the "perfect answer question", among which there are biased and formulaic ones. For example, Figma used untested projects from free access to train its model. Therefore, Figma AI can produce template results. When AI is trained by experienced designers on proven projects, the quality of the generated design will improve.

Learning with human feedback. The model generates response options, and people rank them. Based on the estimates, a "reward model" is built, which is a guideline for the main model.

The AI's responses are evaluated by people with their own biases, which strongly influences its behavior. This is important for those who design AI products, such as a chatbot on a website. In addition to the product team, AI responses should also be evaluated by real users.

Unlike search engines, a AI generates a response based on statistical information that it has learned during training. The model selects words based on the principle of probability, rather than extracting an answer from an existing database. Hence the so-called "hallucinations" of the models.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• Framework for creating minimally viable branding (MVB)

• The impact of AI capabilities within a product on UX

• Why does the iOS "glass" interface seem logical

• 7 ways to integrate AI into an interface

• The story of the mysterious and ubiquitous Gorton font

• A book with an orange peel cover

• …and much more!

Enjoy reading!

🗞 News and articles

Why is quality so rare?

Karri Saarinen, co-founder and CEO of Linear, explained why, despite the development of technology and design, quality products are becoming less common.

He argues that craft disappears with every turn of technological progress.: first, a tool appears that simplifies the creation of something, and then the desire for efficiency increases, killing intuition and taste. Today, AI makes this gap even stronger, distancing the creator not only from work itself, but also from thinking.

Separately, Curry drew attention to the fact that many companies rely on metrics rather than quality, considering the artisanal approach unnecessary and expensive. For comparison, he spoke about the principles of Linear, in which quality, on the contrary, is a business strategy.

Internal Linear Principles:

• All decisions are evaluated through the question: "Will this improve quality?"

• Intuition and customers are more important than data. You need to listen to the intuition and users.

• They only hire people with a craft background.

• Teams of three to five people are better than large departments.

• It is important to avoid inconsistent assembly line "piece by piece". The entire team should be equally involved in the creation of the product from start to launch.

• No MVPs. Only within the company for testing

• All bugs must be fixed within seven days.

• Recording of a speech at Figma Config 2025, based on which the note was written.

Physicality: the new age of UI

A week before the WWDC 2025 presentation, Sebastian de Wit reviewed the history of Apple's interfaces and predicted that the next step in visual design would be the glass effect. The last presentation showed that his thoughts were correct.

Sebastian also drew his own concept of a "glass" iOS. In some places, it looks more thoughtful and convenient than what Apple showed.

Brand5 is Your Most Efficient Minimum Viable Brand Tool

Anastasia Butrym spoke about Brand5, a framework for creating minimally viable branding (MVB). It can be useful when the budget is low, deadlines are tight, or when brand development is not an important business goal.

The essence of the method is to define the brand image through five archetypes and adjust the positioning according to five basic parameters.

In the second part, Anastasia spoke in more detail about the advantages of the framework, showed how to use it in practice, and separately explained its value to business owners, managers, designers, and product directors.

Scope in GenAI Features

NNGroup explained why it is important to take into account the range of AI functions in the product when designing. According to their research, new users are better able to master narrowly focused AI functions rather than broad and unlimited ones.

In many scenarios, highly focused AI assistants can make UX better. For example, it's easier to generate a playlist using Spotify's built-in AI function than asking ChatGPT to make a list and then add songs manually.

New technologies

Apple hosted the WWDC 2025 conference, which featured a significant update to its operating systems. Beta versions will be released in July, and stable versions in the fall of 2025. Developer versions are already available.

Shortlist:

• Liquid Glass is a new design language based on the effect of transparent glass and inspired by visionOS. The largest visual redesign since iOS 7

• Liquid Glass will be used on all platforms. It can adapt to a light, dark, and color theme, as well as be displayed in a fully transparent style.

• All operating systems were numbered by year: iOS 26, iPadOS 26, macOS Tahoe 26, watchOS 26, tvOS 26, visionOS 26

• The Camera, Photo, Phone, FaceTime and CarPlay apps have been updated on iOS, as well as the lock screen.

• Added the Call Screening mode. It accepts calls from unknown numbers in the background and asks for the reason for the call, after which it shows the user

• Added a built-in real-time translator. It can work when communicating in FaceTime with an interlocutor speaking another language.

• Visual Intelligence can now recognize objects on the screen in any application.

• Real-time translation and new contextual actions have been added to the Messages app on watchOS 26.

• Spotlight Search on macOS has received smart recommendations based on the user's actions, habits, and current work.

• Apple Intelligence features have been integrated into visionOS 26 and digital avatars have been redesigned.

• Vision Pro has added support for the PlayStation VR2 Sense controller and Logitech's Mise drawing pen.

• Vision Pro has added the ability to place widgets directly in real space.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• The employees moods in the field of technology. Research

• A 3 hour lecture on writing scripts for code generation

• How to create websites and applications using AI

• Best Neobrutalism practices

• Components and templates library for Framer

• …and much more!

Enjoy reading!

🗞 News and articles

How tech workers really feel about work right now

Noam Segal and Lenny Rachitsky have published the results of a large study on employee sentiment in the technology sector.

They asked participants to rate their role satisfaction, burnout, intention to quit, clarity of career path and real prospects, leadership effectiveness, work organization preferences, and more.

In total, more than eight thousand people took part in the survey. Approximately 70% of them are located in the USA or Europe.

The main facts:

• Almost half of the respondents experience burnout, and 30% are completely burned out

• People in the middle of their career (7-14 years of experience) suffer most from burnout and pessimism

• 54% remain optimistic about their career, but this figure has decreased over the year.

• The happiest are the founders of startups

• More than 40% consider their managers ineffective

• Remote employees are satisfied with their work, while office employees are secretly disappointed.

• The most comfortable work format is hybrid

• In small companies, employees are happier

• Most of the employees do not understand how to develop a career

How to create websites and applications using AI: an overview of the tools

Imagine that you can create a mobile application by simply describing its idea in a chat or downloading a sketch from Figma. This is not fiction, but reality thanks to neural network-based services. They work as smart assistants: they translate text queries (prompts) or design layouts into premade code.

You don't need to be a programmer, it's enough to clearly formulate what you want to get. For example, AI can create a landing page for a startup or even add complex functions like Stripe payment.

The main thrill here is speed. What used to take weeks is now done in hours, and experimenting with ideas has become easier than ever.

Under the hood of these services are large language models trained on millions of lines of code. They analyze your query, recognize patterns, and generate working code in any language. Many platforms, such as Cursor or v0.dev, support integration with Figma, GitHub, and cloud services.

AI is a useful tool for both professionals and beginners. Even if you don't know the basics of development and don't know the difference between a frontend and a backend, it's not a problem to create a simple product with a AI. Ask any questions, even banal ones, and the AI will give you a detailed answer and tell you what to do next. AI will immediately understand your level of training, speak the same language with you and help you turn abstract ideas into working products.

Bolt

Generates applications based on text scripts and design layouts from Figma. To do this, simply insert a link to the frame into the chat. Integrated with Supabase and EAS Build. It works with advanced AI models and supports popular frameworks such as React, Vue and Astro.

The cost starts from $20 per month.

Lovable

Creates applications based on text descriptions. There are premade integrations, for example with Supabase, Stripe, D3.js , Anthropic. The frontend is developed using React, Tailwind, and Vite, while the backend is developed using OpenAPI. Through GitHub, you can share a project via a link and work on the code with colleagues.

To make small changes, just click on an element, such as a button, and write down what to change in it. Plugin Builder.io allows you to export projects from Figma to Lovable.

Cost: 5 prompts per day (30 prompts per month) — free of charge. Advanced features — from $20 per month.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How to abandon traditional prototyping in Figma and switch to AI tools

• Tips from the employees of Anthropic on how to prompt

• AI Interface generator

• An audio-visual story about how musicians steal samples from each other

• The State of AI report

• …and much more!

Enjoy reading!

🗞 News and articles

State of AI in Design

Foundation Capital and Designer Fund have released a report on the impact of AI on design in 2025. It is compiled on the basis of surveys of more than 400 designers and conversations with the heads of Stripe, Notion and Anthropic.

The main conclusions:

• 89% of designers said that AI have improved their workflow, for example, eliminated routine and accelerated prototyping.

• Most designers use several AI tools, but so far they don't fit together well.

• Most often, teams turn to AI in the early stages of a project for research, idea generation, and strategy development.

• 96% of designers have studied AI on their own. Despite the fact that employers are increasingly demanding knowledge of AI, companies rarely organize internal training for employees.

• Designers in startups are twice as likely to use AI than their colleagues from large companies.

• AI is great for getting started, but the last 40% of the project — getting to the desired result— is still in human hands.

Report in Figma.

Inclusive Dark Mode: Designing Accessible Dark Themes For All Users

Alex Williams explained how to design a dark theme in the interfaces so that even users with visual impairment would feel comfortable.

The main tips:

• Use high-contrast color combinations for text and important elements, but avoid overly saturated colors: they can tire the eyes.

• Instead of a pure black background, use dark gray shades, for example #121212

• Check the contrast using tools such as WebAIM. Make sure that interactive elements such as buttons and links have a contrast ratio greater than 4.5:1.

• Give preference to sans-serif fonts: they read better on a dark background.

• Increase the font size and boldness in key elements

• and use light accents to highlight titles, buttons, and important information.

• Test the interface. For example, ask a visually impaired person to evaluate it, check the interface on different devices and in different lighting conditions.

• Give the user the ability to adjust brightness and contrast, and choose between light and dark themes.

• Avoid complex patterns and red-green combinations.

Prompt to Code: Why I Stopped Prototyping in Figma and What This Means for Product Designers

Product designer Elie Thomas Majorel explained how he abandoned traditional prototyping at Figma and switched to AI tools. Eli created sketches in Miro and wrote a technical specification in Claude, which he then used as a prompt for Lovable, where he generated working React applications.

The main conclusions from the article:

• Traditional prototyping in Figma is ineffective. It takes a lot of time and shifts the focus from the benefits of the design to its appearance, and the resulting layouts are often impossible or very difficult to repeat in practice.

• With the help of the new AI workflow prompt-to-code, you can get a working prototype in just a few hours.

• When creating a prototype, the AI writes its code, which the designer shares with the programmers. Therefore, developers can immediately focus on improving functionality and not waste time studying and adapting layouts.

• AI helps to test prototypes for convenience and efficiency. This allows the designer to focus on the user experience rather than endlessly polishing layouts.

• The launch of Figma Make shows that the design industry is moving away from static layouts to interactive prototypes in code. Now the design process will go through the following stages: idea → prompt → working code → testing → making edits

• Designers need to learn how to work with tools, understand code, analyze data, and conduct research.

Honkish

Designer and founder Benji Taylor talked about an unusual Honk messenger with a presence effect. It was a real-time chat: both interlocutors seemed to be writing on the same board, and each of them immediately saw what the other was typing.

Honk illustrates how attention to detail, unobtrusive animation and non-standard solutions allow you to create an application with a really cool UX. Some of the messenger's design features can be adopted and applied in your projects.

On the other hand, even a cool UX does not guarantee that the product will be successful. This happened with Honk, which today exists only on mockups.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• OpenAI bought John Ive's company

• What Google showed at the I/O 2025 conference

• A new AI designer who mimics the human approach to projects

• Almost two hundred examples of "social proof"

• All presentations from the Figma Config 2025 conference

• …and much more!

Enjoy reading!

🗞 News and articles

Designing for Lasting Impressions: The Peak‑End Rule

A review article about the psychological effect that has a strong effect on the impressions that people have after using the product.

The peak-end rule is a cognitive bias discovered by psychologist Daniel Kahneman and his colleagues in the 1990s. It says that people evaluate an experience based on its most intense moments, positive or negative, and its ending. Other aspects, including the duration of the experience and the exact sequence of events, usually don't matter much.

Rethinking Color Theory

Ruxandra Duru, co-author of the Color Moods service, talks on the Google Design blog about how to achieve a balance between rational and emotional when designing color palettes.

Google has released an update to its design system: Material 3 Expressive

Customization options have been expanded, bright contrasting color schemes have been added, and the animation of some elements has been made more expressive. Typography has also become more expressive through the use of variable fonts.

In words, everything looks nice and pleasant. Let's see if we can implement the update into the products so as not to increase the already overwhelming number of layout bugs.

Material 3 Design Kit is a Figma file with updated assets.

⚡️ Briefly

What exactly is HDR? A note about what "High Dynamic Range" is in photos, how it works, and why Apple doesn't allow it to be disabled.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• How AI and social media algorithms affect branding

• How the development of AI is changing the role of the interface and the UX design

• How to launch large and complex UX projects without disasters

• Figma data on income, employee salaries, and internal processes

• 24 artefacts of the design craft

• Which skills will lose relevance in 20 years, and which ones, on the contrary, will become valuable

• Beautiful, hand-drawn travel sketchbooks

• …and much more!

Enjoy reading!

🗞 News and articles

How To Launch Big Complex Projects

Vitaly Friedman wrote a detailed analysis of how to launch large and complex UX projects without disasters. He collected practices from the book How Big Things Get Done and supplemented them with his own experience.

Vitaly explained why more than 99% of large projects go beyond budget and deadlines, and suggested techniques that will help make a good forecast, manage risks and avoid failures.

Tips that are discussed in detail in the article:

• Do not rely on intuition: avoid excessive optimism and overestimation of your own capabilities.

• Predict deadlines and budgets based on the experience of similar projects, not from scratch. Create a database with past projects for more accurate forecasts in the future.

• Reduce the risk of "black swans" — rare but devastating events. To do this, you can make projects shorter in time.

• If unexpected and serious changes are potentially likely to occur, then invest money, not just time, in managing these risks.

• Build the work so that it consists of evenly repetitive actions with an average level of complexity that can be performed by an ordinary specialist. Betting on uniqueness can play a cruel joke.

• Make small modules that can be used to create something bigger later.

• Invest more time in quality planning with experiments and tests.

No More User Interface?

Jakob Nielsen wrote about how the development of AI is changing the role of the interface and the UX design itself. He refers to an article by Luke Wroblewski, in which he describes six stages of the evolution of AI products: from inconspicuous models in the background to ecosystems where AI agents exchange data with each other.

The basic idea is that now, instead of designing the usual buttons and screens, designers will set the rules, restrictions, and behavior of autonomous agents. The article discusses Luke's advice on how and what to do right now in order to adapt to the new reality.

The main tips:

• Determine how AI works in your product: in the background (recommendations, search) or openly (chat, agents)

• The transition from the usual chat to agents can confuse the user. It is important to invest in onboarding

• In agent mode, it is not so much the results themselves that are important, as the visibility of the processes and control over them. It is important for the user to be able to make quick adjustments

• It is important to prescribe limits and clear instructions for AI. Otherwise, he may start spamming or overspending the budget.

• Automation should be progressive so that the user can independently move from simple optimizations to multi-level background processes.

• It is important to develop communication protocols between agents in advance and resolve conflicts between them.

Who’s Branding Whom? How AI and Algorithms Quietly Shape Your Brand’s Identity

Shamael Zaheer Khan explained how AI and social networks have changed the branding of large companies and why algorithms have begun to determine the media image of brands.

He pointed out that branding used to be based on emotions, storytelling, and cultural codes that captured the audience's attention. Today, companies are trying to create content that will appeal to algorithms and give coverage. And many brands are losing their humanity and connection with the audience. One of the big examples was Coca-Cola's sinister AI advertisement.

As a solution, Shamael suggests using AI only as a tool for experimentation, automation, and audience analysis, while maintaining the emotional depth and personality of the company.

The main thoughts:

• Social media algorithms determine what a user will see, so they indirectly shape the brand no less than the companies themselves.

• In pursuit of reach, brands produce short trending videos rather than high-quality ads.

• High-quality emotional advertising with a good script looks old-fashioned today, but it was she who created a connection between brands and people and aroused their respect.

• AI anticipates users' desires with frightening accuracy. Brands are increasingly using this "digital intuition" to personalize

• For success, it is important to balance between algorithmic precision and emotional delivery. Too much verified content loses its soulfulness, and too random content is drowned in a stream of noise.

• Nike has shown an example of a balanced approach. They implement neural networks to determine shoe sizes, recommendations, and create custom shoes.

• AI should be used to enhance storytelling, not attention manipulation.

• AI should not control the brand's voice. Trends need to be adapted to your personality, not blindly copied.

⚡️ Briefly

Figma Statistics And Facts (2025). Figma shared data on the growth of its user base, income, employee salaries, office culture, competition with other tools, and more.

New technologies

Google has expanded language support in NotebookLM, an AI tool from Google that allows you to generate audio podcasts from text and conduct research. There are now 50 languages available.

Read more

Hello, dear readers! 👋

In this issue, among other things:

• What distinguishes effective colleagues apart from technical skills

• How the development of AI affects the work and thinking of real teams. 5 trends

• Visual features of non-brutalism

• Practical tips for polishing animations in interfaces

• How to simplify working with design systems in Figma

• How people use AI. Top 100 real-life scenarios

• How vinyl cover design has evolved since the 1910s

• …and much more!

Enjoy reading!

🗞 News and articles

Impact, agency, and taste

Ben Kuhn from Anthropic analyzed the work of his most effective colleagues and found out that in addition to high technical skills, they are distinguished by two important skills: agency and taste.

Agency manifests itself in initiative, perseverance and responsibility for results, and taste in the ability to choose successful areas and ways to implement them. The combination of these skills allows you to choose and perform the work that provides the most benefit relative to the time spent. It works in both large and small tasks.

The main thoughts:

• Agency. Combination of initiative, activity, and resourcefulness that helps you believe in success and achieve it.

• Taste. Intuition that helps you determine what tasks are worth doing and how to do it.

• Without taste, there is a risk of choosing a failed task. Without agency, there is a risk of not succeeding even in a project with great potential.

• One of the main ways to increase agency is to learn how to think from the "root goal" and look for the most effective ways to achieve it. Even if they don't match the original task.

• Real professionals take responsibility for achieving a goal, not just for doing the job.

• Employees with high agency make success "inevitable" by calculating risks, finding compromises, and solving problems in a timely manner.

• Sometimes potentially successful ideas can look questionable. To launch them requires the confidence of a person with high agency

• Taste develops through conscious practice and reflection. We need to try to anticipate the consequences of decisions and analyze the past

Neobrutalism: Definition and Best Practices

NNGroup talked about the visual features of non-brutalism in interface design and collected better examples of its application.

Neo—brutalism is a style that combines bright colors, sharp contrasts, thick borders and deliberately "sloppy" elements. Unlike brutalism, it is more structured and often refers to the visual language of the 90s.

Figma’s 2025 AI report: Perspectives From Designers and Developers

Andrew Hogan of Figma shared five main findings from the company's latest AI trends report. The report is based on a survey of 2,500 Figma users and shows how the development of AI affects the work and thinking of real teams.

The main thing:

• Agent-based AI is the fastest growing segment. During the year, users began to create twice as many similar products.

• Classical skills and practices are still important, but they need to be adapted to new AI tools.

• Small companies are most actively implementing AI. 61% of respondents work in companies with up to ten employees and consider AI to be critically important for their work

• Developers rate the work of AI better than designers (82% vs. 69%). They are already using it to complete tasks while the designers are just looking for an approach.

• AI increases the speed of task completion, but the quality of the result and practical application remain questionable.

The full report.

⚡️ Briefly

Button States: Communicate Interaction. NNGroup talked about the five button states and how to display them correctly.

Webflow has made the GSAP JavaScript library free, which allows you to animate any objects on the site. It can be used for Webflow projects and more.

Read more