Understanding the diverse forms of User Interfaces and their impact on UX

If you're feeling a little burnt out or uninspired, let me show you how to fix it. 1. Detach from the identity of "designer" 2. Try a new tool or skill (and be bad it) 3. Find some people to learn from I've got you covered: Josh Guo will inspire you with data biz and creative code Phobon will hit you with threejs and shaders Kat will dazzle you with AR prototypes Alan Ang will teach you about computer vision Canvas of Kings will show you what it looks like to solo build an best-in-class mapmaking UX (without AI) Sundays are for the gamers.

If you've spoken with Noam Tenne or me lately, chances are we've been talking about "dynamic software", or "software you can summon". Spin it up on the fly and it appears. Then, as you interact, it grows and evolves without explicit prompts or a traditional feature roadmap. With Anthropic's new model release today, they also released a demo of "Imagine with Claude", bringing that idea closer to reality. What’s different from typical chatbot & vibe coding UX: you prompt once, and the system keeps generating new UI components as you interact, expanding piece by piece from the initial element. In the video below, I asked for an AI Tamagotchi landing page in pixel-art. As I clicked buttons, new sections emerged—no planning, no additional prompts. Curious to see what new experiences this unlocks.

UX - with sticky notes or algorithms? Lately I’ve been experimenting by creating UX Scaffold — a small web app built on the basics of #OOUX Object Maps. 👉 Try it here (desktop only): https://lnkd.in/di3J2KMC As a UX Designer, I’ve used OOUX for years and I love its analog clarity: sticky notes, markers, and no screens pulling focus. But I’ve also wondered — can we let digital tools handle some of the structure work? Especially when it comes to visualizing relationship diagrams or enabling AI assistance. UX Scaffold adds a JSON-based foundation that makes it possible to: 🤖 Generate and refine Object Maps with AI 🔗 Automatically build relationship diagrams 🔄 Keep everything in sync across views I’d love to hear what the OOUX community — and Sophia V Prater — think about the approach. Can we use AI to help us think better when reflection is needed — and move faster when we’re exploring? And should we? #UXDesign #AI #DesignTools #Lovable #ProductDesign

Yesterday evening I stumbled upon Enrico Tartarotti’s YouTube channel. He makes great videos about new technologies, AI, design, and product development. And one video really caught my attention. It’s about how AI might change the future of user interfaces. The main idea is that while we used to obsess over optimizing UX: simplifying actions, reducing clicks, making everything as efficient as possible, people are now moving away from traditional interfaces altogether. For example, instead of Googling something, we just ask ChatGPT. Which means optimizing search results pages may no longer make sense. The takeaway for me is that soon, the user interface as we know it might be replaced by AI, which will generate a personalized interface based on each specific query. So, the designer’s role will shift: from creating the “best UX” to teaching AI how to respond to users in the most intuitive and helpful way. As a user, this sounds incredibly exciting - finally, for example, I could just see the plane tickets I want right away instead of fiddling with Skyscanner’s search and filters. However, as a product manager, this really makes me think - what will be the future or my role, of UX roles and what can I do now to prepare for it. Not yet sure, but very curious to find out. What do you think: will the “traditional” UI ever go away? 🎥 https://lnkd.in/e5ZKcVtE

🔥 Design Hot-takes — with Morten Rand-Hendriksen We’ve been asking our George UX Conf speakers to share their spiciest design opinions — and Morten didn’t hold back. From how we (wrongly) expect users to find help, to what AI should — and absolutely shouldn’t — touch, his takes cut right to the core of designing for clarity and trust. 👉 Swipe through the carousel to see what he said — and decide for yourself: is he spot on, or stirring the pot? 📅 Join the livestream on November 5th to hear more from Morten https://lnkd.in/d9m4-yBF #DesignHotTakes #GeorgeUXConf #UXDesign #Fintech #AIinFinance #ProductDesign

🎯 AI Accessibility Analysis — My MVP Project I recently built a small AI-powered app designed to help designers and teams identify accessibility and design issues early in their process. The idea was simple: ⚡️ Upload a UI screen → let AI analyze it → receive structured, actionable feedback based on accessibility (WCAG) and design best practices. This project began as part of an AI Challenge by WIT, exploring how AI can be applied to real design workflows - turning curiosity into something tangible. 📌 I built the MVP using n8n automation, a LLM for the analysis logic, and Lovable to bring the interface to life. My goal was to create a proof of concept that shows how AI can assist with: ✔ Visual accessibility reviews (contrast, legibility, hierarchy) ✔ Structured feedback organized by severity and issue type ✔ Faster, more objective design evaluations You can try it here 👉 https://lnkd.in/damwVr5s I’d love to hear your thoughts, impressions, or ideas for improvement - every perspective helps shape what this could become next 🚀

We often focus on beautiful UI elements and smooth interactions, but what about the foundation that makes everything findable and usable? We're talking about Information Architecture (IA)—the art and science of organizing and labeling content so users can efficiently find what they need. Think of IA as the blueprint for a building. Without a clear IA, even the prettiest design is just a pile of bricks. Good IA involves things like: Taxonomies: How we group and label content. Navigation Systems: The paths users take. Sitemaps: The overall structure of the product. Where in your current project do you feel the IA is the strongest? Or conversely, where did a challenging IA problem force you to get creative? Share your experiences and any great examples of brilliant IA you've encountered! #dc #DesignCommunity #Design #Community #LH #InformationArchitecture #IA #UXDesign #ContentStrategy #DesignSystems

⚡️𝐅𝐚𝐬𝐭𝐞𝐬𝐭 ≠ 𝐁𝐞𝐬𝐭 𝐕𝐨𝐢𝐜𝐞 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭𝐬⚡️   Speed looks great in a demo or benchmark. But in a real interaction, it often means mid-sentence interruptions, frustrated users, and worst of all - missed critical words. That’s speed at the expense of accuracy and UX, and it can stall the whole conversation.   𝐓𝐡𝐞 𝐆𝐨𝐥𝐝𝐢𝐥𝐨𝐜𝐤𝐬 𝐚𝐫𝐞𝐚: I believe that speed 𝐝𝐨𝐞𝐬 matter for voice UX, to a point. After that, the accuracy of the Speech-to-text and other components outweighs the benefit of only going fast. 🎯 Garbage in == Garbage out (that old adage)   I’ve been following the work from people like James Zammit - Roark (YC W25) and Brooke Hopkins (Coval), and can see the depth of the challenge to rigorously test these agents, which goes beyond simple speed (e.g observability and real-world testing - links below)   Great agents don’t just react fast. They 𝘭𝘪𝘴𝘵𝘦𝘯 well, wait for the right time to answer, and respond appropriately.   Benchmarks measure millisecond speeds across STT/LLM/TTS components. Users measure how the conversation 𝘧𝘦𝘦𝘭𝘴.   𝐆𝐨𝐨𝐝 𝐯𝐨𝐢𝐜𝐞 𝐚𝐠𝐞𝐧𝐭𝐬 𝐚𝐫𝐞 𝐟𝐚𝐬𝐭 𝐞𝐧𝐨𝐮𝐠𝐡 - 𝐚𝐧𝐝 𝐩𝐫𝐞𝐜𝐢𝐬𝐞.   Curious to hear from a wider group: ★ Have you ever slowed an agent down to make it feel more natural? ★ Where do you draw the line between latency and UX? ★ Any tricks for better turn-taking without adding lag? 👉 Drop your tips, stories, or insights below. Would love to learn from you.   Further reading 📖 James about the need for observability in voice AI https://lnkd.in/eMFWcuqD Brooke on the case for testing agents like you would a self-driving car, not just in a demo. https://lnkd.in/enttT7M8 https://lnkd.in/e-dKb_xy

Love this post from Stuart Wood at Speechmatics on the tradeoff between speed and accuracy in voice AI. ⚡️ We’ve seen the same: the fastest agent isn’t always the best agent. A slight pause can make the conversation feel more natural, but without observability, it’s impossible to know whether that change is improving or hurting UX. That’s exactly why we built Roark (YC W25) - to make sure teams can see when their agents interrupt, mishear, or regress, and test changes before they go live.

Source: https://lnkd.in/diX5tXsc 🚀 The web isn’t just evolving—it’s redefining itself. From data-first to model-first design, AI-first development is no longer a niche experiment but a seismic shift in how we build digital experiences. 💡 Why it matters: Model-first design centers intelligence as the core of architecture, letting AI shape routing, state, and UX dynamically. Remix v3’s AI-aware loaders and route-based mental models are paving the way for apps that anticipate user needs—no bolt-ons required. This isn’t just about tools; it’s about rethinking how we architect context, adaptability, and real-time intelligence. 🌐 The call to action: Are you ready to move beyond static schemas and embrace fluid, AI-driven workflows? The future belongs to those who build for relevance, not rigidity. Let’s rethink the web—what will your next project look like? 🌐✨ #AIFirstDesign #WebDevelopment #RemixV3