TLDR Design 2025-12-24
Luma AI Start to End Video 🎥, ChatGPT Wrapped ✨, Alibaba AI Image Layers 🖼️
ChatGPT launches a year-end review like Spotify Wrapped (2 minute read)
ChatGPT is rolling out a Spotify Wrapped–style annual review called “Your Year with ChatGPT” for eligible free, Plus, and Pro users in select English-speaking markets who have chat history and saved memories enabled. The optional, privacy-forward feature uses personalized visuals, awards, and creative outputs like poems and images to reflect how individuals used ChatGPT over the year, and is available on web and mobile.
Alibaba's Qwen Releases AI Model that Splits Images into Editable Layers Like Photoshop (1 minute read)
Alibaba's Qwen has released an AI model that automatically splits images into separate editable layers with transparent backgrounds, similar to Photoshop's layer system. The model allows users to independently resize, reposition, recolor, or replace individual elements, such as backgrounds, people, or objects, across 3 or 8 layers. Code and demos are available on GitHub, Hugging Face, and ModelScope for testing.
Luma Releases a New AI Model that Lets Users Generate a Video from a Start and End Frame (2 minute read)
Luma AI released Ray3 Modify, a new model that enables users to generate videos from start-to-end frames while modifying existing footage using character reference images that preserve the original performance. The model allows creative studios to transform human actors' appearances while retaining motion, timing, and emotional delivery, available through Luma's Dream Machine platform. This release follows Luma's $900 million funding round led by Saudi Arabia's Humain in November, with plans to build a 2GW AI cluster there.
Why AI is Exposing Design's Craft Crisis (13 minute read)
AI design tools exposed a technical literacy crisis that has been eroding designers' strategic influence since the early 2010s, when the industry collectively decided that designers don't need technical knowledge. When tools like Figma AI and Figma Sites launched with serious flaws, many designers couldn't evaluate whether outputs were functional or broken. Designers who maintain strategic relevance understand technical systems deeply enough to participate in product decisions, evaluate AI-generated work, and advocate for users in technical conversations.
The Art of Vibe Design (4 minute read)
AI has collapsed the gap between idea and execution, making it possible to build polished, expressive products through conversation alone, without traditional design or coding skills. The new bottleneck is taste: clear vision, judgment, and the ability to articulate what “feels right,” with AI acting as a force multiplier rather than a creative substitute.
Why AI Products Fail Without Proper UX? (5 minute read)
Many AI products fail because of a disconnect between their technological capabilities and user experience design, with 42% of businesses struggling due to poor market fit. The core issues include AI's "black box" nature, automation bias, blank-canvas paralysis from unclear guidance, workflow friction, hallucinations, and a lack of feedback loops for improvement. Good UX design can address these failures by positioning AI as an assistant rather than an expert, using confidence scores and transparency measures, providing templates and suggestions, ensuring seamless integration, acknowledging imperfection, and implementing user feedback systems.
Six Surprising Illustration Trends for 2026 (9 minute read)
Industry professionals predict that illustration in 2026 will emphasize handcrafted techniques such as woodblock printing and linocut, while artists bring authentic, personal perspectives reflecting diverse cultural experiences. Realism must evolve beyond technical accuracy to incorporate unique concepts, atmospheres, and emotional resonance that AI cannot replicate. The field calls for experimental aesthetics using new technologies, positivity-focused imagery, and vibrant color palettes featuring neon hues and bold combinations.
AI Image Generators Resort Back to the Same 12 Photo Styles, Study Calls It ‘Visual Elevator Music' (3 minute read)
Researchers found that when AI image models repeatedly generate and re-describe images, they rapidly converge on a small set of generic visual styles regardless of the original prompt. Across thousands of runs, outputs consistently settled into about a dozen recurring motifs, which the authors liken to bland, stock “visual elevator music.”
Apple UX Pioneer on Reviving Computer Desktop Design (5 minute read)
Scott Jenson, Apple's first member of the Human Interface group, argues that desktop UX has stagnated for 20 years despite its untapped potential for productivity.
This Designer Turned Road Material Into Stunning Furniture (3 minute read)
Designer So Koizumi created the "As" furniture collection using asphalt as a binding agent to connect metal, stone, and resin into stools, tables, and lighting fixtures.
A sharp tool can still ruin the cut (11 minute read)
The piece argues that judgment, taste, timing, and human friction matter more than ever, with AI best used as a partner after intent is formed, not as a substitute for thinking, learning, or wandering.
Curated tools 🖌️ , trends 🎨, and inspiration 💡 for design professionals
Join 225,000 readers for
one daily email