TLDR Design 2026-08-24
Apple Layoffs ๐, ChatGPT Photo Shortcut ๐ธ, Slack Vibe Coding ๐งโ๐ป
Apple lays off 200+ people across Vision Pro and Siri teams (3 minute read)
Apple has laid off more than 200 employees across its Vision Pro, Siri, and software engineering teams as part of a broader restructuring effort that includes scaling back Vision Pro gaming and immersive video initiatives while reaffirming its long-term commitment to visionOS and future hardware. The company is also reorganizing Siri and AI-related teams around its new Siri AI platform, eliminating some roles, creating new ones, and shifting resources to accelerate development of AI-powered features across its products.
ChatGPT for iPhone now lets users grab recent photos with a long press (2 minute read)
OpenAI has added a new iPhone shortcut that lets ChatGPT users long-press the + button to instantly access and attach their four most recent photos, streamlining a process that previously required several taps through the Photos menu. The update reflects the growing importance of image-based interactions with AI and joins other recent improvements aimed at making photo uploads, visual questions, and longer conversations faster and more seamless.
Slack has (of Course) Launched a Vibe Coding Tool (2 minute read)
Slack Code is a feature that lets teams tag AI coding agents like Claude Code, Devin, or GitHub Copilot directly in group chats to build and ship software fixes. Available now on all Slack plans, it requires human approval for high-stakes actions like production merges and archives the channel once work is approved. The launch continues Slack and Salesforce's broader push to market AI agents as autonomous digital coworkers, following prior products like Slackbot and Agentforce Coworker.
Figma is the New Dreamweaver: How Modern Prototyping Tools are Trapping Us in 2018 (6 minute read)
Figma excels at visual design and collaboration, but treating it as a faithful representation of the web can lead designers to overlook how interfaces actually behave in real-world conditions. Static artboards and Auto Layout can mask challenges such as responsive behavior, loading performance, accessibility, unpredictable content, and dynamic states. As products become increasingly adaptive and AI-driven, designers need to think more like systems architects, testing ideas in code, designing for edge cases, and grounding design systems in production-ready components rather than purely visual artifacts.
One AI Output is an Example, Not an Evaluation (11 minute read)
A single AI output only demonstrates a system can perform a task onceโit cannot show how reliably that performance holds across different situations. Rigorous evaluation, much like quantitative UX research, requires testing multiple representative inputs across repeated runs, then reporting averages with confidence intervals to capture both test-input and run-to-run variability. Systems with identical overall scores can fail in very different waysโconsistently on specific inputs versus unpredictably across all inputsโso understanding failure patterns matters as much as the aggregate score itself.
How to Check Your AI System Still Matches Your Values (8 minute read)
As AI becomes more deeply embedded in products and services, designers have a growing responsibility to help ensure that technology reflects human values, ethics, and societal needs. Building trustworthy AI requires close collaboration between design, research, engineering, and governance, with values such as privacy, transparency, and safety translated into concrete decisions throughout the development process. A human-centered approach treats design as the connective thread that helps teams align strategy, implementation, and outcomes while continuously evaluating whether systems deliver on the promises made to users and society.
Watch Susan Kare revisit the icons and graphics for the original Mac (2 minute read)
Susan Kare reflects on designing the original Macintosh's iconic visual language and shares lessons on creativity, problem-solving, and the lasting impact of thoughtful interface design.
A Sloppy Interface is a Security Liability (2 minute read)
AI now makes it trivial to fake polished interfaces, as seen in phishing attacks like the one behind the Axios npm incident, raising interface craftsmanship into a security consideration. High-fidelity design detailsโmicro-interactions, loading behavior, UI voice, edge-case handlingโare hard and costly to replicate, making sloppy products easier targets for imitation-based attacks. Investing in above-average craft therefore serves as both a brand asset and a practical layer of defense against exploitation.
Vibe Coding and Quality (2 minute read)
Vibe coding was once dismissed as producing only prototype-quality, "duct-taped" results unsuitable for reliable software. A Mac app built with Codex reportedly outperformed existing human-made alternatives within a day, with clear room for further refinement. This experience led to the conclusion that achieving exacting standards of quality is realistically possible only through vibe coding, not conventional programming.
Curated tools ๐๏ธ , trends ๐จ, and inspiration ๐ก for design professionals
Join 320,000 readers for
one daily email