Real teams, real workflows: what senior designers at major companies actually changed.
Designers at major companies are using AI for research synthesis, generating design variants, writing microcopy, and automating asset preparation. Most report AI saves time on routine tasks but still requires significant human oversight and refinement to meet quality standards.
A concrete look at moving AI-generated code back into Figma for collaborative editing.
Figma built a tool that converts Claude-generated production code back into editable Figma components, preserving layer structure and design tokens. The workflow lets developers generate functional code, then hand it to designers for visual refinement without starting from scratch.
A well-curated NN/G reading map for practitioners figuring out where AI fits in UX.
NN/G categorizes AI use cases across research, ideation, prototyping, and testing, with links to studies on each. The guide emphasizes when AI accelerates work versus when it introduces quality or ethical risks that require careful human judgment.
Honest breakdown of Lovable, Cursor, and Bolt, including the hidden costs they don't advertise.
Lovable, Cursor, and Bolt let you prompt for full apps but lock you into their hosting, charge for every iteration, and produce code that's hard to customize or migrate. The tools excel at prototypes but struggle with auth, complex state management, and maintenance over time.
A grounded look at what Bolt, V0, Lovable, and Figma Make actually change about who designs software.
Bolt, V0, Lovable, and Figma Make lower the skill floor for building interfaces, letting non-coders ship functional prototypes. They don't replace professional designers but shift the work toward prompt crafting, component curation, and quality control rather than pixel-pushing or writing basic markup.
Survey-backed reality check on how practitioners actually feel about prompt-to-code workflows.
Survey data shows designers and developers see prompt-to-code tools as useful for rapid prototyping and exploring ideas, but most wouldn't trust the output for production without significant rework. Concerns center on code quality, accessibility gaps, and difficulty maintaining AI-generated codebases long-term.
Show 1 reply