Design, knowledge visualization and AI content taste

Designers discuss restraint in AI-generated visuals and the role overlap between design and code. - Advice: use knowledge visualization sparingly and only for complex or unfamiliar concepts - Overuse of AI visuals can fatigue audiences; creators should self-check repeat viewings - Emphasis on taste, deep thinking, and long-term intellectual investment over mass AI output - Debate on whether designers should code reframed: ask what the organization expects of design roles - Karri Saarinen's method: question the problem, explore concepts, then execute; warns against skipping slow thinking - Design tools shape behavior; shifting the default to code might suppress pre-execution conceptual work
Key insights
Designers urged to be restrained with AI-generated visualizations
Design commentary emphasizes that AI-generated visuals (comics, infographics) should be used only when text alone cannot explain complex or unfamiliar topics. Overuse risks audience fatigue; the test suggested is whether the author would re-open the content multiple times. The recommendation stresses taste and long-term attention investment.
The question 'should designers code?' is reframed to focus on expected design value
Karri Saarinen argues the debate is misframed; instead of asking whether designers should code, teams should ask what they expect designers to deliver. He prescribes a three-stage method—question the problem, explore concepts, then execute—and warns that tools influence behavior and can erode 'slow thinking' if execution is too easy.