🔒Открыть пост
и посмотреть медиа
и посмотреть медиа
Researching Signals in the Age of ML & Personalization
User actions (clicks, saves) are signals, but meaning isn't obvious — a click could signal interest or just price-checking. The framework distinguishes: evergreen (durable), task-specific (situational), and live-moment (immediate). Goal: help systems avoid misinterpreting temporary behavior as permanent preference. Not every signal carries the same weight
Does AI Find Real UI Problems or Just Hallucinations?
AI found half the human-identified usability problems plus 11 extra. Of those 11, only 1 was real, 7 false alarms, 3 hallucinations. Bottom line: AI adds value as a junior researcher — can find real issues — but requires human oversight. 90% of AI-only problems needed correction
📹 NNG: Demystifying Clickbait - You’ll NEVER Believe What I Learned
Instead of understanding clickbait, writers often avoid anything associated with the practice, to the detriment of their writing
AI: From Fear to Empowerment - How to Supercharge UX Research with AI — a framework
A framework to move UX researchers from fearing AI to strategically delegating tasks. Four quadrants: Safe to Automate (transcripts), AI Assist → Human Refine (synthesis), AI Draft → Human Verify (screeners), and Keep Human Only (strategy, ethics). Use AI for heavy lifting, keep human judgment for high-risk core work
Case Study: Cineo — Discover movies through mood based curation
Cineo is a movie discovery app using mood-based carousels and social recommendations to reduce decision fatigue. Based on mood-congruence theory, visual mood cues replace traditional genre browsing. Features: community recommendations, "hidden gems" for underrated movies. Core insight: mood matching helps users decide faster
@uxdigest