Is it still worth learning UI/UX as a Beginner?

Main idea

The speaker evaluates how well “latest frontier” LLMs (examples mentioned: Gemini, Codex, Opus) perform for UI/UX design versus coding. They conclude that UI/UX output is currently noticeably weaker than code output, and that beginners should use AI as an assistive tool, not a replacement for learning design fundamentals.

Model performance ratings (speaker’s estimates)

• UI/UX quality: ~6/10
  • Assessed on a scale where 0 = worst possible novice and 10 = best-in-class UI/UX designer.
• Code quality: ~7–8/10
  • The speaker argues code is the stronger area for current frontier models.

Evidence / analysis: reviewing “vibe coding” projects

The speaker examines real projects shared publicly (mentions Reddit, Facebook, and “vibe coding groups”) and critiques them the way a human UI/UX designer would.

Recurring issues flagged:

• Overly large or unnecessary branding elements
  • Example: a huge logo treated as a UX/spacing problem.
• Spacing, alignment, and layout problems
  • “Spatial issues” and misalignment are highlighted as detectable by trained designers.
• Desktop real estate not used effectively
  • Example: designs that look “mobile-only” even when displayed on desktop.
• Color palette inconsistency
  • Multiple color schemes and mismatched stroke/element colors are cited.
• Typography and hierarchy problems
  • Poor hero sections (unclear messaging), card typography issues, and overall readability/structure concerns.
• “Typical vibe coded design” look
  • Some projects are described as lacking anything “special” or thoughtfully designed.

Core point: even when prompted to be “the best UI UX designer ever,” AI can still produce UI that is aesthetically and structurally inadequate. Details like spacing, alignment, and typography often require expert refinement.

Guidance for beginners (learning / UI-UX takeaways)

• Beginners won’t be as good as current LLMs immediately, so AI can be used to speed up iteration and idea generation, not to “raw dog” everything without learning.
• Beginners still need to develop the “eye” to recognize problems like spacing, alignment, hierarchy, and palette issues—so they can improve beyond AI output over time.
• Re: “are designers cooked?”
  • The speaker argues no—entry-level designers can still succeed if they apply themselves and learn the craft.
  • A beginner can potentially become better than AI with effort and practice.

How the speaker personally uses AI in design workflow

The speaker says they are currently better than AI at UI/UX, but still use AI heavily:

• Use AI early for macro/high-level structure and rapid exploration.
• Once functionality works, fine-tune UI details (the polish stage).
• Position AI like an entry-level helper: fast at generating, but lacking expert refinement.

Product / channel mentions

• Mentions courses at designcourse.com, emphasizing that the channel focuses on maximizing AI usage while building specific human skills.
• Notes that AI becoming universally better than everyone soon is unlikely (“years in the distance”).

Main speakers / sources (as mentioned)

• Primary speaker: a solo UI/UX designer (narrator/reviewer assessing AI outputs)
• Referenced sources for examples: community-shared projects from Reddit, Facebook, and “vibe coding groups” (no specific account names given)