Exposure Time for Taste
Deliberately spend more time learning than building to develop the judgment AI can't give you.
- Difficulty
- Easy
- Time to result
- ~ongoing to results
- Steps
- 4
- Confidence
- 88%
In a world where anyone can produce 'good enough' with AI, the scarce, rewarded skill is judgment and taste — knowing what world-class and magical look like. You build that by deliberately budgeting more time to learning than to building: reading the agent's output to learn how it thinks, and immersing yourself in exquisite design and copy from world-class creators so your standard of 'good' rises.
Origin
Lazar credits the 'exposure time / exposure hours' concept to Lenny Rachitsky (who in turn attributes it to guest Jason M. / GM), and pairs it with the observation that 'the ceiling on the AI isn't the model intelligence, it's what the model sees before it acts.'
Core principles
- 01AI amplifies whatever you bring: 'if you don't know what you're doing, you're just going to produce garbage faster.'
- 02The gap that matters shifted — everyone now clears 'good enough,' so the reward moved to 'world-class and magic.'
- 03You can code garbage fast just as easily as you code magic fast; the input and judgment are what differ, not the time.
- 04Taste is a muscle built by exposure to great work, not by talent.
- 05Reading agent output (not code) teaches you what's possible and how the tools reason.
How to run it
- 1
Budget more time to learning than to building
Deliberately set aside more hours for consuming great work and reading agent output than for shipping. Building is now fast; judgment is the constraint.
Watch out Don't over-index on learning to the point of never shipping — you also get better by building. Do both, but weight toward learning.
- 2
Read the agent output to learn how it thinks
Ignore the code; read what the agent tells you it's doing and reasoning through. This teaches you what the tools can and can't do so you know what's genuinely possible.
Pro tip Watching the agent conversation is the emerging layer 'over code' — treat it as your primary teacher.
- 3
Immerse in world-class design and copy
Follow elite designers, subscribe to design newsletters, and study tools where great design is produced — even follow the individual creators who post long videos of themselves designing and prompting.
Pro tip Watch a world-class designer talk to the tool in real time (40-50 minute videos) to absorb how they prompt for quality.
- 4
Deconstruct excellence to recalibrate your standard
When something looks great, open it up and see what it actually took. Lazar tried to lift a 'simple' gradient from a Figma file and found it was 50 layers of colors at varying opacities — resetting his sense of what world-class requires.
Pro tip Build a reference app of styles and their prompts (he built a public one with ~18 design styles) so you can learn and reuse the vocabulary you lack.
In the wild
Lazar went into a Lovable designer's Figma file to copy a background he assumed was a simple three-color gradient. On inspection it was ~50 colors with different gradients and opacity levels.
→ The discovery exposed the gap between his eye and world-class execution and became a core lesson in why exposure to elite design matters.
As a non-designer who didn't know terms like glassmorphism, Lazar built a public Lovable app cataloguing ~18 design styles with the prompts to replicate each.
→ He turned his own lack of design vocabulary into a reusable learning tool that teaches the styles and their prompts.
Common mistakes
Optimizing for raw output speed
People rush to produce faster, but 'we won't be rewarded in the world of AI for faster raw output, we will be rewarded for better judgment.' Speed without taste just floods the crowd with more good-enough slop.
Trying to learn to code instead of learning taste
For non-coders, Lazar advises against learning to code now — it's optimizing for the wrong, soon-commoditized skill. The durable investment is design, judgment, and clarity, not syntax.
Is it for you?
Best for
Non-technical builders and PMs who can already ship with AI but whose output stalls at 'good enough' and needs to reach world-class.
Not ideal for
People who haven't built anything yet — they should start building first; taste-building compounds best once you have hands-on reps.
From the transcript
“Set aside more time on learning than building”
“the ceiling on the AI isn't the model intelligence. It's what the model sees before it acts”
“you can code fast but you can code garbage fast as well as magic fast”
From the episode
The rise of the professional vibe coder (a new AI-era job)
Lazar Jovanovic (Professional Vibe Coder)