Taste as Shedding Social Proof
Real taste is recognizing what's great before results prove it — by evaluating ideas stripped of their social and authority proof
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 3
- Confidence
- 90%
Doshi reframes 'taste' from polished product output to the ability to identify what is really good without needing to see its results. The practice is to deliberately strip an idea of its social proof, authority proof, catchy metaphors, and impressive-looking math, then judge the idea on its own merits — which is what actually makes you a first-principles thinker rather than someone who just claims to be one.
Origin
Shreyas Doshi; he uses 'game recognize game' and the Jensen Huang / Nvidia example (obvious as genius in 2024, hard to call in 2010) to illustrate taste as pre-result recognition.
Core principles
- 01Taste is identifying what's really good WITHOUT needing to see its results
- 02Calling a proven winner great (Jensen in 2024, Mahomes today) requires zero taste
- 03Being a harsh grader who dismisses everything also requires zero taste
- 04We are unduly swayed by catchy metaphors, alliteration, authority, and complex charts we don't understand
- 05Everyone thinks they're a critical / first-principles thinker; almost no one actually is
How to run it
- 1
Audit what you take in, not just what you output
Turn the taste question inward: examine how you choose what to believe, how you learn, who you learn from, and what content you resonate with — because those beliefs dictate how you manage, lead, and decide.
Watch out You can have great taste in UI flows and still have poor taste in the beliefs you adopt.
- 2
Spot your susceptibility triggers
Notice when you're being persuaded by a catchy metaphor, an alliteration ('fail fast', 'fast follow'), an authority ('Bezos said it'), or complicated charts and math you don't actually understand.
Pro tip Test an idea by renaming it plainly — would 'fail quickly' attract you as much as 'fail fast'? If not, you were seduced by the alliteration, not the idea.
Watch out Some product leaders weaponize complicated charts and math specifically because it bypasses critical evaluation.
- 3
Evaluate the idea stripped of its proof
Deliberately separate the idea from its social proof and authority proof and judge it on its own logic. Practice recognizing quality before the results are in — 'game recognize game' in the practice session, not after the game is called.
In the wild
Doshi argues calling Jensen Huang a genius in 2024 requires zero taste because you can just look up Nvidia's stock price — but Huang didn't change much between 2010 and 2024, so recognizing it in 2010 is where real taste lived.
→ Illustrates that taste is pre-result recognition, not post-result acknowledgement.
The identical idea framed as 'reversible and irreversible decisions' didn't catch on, while 'one-way door / two-way door' did — the only difference being the catchier metaphor and Bezos's authority.
→ Demonstrates how metaphor and authority, not merit, drive which ideas spread.
Common mistakes
Confusing cynicism with taste
As people get senior and successful they become 'stuff graders' who dismiss everything as crap — but blanket negativity requires zero taste and signals the opposite of discernment.
Claiming first-principles thinking without shedding proofs
Everyone says they're a rigorous first-principles thinker, but until you strip ideas of social and authority proof you're just pattern-matching on prestige.
Is it for you?
Best for
Experienced leaders who want to sharpen judgement and stop outsourcing conviction to authority and hype
Not ideal for
Domains where you lack any grounding to evaluate ideas independently and social proof is a reasonable heuristic
From the transcript
“taste is about the ability to identify what is really good without needing to see its results”
“game recognize game before the game is called”
“we get overly excited about cool metaphors”
“how many of us love fail fast”
“what if that thing were called fail quickly it's the same meaning do you think you would be as attracted to that idea”
“I kind of have to evaluate the idea separate from all of its like social proof and Authority proof”
“we become these stuff graders right like I don't like anything”
From the episode
4 questions Shreyas Doshi wishes he’d asked himself sooner
Former PM leader at Stripe, Twitter, Google