Turn Advice Into a First-Principles Framework
Don't collect rules — ask why, triangulate three people, and rebuild the reasoning underneath.
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 4
- Confidence
- 85%
Most advice is anecdata: someone extrapolating a hard 'never/always' rule from one or two experiences. To use advice well, interrogate the reasoning behind it, gather it from multiple sources, and reconstruct the underlying framework so you can apply it with nuance instead of following a brittle rule. Confidence of delivery is not correlated with accuracy.
Origin
Bret Taylor's reflection on losing FriendFeed to Twitter partly because he sought advice from the wrong people, plus his general model for evaluating advisors.
Core principles
- 01Confidence of an opinion is not correlated with its quality — sometimes inversely
- 02A single person's advice is almost never statistically significant
- 03Advice delivered as a rule usually hides a single backfiring experience
- 04Good judgment about people is itself the meta-skill you're hiring and cultivating
How to run it
- 1
Ask who to talk to, not just what to do
Beyond asking for advice directly, ask people whom you should talk to for good advice. Common names recurring across sources is a strong signal of genuine quality.
- 2
Interrogate the why like a two-year-old
Don't stop at the recommendation — keep asking why, why, why, to surface the actual experience and reasoning that produced it.
Pro tip Understanding that a 'rule' came from one backfiring incident lets you weight it correctly instead of obeying it blindly.
Watch out A useful anecdote presented as a universal law will mislead you if you never ask where it came from.
- 3
Triangulate across three people
Gather advice from multiple people on the same question. When several report similar underlying experiences, you can synthesize a first-principles framework from which the advice emerges.
- 4
Apply with nuance, not as a rule
Use the reconstructed framework so you apply judgment with context, rather than following an inherited rule you don't understand.
Watch out Discount eloquence — the most fluent, confident statements in a field you know well are often the least accurate.
In the wild
FriendFeed (11 of 12 employees were engineers) out-shipped Twitter, had more features and 100% uptime during Twitter's 'fail whale' era, yet lost when Twitter loaded celebrities like Oprah and Ashton Kutcher. Taylor was myopic, polishing product, and wasn't soliciting outside-in advice on distribution.
→ He concluded the missing ingredient was good advisors and boards asking what he couldn't see — 'the advisors who actually tell you what you need to hear' — and became a strong believer in boards.
Common mistakes
Treating one anecdote as a rule
Someone's 'never do X' often comes from a single bad experience; obeyed as law it removes nuance and can steer you wrong in a different context.
Weighting advice by confidence
The most eloquent, self-assured statements are frequently the least accurate; selecting advisors by charisma rather than judgment corrupts your inputs.
Is it for you?
Best for
Founders and operators building an advisory network and deciding whose input to weight
Not ideal for
Truly statistically-grounded domains where an expert's advice reflects large samples (e.g., Buffett on investing) rather than anecdote
From the transcript
“there's not a strong correlation between the confidence uh with which someone expresses an opinion and the quality of that opinion”
“don't just ask what (29:30) to do, but why. Like be it like an obnoxious 2-year-old kid, you know, why why why?”
“if you ask advice of three people and they all have very similar interactions, you can create kind of like a first principles framework from…”
“no one's advice is statistically significant or very rarely is it”
From the episode
He saved OpenAI, invented the “Like” button, and built Google Maps: Bret Taylor on the future of careers, coding, agents, and more