The Advice Problem

Sunday morning, March 1st. A founder is having coffee with a successful entrepreneur—a genuine success story, built and sold, now mentoring. The advice is confident and specific, delivered with the authority of lived experience: trust your instincts, hire slow, don't let perfect be the enemy of good. The founder nods, takes notes. He leaves feeling inspired. Two years later, having followed the advice carefully, he wonders what went wrong. The answer isn't that the advice was cynically given. It's that it was structurally unreliable from the start.

The people most positioned to give you advice—successful, experienced, willing to mentor—are systematically biased in four ways that make their advice unreliable. Not because they're bad people. Because of how advice structurally works. The bias isn't a personality defect you can screen for. It's baked into the structure of success stories. Learning to account for it is more useful than learning to find better advisors.

You Only Hear From the Survivors

The advice ecosystem has a selection problem. The people who talk about what worked are the people for whom something worked. The people for whom the same thing didn't work are mostly silent—they left the field, pivoted, gave up, or never built a platform.

When a successful founder tells you "I bet everything on my conviction and it paid off," they are describing one branch of a probability tree. They are not describing—because they don't know, because they weren't there—the much larger population of people who bet everything on their conviction and lost everything. Both groups tried the same thing. You're getting data from one.

This is survivorship bias. It's well-documented in finance, where mutual funds with bad track records quietly close; in medicine, where drug trials publish positive results more than negative ones; in military history, where Abraham Wald famously figured out that the planes returning from combat with bullet holes in their wings needed reinforcement in the areas with no bullet holes—because the planes hit there didn't come back.

It is almost never accounted for in mentorship conversations.

The entrepreneur who made it survives to give advice. The entrepreneur who followed identical advice and didn't make it is not in the room. His absence shapes the conversation as surely as the survivor's presence does.

Winners Misremember Why They Won

Even when advisors are genuinely trying to help, they're telling you their narrative, not the actual causal chain.

People do not accurately remember why they succeeded. They construct a coherent story afterward. The things an advisor remembers as pivotal are often not the things that actually drove the outcome. The startup founder recalls the key hire, the strategic pivot, the moment of conviction—not the favorable economic environment that lifted all boats, the warm introduction from a college friend that unlocked the critical meeting, the competitor that happened to implode at a useful moment.

Causality in a career is almost impossible to establish from inside it. Too many variables move simultaneously. Too much time passes between action and outcome. The mind, left to interpret this mess, does what minds do: constructs a narrative with a protagonist whose choices led to the result.

The narrative is usually psychologically satisfying—the founder trusted their gut, made the hard call, refused to give up—and causally suspect. The things people remember doing are systematically biased toward the things that felt meaningful, not the things that actually mattered.

Context Doesn't Travel With Advice

Advice gets transmitted without the conditions that made it work.

"Network relentlessly" meant something different in 1992 than 2025—different industries, different platforms, different gate structures, different norms around cold outreach. "Just ship it and iterate" means something different at an eight-person seed-stage company than at a regulated financial institution with a million users. "Trust your gut" applies differently to someone whose gut has been trained by fifteen years of domain experience than to someone whose gut mostly reflects optimism.

The advisor experienced the advice in specific conditions. Those conditions are not transmitted. The receiver applies it in their own specific conditions, which may or may not match. When the advice fails, this is attributed to the receiver's execution rather than to the mismatch between context and prescription. Which makes sense from the inside—the advisor knows it worked for them—but is epistemically wrong.

Generic advice ("work hard," "be persistent," "find your niche") is the extreme version of this problem. It's so stripped of context that it can't be falsified. It sounds like wisdom. It contains almost no information.

The Advice Giver Is Rewarded for Sounding Wise

Giving advice is rewarding. It confers status. It builds relationships. It makes the giver feel useful and the receiver feel validated. These are real incentives, and they're not aligned with giving accurate advice.

This doesn't mean advisors lie or consciously perform. It means that, at the margin, advice gets optimized for reception. Advice that is clear, confident, and actionable tends to land well and be remembered as good advice. Advice that is genuinely conditional—"it depends on five variables I can't assess from here, and by the way here are three cases where the opposite was true"—is harder to deliver and easier to forget. So advisors drift, without realizing it, toward advice that sounds wise rather than advice that is accurate.

The structural tell: most mentorship advice is unfalsifiable. It can't be wrong in a way the advisor would acknowledge. If you follow it and succeed, the advice worked. If you follow it and fail, you must have executed poorly, or the environment was against you, or you didn't have the founder's specific insight. The advice is never the problem. Advice that can't be disproved can't really be evaluated.

What Actually Helps

The most useful sources of guidance are usually not the most successful people in your field.

People who failed at what you're attempting. They've learned the failure modes from the inside. The successful advisor can tell you what worked; the person who tried and failed can tell you what breaks—which is a different and often more useful category of information. Failure narratives are rare because nobody builds a platform on having failed, but they're worth actively seeking.

Base rates rather than exemplars. "What happened to the median person who tried this?" is a better question than "What happened to the best person who tried this?" The median restaurant survival rate, the typical seed-to-Series-A conversion rate, the usual outcome for career changers in a given field—these numbers are harder to find and less inspiring than success stories, and considerably more useful for planning.

People five years ahead of you rather than twenty. Closer conditions, more recent information, less distance between their experience and yours. The person who navigated your industry five years ago knows things that the legendary founder from fifteen years ago doesn't—and their knowledge is more likely to still apply.

Specific adversarial questions. "Under what conditions would this advice be wrong?" forces the advisor past received wisdom into actual reasoning. "What did you try that didn't work?" gets past the highlight reel. "What assumptions am I making that you'd challenge?" surfaces your own blind spots rather than adding more confident advice on top of them.

The Takeaway

Most of the advice you've received from successful people has made you feel inspired rather than better positioned to succeed. That's not a coincidence—it's what the structural dynamics above produce.

The adjustment is not to stop seeking advice. It's to weight it correctly. Treat success narratives as one data point from a biased sample, not as a prescription. Actively seek out failure narratives, base rates, and advisors with incentives to be honest rather than impressive. Ask specifically about failure modes and limiting conditions rather than about what worked.

Good advice is usually conditional, messy, and delivered with uncertainty about whether it applies to your situation. If the advice you're receiving sounds clean, universally applicable, and maximally confident, that's often a sign it's been processed by the same filters that make it unreliable.

The most useful thing a mentor ever told me wasn't advice. It was a specific story about a decision they made that they still regret, and why they made it, and what they would need to believe differently to make the other choice. That's the information that actually transfers. Everything else is a story they've told themselves enough times to believe.