A junior analyst gets a coffee meeting with a senior partner. The partner tells her the key to success is developing a specialty — that's what made the difference for him. Two months earlier, the same partner gave identical advice to a different junior analyst. Last year, he gave it to three others. His advice is not wrong, exactly. But it has nothing to do with any of these people's actual situations. It is a story about him.

The mentorship industrial complex has a problem it doesn't want to name: most mentorship is bad advice delivered confidently by people who have no particular reason to give good advice.

This sounds harsh. But the alternative — pretending that experience reliably generates transferable wisdom — produces something worse: people who follow advice meant for someone else's life, handed down by someone whose incentives are more complicated than they appear.

The Narrative Transmission Problem

When a mentor explains how they became successful, they are doing something cognitively difficult: reconstructing a causal story from a sequence of events. The events are real. The causal story is largely constructed after the fact.

Psychologists call this the "hindsight bias" — the tendency to perceive past events as more predictable than they were. Once something has happened, our brains reassemble the sequence of decisions into something that looks like a clear path. The detours become "learning experiences." The lucky breaks become "positioning." The random advice that happened to be followed becomes "the key."

What a mentor transmits is not their decision-making process. It's the narrative they've constructed to explain their outcomes. These are different things. The narrative is optimized for coherence and meaning, not for accuracy. And it is almost entirely about them.

This is not cynicism about mentors. Most are genuinely trying to help. The problem is that we are not actually very good at extracting principles from our own experience. We extract stories. And our stories are first-person.

Survivorship Bias With a Face

Mentors are, by definition, people who succeeded through some path. The people who tried the same path and failed are not sitting across from you at coffee. They are not writing the business books. They do not have the LinkedIn follower counts that get them invited to speak.

This is survivorship bias at the interpersonal scale — and it is more pernicious than the statistical version because it has a human face, a compelling narrative, and the implicit social weight of the mentor's status.

When a mentor tells you to "take risks early in your career," they are speaking from a sample of one — and a selected one at that. The people who took risks early and failed are, on average, less likely to become your mentor. The confirmation bias is baked into the selection process for who ends up doing the mentoring.

Philip Tetlock's research on expert prediction found that domain expertise correlates weakly with forecasting accuracy in complex domains. The relationship between "has succeeded in a domain" and "can explain why they succeeded in a generalizable way" is weaker still.

A surgeon who has performed ten thousand successful operations has genuine expertise in surgical technique. That same surgeon has no particular expertise in advising young surgeons on career strategy — except through the lens of their own specific career, which may or may not generalize.

The Status Dependency Problem

Here is the part that is genuinely awkward to say: mentors have incentives that are not perfectly aligned with your interests.

A mentor's status, at least in the context of the mentoring relationship, depends on their advice being valuable. If you follow a mentor's advice and succeed, the mentor's status is confirmed. If you ignore the mentor's advice and succeed anyway, the mentor's status is not. This creates subtle — usually unconscious — pressure toward confident prescription over honest uncertainty, toward advice-giving over question-asking, toward direction over exploration.

The power differential reinforces this. You are less likely to push back on advice from someone more senior than you. They are less likely to notice when they're wrong because the feedback loop is weak — you probably won't return to the coffee meeting six months later to report that their advice didn't apply to you.

This is not a critique of individuals. It is a structural problem with the mentorship form. The incentives don't line up well for the production of accurate, tailored advice.

None of this applies to sponsorship — where someone with power actively advocates for you in rooms you're not in. Sponsorship is structurally different and substantially more valuable, because the sponsor's reputation is at stake in a way that sharpens their judgment about your actual capabilities.

What Actually Transfers

Not all mentorship is equally bad. Some types of knowledge transfer reliably.

Tactical, specific knowledge — how to navigate a particular organization's politics, what a particular skill actually requires in practice, what interviewers in a particular field actually care about — transfers well. This is knowledge that doesn't require generalizing from biography. It's more like local information.

Network access — introductions, referrals, endorsements — transfers extremely well and is probably the most valuable thing a senior person can offer a junior one. This requires the least interpretation and the least inference about whether the mentor's experience maps to yours.

Challenge without direction — the mentor who asks hard questions, points out where your reasoning is weak, and refuses to give easy reassurance is more useful than the mentor who tells you what they would do. You are not them. What they would do is only weakly predictive of what you should do.

Pattern recognition within a narrow domain — an experienced negotiator who tells you what they see in your specific negotiation script, a senior writer who marks up your actual prose, a veteran product manager who reviews your specific roadmap. Expertise applied to your specific artifact is different from expertise applied to the general question of how you should live your professional life.

How to Extract Value Anyway

Given all this, the corrective is not to avoid mentors. It's to use them differently.

Treat mentor advice as data, not instruction. When a mentor says "you need to develop a specialty," the useful question is not "should I develop a specialty?" but "what is this person seeing in my situation that makes them think this?" The advice is a signal about what they notice. It doesn't determine what you should do.

Ask for stories, not prescriptions. "What did you do when you faced a situation like this?" produces more useful information than "what should I do?" The story is real. The prescription is a narrative construction.

Seek out mentors with different paths. If your five mentors all succeeded through the same general route, you're getting five versions of the same survivor testimony. Deliberately seek advisors who took different paths, including ones that look unusual from the outside.

Weight tactical advice heavily, strategic life advice lightly. The mentor who tells you how to handle the specific conversation you have next Tuesday is much more valuable than the mentor who tells you what to optimize your career for over the next decade.

Separate access from advice. The most valuable thing a senior person can offer is often a door, not a direction. Be explicit about asking for introductions, referrals, and endorsements. Don't outsource your strategic thinking in exchange for these.

Takeaways

The failure mode of mentorship is this: you get someone else's autobiography dressed up as universal principle, and you follow it because they are senior and confident and it would be rude to push back.

The corrective is to treat mentor advice as you'd treat any other single data point — useful as one input among many, requiring interpretation, subject to the source's particular biases and experiences. Mentors who succeeded through a specific path will, reliably and non-maliciously, recommend that path. Your job is to figure out whether the path fits your terrain, not theirs.

The most useful mentors are the ones who seem to know this about themselves — who say "here's what I did, but I'm not sure it generalizes" and then ask you hard questions instead of answering them. These people are rarer than the confident prescribers. When you find them, hold on.