People’s advice for others often fall into two categories: general platitudes or personal anecdotes. The former are mostly too vague to be practically useful, while the latter are likely to lead people astray as the circumstances may differ significantly. In statistical terms, they are forms of under and over-fitting respectively, and can manifest through many different ways. Post-hoc justification for example, tend to draw over-fitted conclusions. When the outcome is given, anything that happened can be seen as an important contributing factor, and while the conclusion may be correct in this specific case, it is usually not generalizable.

This is why asking “what would you do in my situation” is one of the better ways to seek advice. It protects against badly fitted guidance by explicitly forcing it to fit the current relevant situation. However, this type of advice is very costly, both to the seeker and the giver. It requires that the circumstances and necessary context be explained adequately, and demands a certain level of intellectual investment from the advisor. In most cases, ain’t nobody got time for that, so people default to philosophized tropes or telling their own stories, resulting in poor advice.

Because of the effort required, well-fitted advice cannot be scaled easily. Since answering “what I would do in your situation” is essentially a form of outsourced thinking, to scale this is equivalent to scaling the mind of the mentor. Under or over-fitted advice does not suffer from this problem, and as a result, anything distributed via one-to-many broadcasts - large mentorship programs, talks, blog articles, etc… - is limited to these kinds of lower quality advice.

This is why even top-tier startup accelerator programs (such as Y Combinator) decrease in quality as they scale. Part of it is the inevitable regression to the mean when letting more people into the program, but a bigger reason is the impossibility to scale good mentorship. In the early days of YC, when there were only a dozen or so companies in the program at once, there was a much greater availability of the mentors’ time. Each company was able to “borrow” a significant share of these expert minds to help make decisions. In a way, these companies effectively had the mentors, including people like Paul Graham, as part-time co-founders in their early days. Now, with hundreds of companies in the program batch, this dynamic is all but a distant memory.

When receiving well-fitted “what would you do in my situation” style advice, there are two common scenarios. One is when you are looking for genuinely new ideas to help solve a difficult problem, and the other is when you are looking for encouragement on something that you already knew but have trouble committing.

The first is about outsourcing intellect. Since you are looking for new ideas, in this case, divergence in thinking is key: you want someone who is intelligent in a very different way from yourself. It’s also critical to have the mentor not just provide their answer, but also the why. By walking through their decision-making process, you can get a tiny glimpse into their psyche. The answer itself will be over-fitted to this particular situation, so while it is very useful now, it won’t be as useful in the future. The mental process on the other hand, is more generic, which can be incorporated into your own thinking and reused later.

The second is ultimately a form of outsourced therapy or self-help. You aren’t looking for anything radically new, but are still nonetheless stuck with a problem because of your inability to carry out the solution (perhaps due to doubt or lack of will). In this scenario, having the appropriate relationship with the mentor is crucial. The person needs to be willing and able to actually compel you to act.

Regardless of the type or situation, getting high quality advice requires a high degree of trust and familiarity from both sides. The ability to solicit well-fitted advice whenever needed is probably highly predictive of success in general. It depends on both the quality and number of people that one has good relationships with, relationships at the level where the high mental cost interactions needed to give good advice could occur. High quality relationships with extremely competent people is one of the most valuable and scarce resources that one can possess, precisely because of its lack of scalability - everyone only has so much time and mental capacity in a day.