Interviewing Candidates

With the school year rapidly coming to a close, I was reflecting on how fortunate we were to find such an amazing woman to teach our pod comprised of our boys and two of their friends who are matched in age. I feel fortunate because this woman who we entrusted 40 hours a week of exposure to our kids was chosen on the basis of 2 hours of interviews and some reference checks. She was selected from just 4 possible candidates, sourced mostly from There’s no way to objectively feel other than lucky.

If our odds were improved in any way, it was because of the other set of parents in our pod. I was blown away about how good they were at asking questions and reading the candidates. Yinh, who has now spent countless hours both screening and interviewing podcasts guests, felt the same. This couple is masterful in finding efficient questions to cut to the core of the candidates. It gave me so much appreciation for that skill especially when we consider the stakes.

Well, this week I came across Graham Duncan’s outstanding post about interviewing job candidates. It was one of those posts I added to my list of influential readings and re-factored for my own future reference. It’s full of practical advice but philosophically it really starts with the idea that interviewing is really a narrow application of a broader art.

From the intro:

The philosopher Kwame Appiah writes that “in life, the challenge is not so much to figure out how best to play the game; the challenge is to figure out what game you’re playing.”

When I try to figure out what game I’m playing, I see that for the last 25 years I have been playing a game of strategy applied to people, a game where over and over I try to answer the question “what’s going on here, with this human?”  In this essay, I make recommendations about candidate selection based on thousands of assessments I have made and my somewhat obsessive interest in the topic.

My goal in this essay is to help others make better decisions on a potential hire, business partner, or even life partner as quickly and as accurately as possible.  It’s made up of suggested action steps and some of the ruminations that underlie them. At the end I include my own assessment of different personality assessments and some of my go-to interview and reference questions.

My single favorite line is an idea I take seriously:

One of the greatest gifts we have for each other, for our children and spouses, for our teammates, is the positive feedback loop we can put someone into purely by believing in them, by seeing their genius and their dysfunction clearly and then helping them construct conditions for the former to flourish.

That emphasis is mine. I consider this to be one of the cheapest forms of human capital and this essay is ultimately about directing that capital.

Towards the end of the post you get a nice primer on personality tests as well as guides for conducting interviews and calling references.

Continue to my own takeaways and the link directly to Duncan’s post, Graham Duncan’s “What’s Going On Here, With This Human?” (Link)

Interviewing Engineers

I enjoyed tech founder Slava Akhmechet’s “super secret proprietary no-nonsense guide on how to interview engineers.” I am unqualified to know how valid it is, but it was fun to read. Find the full post here.


This guide is for interviewing very talented people.

It’s applicable if you’re building an extraordinary team at a hard technology startup. If your startup is technology-enabled, or you’re designing an interview process at a large company, or you’re hiring for well-established roles to do specialized tasks, this guide isn’t for you. You will not find references to “junior vs senior” or “front-end vs back-end” here. From the perspective of what we’re trying to accomplish, specialization is for insects.

There are three things you need to determine about a candidate: talent, judgement, and personality.

Think of hiring an engineer as if you’re buying a race car. The first thing you must look for in a race car is horsepower, because without horsepower the car is useless for racing. The horsepower of engineers is talent. Without talent, engineers are useless for building products, so it’s the first thing you must look for in a candidate. It doesn’t matter how nice the person is, or how hard-working. No horsepower, no race.Talent alone is insufficient. The world is filled with talented people who never get anywhere for a myriad reasons. Laziness, anxiety, fragility, impulsivity, egotism, victimhood, just to list a few. So once you’ve identified talent, you have to determine the shape and quality of its vessel. Where will the person direct their talent? And are they well-adapted to the demands of the external world?


  • Talent is a combination of speed, working memory, taste, knowledge of the toolchain, understanding how computers work, and ability to program. It’s IQ, but specialized for engineers. IQ is 50-80% heritable, impossible to improve, normally distributed, and strongly correlated with success in fields like science and engineering.
  • This matters for candidate selection because someone can improve within their talent band, but they can’t jump talent bands. A person with IQ of 145 (σ=3) is dramatically better than a person with IQ of 115 (σ=1). If you watch both people work, it’s like they’re from different galaxies. So your job as an interviewer is to find out the candidate’s talent band.


  • Judgement tends to be weakly correlated with talent, and comes down to this: there is a difference between a tinkerer and an engineer. They’re close, but they aren’t the same thing. Tinkering is building a Rube Goldberg machine for the sheer delight of building it. Engineering is discovering and satisfying (often unintuitive) constraints. The tinkerer works for the machine. The engineer makes the machine work for him.
  • Most engineers aren’t tinkerers, they’re in it for the money. Don’t hire those because they have no soul and hanging around them will slowly poison your own soul by osmosis. Conversely, many tinkerers aren’t engineers. Don’t hire those either, because they’ll build beautifully complex structures that serve no purpose other than their own existence. You want people who take great delight in building Rude Goldberg machines, but balance it with a broader sense of what they’re trying to accomplish.
  • Another way of thinking about it is that talent is a combination of general aptitude and programming tactics. Judgement is programming strategy.


  • The easiest way to think about it is in terms of the big five personality traits. These are kind of like Myers Briggs, except real. The three traits you especially care about are conscientiousness, agreeableness and neuroticism. Psychologists have precise technical definitions for these terms, but in plain language you’re trying to find out (a) whether the candidate is lazy or hard working, (b) are they an asshole, and (c) are they going to be stuck in analysis paralysis and invent life emergencies for themselves all the time instead of working.
  • The bad news is that you can’t find any of this information out until after you’ve hired the candidate. You can set up a low pass filter that might trap a few bad apples, but almost everyone is on their best behavior during interviews. Personal flaws rarely come out until long after the person is working for you. My sixth sense for picking out talent and judgement is pretty good. But for personality it’s only slightly better than random. So unless you have magic powers in this area, set up a low pass filter and later fire bad personalities as they reveal themselves.

Random observation

I’ve now noticed the respect for the Big 5 Personality Traits in multiple places. They’re beloved in tech world.  Slava said the Big 5 “are kind of like Myers Briggs, except real.” Marc Andreesen focuses on ‘conscientiousness’ trait in his interview on education (my notes here). And as we saw earlier, Duncan wrote “Within psychology, it’s the equivalent of gravity, and at this point, nearly everyone in academia finds it a useful mental model for personality.”

Personally, I scored around average for neuroticism, extraversion, and openness. I scored over the 90th percentile in conscientiousness (yay, allegedly) and agreeableness. Of course, as a trader I find agreeableness to be a backhanded compliment (see my post Being a Disagreeable Investor).

At any rate, when I hear the words “Big Five” I still think of going on safari.

More useful posts on interviewing:

  • Cedric Chin’s Using Head Fake Questions To Achieve Your Career Goals (Link)
  • First Round’s 40 Favorite Interview Questions from Some of the Sharpest Folks We Know (Link)

Leave a Reply