It’s not going to be easy

Michel Trottier-McDonald
so many slugs
Published in
4 min readJul 19, 2016
24094635266_ffe4ef3865_k
Muir Woods, January 2016. That’s easy. Understanding backpropagation in feed-forward neural networks isn’t.

Seriously, it won’t be. I know you’re saying that just to reassure me, but all you’re doing is giving me very poorly adjusted expectations.What am I talking about?

3 years ago I was doing a summer school on particle physics. I knew from people who took it before that it was hard. Like, very hard. The program organizers were aware of this, and they tried working with the instructors to diminish the workload. Confident that it would be more tame this time around, they prefaced the school by telling us to expect 30 min of homework for each hour of lecture. It turns out they were still off by a factor of 8. In the bad direction.

Quickly realizing the mountain of work that was piling up ahead of us for this short 2-week program, we started stressing out and being defensive. The instructors themselves kept reassuring us that the next problems they were gonna drop on us would be easy and doable in a short amount of time. They kept on being wrong over and over again.

We were supposed to be all wrapped-up at the end of the 2 weeks. The school organizers however decided to postpone the due date for all assignments for an extra 2 weeks. When the final deadline finally arrived, many of us still didn’t manage to get all the assignments in. Actually, I think only one participant managed to do everything. We all learned a lot, but it wasn’t as fun as it should have been. I don’t think any of us will be excited to do a summer school ever again.

This wasn’t my first time (and far from my last) seeing people outrageously underestimate how hard it is for other people to solve certain problems. It happened in class throughout my education, and it’s happening in interviews now. Why is this kind of situation so prevalent in technical fields?

I believe the reason this keeps happening is that there exists a psychological trap that we fail to recognize way too often. Once you’ve solved a problem, you forget the path you took and the time you spent to arrive at the solution. When you give the same problem to someone else, you think it will take them just somewhat longer than it would take you to solve it. You underestimate dramatically the legwork it takes just to understand how the pieces of the problem fit together, let alone solve it. You’re failing to take into account your privilege of knowing the solution.

The cause behind this may be impostor syndrome. Maybe you think people are smarter than you are, so they’ll be able to do it faster than you. If that’s the reason, it’s redeemable. But you should be keenly aware that you might be passing your own impostor syndrome to the next generation. Your students will feel thoroughly inadequate as soon as they realize they can’t solve the problem within the time frame you gave them. A few students might still make it, but it will only make the others feel even worse. I don’t think that’s what you want to achieve.

Another, nastier reason for you to think the solution (which you know) is easy to arrive at might be pure negligence. You might not care about giving your students realistic expectations, and you might not care about how they feel about themselves. You might actually think they ought to complete the challenge in the time frame you gave them. It might be your way of picking out the best of the best. If that’s the case, never put yourself in a mentoring position again, you’ve probably done enough damage already. Selecting out fragile egos instead of giving them a chance to get stronger is… We need more smart and educated people, not less.

An even worse reason is because it may flatter your ego to say that a particular problem is easy and that it shouldn’t take long to solve. You know deep inside it’s a hard problem, but it’s easy for you now and you’re proud of it. It’s fine to be proud of what you can do, but it’s not fine to use that to position yourself above others (I thought of editing that sentence out since it sounds so trivial when you say it out loud). If you don’t understand that mentoring is not about you but about your students, the last part of the previous paragraph apply to you.

One of the first thing they told us at Insight Data Science was that getting through the program was going to be really hard. We were going to struggle, get frustrated, and feel inadequate. The deadlines for each milestone looked absolutely ludicrous up front. Yet, everyone made it through on time and ended up feeling terrific about their project. The program works extremely well. You should see the cutting edge things people accomplished. Better yet, everyone comes out of it with their self-esteem intact, if not invigorated.

What I’ve learned is that it very rarely helps to try to reassure people by telling them a hard thing is going to be easy. It won’t be. But if you take on the challenge anyway and suffer through the setbacks, you will succeed and you will feel damn good about yourself when you do. As a mentor, what you want at the end of the day are students who have the courage and confidence to tackle any challenge. Do not get in the way of that.

--

--

Michel Trottier-McDonald
so many slugs

ex-particle physicist turned data scientist who spends way too much time reading about North American politics