Recruiting, fast and slow

Photo by Ahsan Avi on Unsplash

TL;DR Maybe we’re doing something wrong

First of all, let me break the bad news: you are discriminating.

As long as your recruiting ad contains any other words than “I’m looking into hiring a human being”, you are discriminating. Change my mind.

There’s a lot of emulation out there about stopping discrimination in the hiring process, about being an equal-opportunity employer. Which is good. Absolutely great. Nobody should be judged on criteria like skin colour, gender, ethnicity, etc. Everybody deserves a chance. What I find puzzling in this is the fact that most of the “features” used in fight against discrimination are extrinsic. Nobody talks about the mind. It’s almost the other way around.

First, you go through the screening process, then a few technical interviews, then maybe the values interview. NB, I’m talking here about my view of the world, IT hiring, but I doubt the process is much different in other fields.

So what’s the purpose ?

I’d argue that we are checking your level to fit our echo chamber. We’re looking for our like-minded peers. While echo chambers are not necessary the personified evil (see https://www.economics.ox.ac.uk/materials/jm_papers/921/echochambers.pdf), I think that in the current context, and looking at the future, they are not exactly the best thing since sliced bread.

Let me try to explain.

Since the industrial revolution, mechanical and repetitive labour has been constantly under replacement by smarter and smarter machines. Lately, with the advent of deep learning, the process is being accelerated a lot. True, we won’t be replaced in the next 5 years. But I think it’s an exponential curve, and you can find the writings on the wall all around us: self driving cars, AlphaZero, specialised AIs being better at interpreting legal documents, X-ray imagistic, etc. Software may seem untouchable from an outside view, but wait… there’s more.

Facebook has already in production a couple of AIs that assist developers in writing code and in fixing bugs. Microsoft has bought Github, so they have at their fingertips a huge codebase for training purposes, and in case you don’t know, when you have enough training data (read that huge), the algorithm you use is (almost) irrelevant. Imagine the existing system that can generate images from a textual description, feed it with enough code, and maybe he’ll spit up a software component, based on a concept’s text description.

No, we are not there. Yet. Look at TabNine. It’s scary. Almost like a mind reader.

Until now it has been Marc Anderssen’s “software is eating the world”, I think that just around the corner lurks “software is eating software”.

Why this digression ? Well, I’m surely biased, but I still think that writing software is a creative activity. Meanwhile, machines are getting better by the day in copying human behaviour. And what we are doing ? We (mostly) hire the ones that are close to our mindset. We build our own company-wide echo chamber. We seek for gender diversity (again, very happy to see that), but for mindset uniformity. We are throwing out people with ideas that don’t look like ours.

I would argue that not very far in the future this would be the most destructive behaviour for any project out there. We are moving more and more towards the creativity-driven world, yet we are not ready to accept different views. Unless the person trying to do that is willing to put a considerable effort in fighting our inherent confirmation bias (yes, I’m human, I have biases. We all have. I’m trying to be as aware as I can about them, and try to fight them as much as I can. You ?)

True, there’s also the part where the lateral thinker is at the top. But have you ever wondered how many Jobses lurk unknown in the belly of Apple, because they have been shut down by their own echo chambers ? (Apple and Steve used here as examples only)

So, what to do about it ? I don’t know. And I’m not in the position of knowing or doing. I just watch the game and counting points.

Maybe we should reconsider the process. Maybe we should look for more T-shaped personalities. Maybe we should give more weight to lateral thinking. I don’t know. We people are kinda bad at making predictions ( see Daniel Kahneman’s works, especially Thinking fast and slow, which I shamelessly “stole” for my title). My crystal ball is foggy. Maybe anything that I wrote about in here won’t happen. Maybe we will happily ever after live in our chambers.

But what if not ?

PS: These are my views and mine only. No connection with any past, present or future employers :D

I write code