Work on something hard and interesting
the lesson every young person in the valley needs to internalize is nothing is as sexy as it seems and everything is hard, so you might as well do what you like doing
— amir (@amirbolous) May 22, 2024
oh ai is cool? have u seen how much grunt work scraping, cleaning data, running experiments, training and…
A few months back, I was laid off from my job (I took a break), which prompted me to reflect on my career and what kind of work I find fulfilling. I read this tweet and it prompted me to write a twitter thread. I try to summarise my thoughts in greater detail here.
It's a bit incoherent right now. To give you a hint: I am trying to justify why one should work on something hard and interesting and how to figure out what is interesting. I also try to decode the tweet I mentioned in the beginning of the blog.
Paul Graham in his essay "How to work hard" says the best test of whether it's worthwhile to work on something is whether you find it interesting. I think it's true.
It was easy to work hard when I had just started. I was into backend engineering. First 1 year or so, I was new to working in production environment and pushing code that impacted several customers which were paying my company millions of dollars annually.
It was intimidating (especially turning off those feature flags haha). However after an year or so, my work started becoming a bit repetitive. I like backend engineering but the work evolved to be more of writing a lot of business logic in Java. I was still creating impact but my learning curve had declined plus I was getting bored and tired of navigating deep Java codebases with no proper documentation to write lots of business logic with lots of edge cases.
The amount of novel stuff reduced so the work felt uninteresting. Familiarity and repetitiveness creates a sense of boredom. I was still productive but my mind felt dull at the end of the day.
In my break, I realised that I need to work on something more interesting. I have always been interested in AI (since college) and these days LLMs are booming so I am trying to get into the applied ml + backend engineering space for my next role.
However, working in AI is not all hunky dory. I have trained models in the past and worked on projects. They involved a lot of grunt work and babysitting the models. My mental health would almost always crash after big projects.
However, I pulled through my side projects because I was interested in them. But then, then you don't have the same energy when you work for an employer so you gotta carefully choose the next job or hard problem to work on.
Decoding the above tweet
"oh AI is cool? have you seen how much grunt work scraping, cleaning data, running experiments, training and testing models is?"
You can get this picture from working on side projects, or a while in production. Getting a broad idea of things leads to familiarity. It can make things seem boring as I mentioned earlier, even if it is a field of interest. I have had several panic attacks because of such thoughts (what if I get bored again or what am I really interested in)
Everything involves grunt work so you might as well choose to work in an area of highest interest to work in.
"Nothing is as sexy as it seems and everything is hard, so you might as well do what you like doing"
If the problem you are working towards is not challenging enough, then you will get bored. So you might as well choose a hard and interesting problem to work on. Smart and ambitious people are attracted to these so you won't have to worry a lot about finding good company of people to learn from.
On identifying if you are bored and need a change
There are two things - either you are really bored and you need to change your outlook.
You will need to be very honest with yourself in order to identify if you are really bored. These things usually show up in a matter of months or an year.
You may find things boring because not enough novelty
Maybe you are working on the same project for a long time, not trying out new things
You are stuck on stuff often (we crave progress so if minimal progress, things become boring)
OR
It could be an outlook problem - you are zoomed in too much into the tools - programming languages, frameworks, LLMs, agents
You need to zoom out and think of looking at your work as solving problems. we, engineers and builders seek enjoyment from solving hard problems that ideally have an impact in reality (changing people's lives, making something more convenient, stakeholder value etc.)
In the case you have concluded that you are really bored, you gotta FAFO (fuck around and find out) and try something different. This may mean changing companies where they work on a problem of your interest, trying out different industry in same field (like going from E-commerce to Fintech, Fintech to AI etc.)
Thanks for reading till here.