The rush of productivity and fast output

I would call myself a lucky person in the tech field. The reason, being the fact that I had the opportunity to learn programming the hard way(although some people who did not have access to the internet my contradict me on that).

I remember the days when I was watching youtube tutorials and searching on the internet for answers, when sometimes I was taking an entire day to solve a leetcode problem, and not because it was extremly hard (well ye it was for me at that time) but because finding answers was hard, finding answers was took time, it meant browsing multiple websites, it meant sometimes even reaching out to Google’s second search page, pff, even third one, because the answer was not there, because the answer was not easy to find. It was often a struggle, it often happened that not that many people encountered my error, and moreover a solution to it.

I even remember when I found a solution to my problem on a Chinese website, and had to spam google translate to make sure that I understand what is happening. Not knowing how to solve a problem meant that you had to try again, and because there were no easy answers available around, or someone to tell me how things should be done, I have to dig by myself, I had to read some irrelevant information sometimes, and filter a lot of content throughout the journey of searching for the truth, and everything added up, everything took time, a lot of it.

And I wouldn’t want it to be any other way, because through all those countless random articles and irrelevant information I was discovering new things, I was intersecting with new concepts, with techniques that I never though of, or frameworks, that made me try them, libraries that I wouldn’t have heard of otherwise, I was consuming and processing so much valuable stuff, I was learning, in the truthful meaning of the word.

Ye, but it was a big struggle wasn't it? - Random Reader

You bet it was, but that was the magic behind it, all that struggle, all the pain and countless hours it was all adding up, and the harder it was, the better it felt at the end. The feeling of solving an issue or an error, all by yourself, after so much work put in was undefinably (if that’s even a word) sensational.

The reason behind me writing this article is that I must admit, for some time after getting in the tech industry now, this element slowly faded out, this drive like moment of learning, this feeling of discovery, it faded away behind the darkness of expectations, of deadlines and story points, behind the idea that time to result should be fast, and action should be quick, to maintain adaptability and flexibility to the market and the client. But that squeezed a lot of juice (and for all of the persons that would tell me I burnt out, ye, I did not), I was still finding this glimpse of light, when I was doing some random course, or when I was preparing for some certifications, but as time passed and the new LLM era started upon us, this only got worse.

The best part about LLMs is the fact that they have a lot of knowledge compressed behind them, they know things, and while not always correct, they can definitely be a lead creator when you are searching for answers.

The worst part about LLMs is the fact that they can alaways answer when you are searching for answers, and while not always correct, they know things, because they have a lot of knowledge compressed behind them.

For me they seemed like a blessing while being a curse at the same time. And I will make this as straight as possible, when time is pressing you, when expectations from all around are coming towards you and the output is what’s measured, you will be more inclined to make the “logical” decision of using them, but this comes at a cost. I feel like being able to find an answer fast is a quick win and a slow loss, I observed this at least in my behaviour, and I don’t know if this is because I am getting older, or lack of sleep, but my ability to remember the solutions and the logic behind the solution to my problems decreased as I started using LLMs more and more. I believe it came out as a comodity more since I can always just prompt the model again and find the answer I am looking for, but is that what we are looking for? There is definitely a group of people that would agree and think that abstractization is okay and it is part of the progress, and while I would like to think so, I don’t think progress comes with downgrades.

I felt it myself when I did not have access to any wifi, and while writing code, I realize that I was struggling programming by myself without any support from a model, it felt weird, definitely not in the good way.

I started programming and I got into the world of computer science because I liked this craft, it enabled me to create and provided me limitless capabilities for growing. Thinking that my passion would transform into an ordinary activity, was not gonna stick to me, so that’s why I went against the wind. Now this is not gonna be your moment of:

I decided to not use LLMs at all and I am going all against AI.

I started making my life harder a bit, and as per some inspiration from one good friend I started using vim. (Ye, he wrote this whole article just to brag about the fact that he is using vim) I find it a bit uncomfortable from time to time, and it makes me slower (haha, gotcha here;) and because I am still learning it forces me to actually think through what I am writing once again. I also listened to a suggestion of the same good friend @danielavornic and I customized my terminal, it was not rocket science, but it definetely sparked that teenage drive since when I was started programming, so thank you for that Daniela!

And I would be a fool to admit that I am not using LLMs anymore, I definitely do, but I am being more cautious about it, I allocate myself time without them, where I am going old school, and it feels so much more enjoyable to be honest. I did not manage to find a clear way of how they should be treated, and I do not believe the vast majority of AI gurus out there that are saying they know the key. What I can say though, is that learning should never feel easy, it should create discomfort, it should feel hard sometimes (say that to all the gamification first people out there) and while I realize that on some notes I can’t totally escape this output driven approach, I would be naive to think that it should be the way to go.

And for those of the people that got to the end, especially teachers, employers, mentors, please do me a favour, and if you really care about those students/employees/people and their future. Make sure you give them space and time to do it the slow way sometimes, it really makes a difference!