From low-level to high-level
Hey guys! Welcome back to the logs! Today's content is gonna be a little different, as you can see I'm trying a new way to express my thoughts to you guys and the world. So, today's topic is about the future evolution of software engineering through my vision, let's dive into it.
Back then
At the beginning of programming, people were trying to communicate with machines through languages they could understand, which were basically different combinations of 0s and 1s. That's Machine Languages. It doesn't need a compiler to translate or any additional processes compared to modern programming languages. That's the reason why it's fast.
But here it comes the cons. One of the big problem is, machine languages is not readable for human, so it's hard to maintain and debug. You wouldn't expect your manager to type out a bunch of 0s and 1s on Monday morning and you can't even understand it. So there it comes, Assembly Languages.
It can map 0s and 1s into more human-readable syntaxes through an assembler, like MOV, ADD, etc. And the humans think, "Our hardware is going to be more powerful than ever every 12 months. And these languages are still not readable enough — we're spending way more time on debugging than the actual processing time!". That's the birth of Modern Languages.
In this new era, the language syntaxes are more readable than before, making it easier to maintain and debug. But I'm not going to explain any further — the details aren't the main subject in this article, so I'll cut it right here.
The pattern
You can see from the past to these days, the level of our thoughts, tools, and connections is higher than before. From the machine language to modern high-level languages further to the current trend — using natural language to communicate with machines. It looks like a pattern.
In my opinion, this pattern will continue and lead us to a world that focuses on more abstract and vague thoughts. In software engineering, I think we will be more focused on the system architecture than the low-level algorithms. But I'm not saying you don't have to understand the fundamentals of programming or the researchers are going to be useless, they are still needed and you'll still need to understand the pros and cons, known as tradeoffs between different decisions.
Shifting
In the past year, the LLMs have been evolving at an insane speed that we have never seen. I'm not guaranteeing but probably 70-80% of code now can be generated by the SOTA model, even without a single mistake. It's really unbelievable, right? So, now we shift our focus from the writing to considering the whole circumstance factors, tradeoffs, or like I said, more abstract things, even the company culture.
Based on the things I mentioned above, it sounds like the new grads are going to doom scroll and be unemployed til life ends. I mean it sounds like it but not really (I admit I'm exaggerating LOL), the roles are also shifting when the dev game changes.
Because the focus is shifting, you can't still think the way a junior used to, you need to start thinking like a senior, not a junior. Thinking about how to make the system more robust, including stability, scalability, reliability, etc. Not just completing the features your supervisor asks for. That's what I wanted to get across.
I think that's just probably a fraction of it. AI opens up more chances and opportunities in life, but it's all up to you. If you want to, remember time waits for no one.