I've been a developer for 8 years and in the first 7 years I've always heard from people that I was fast at building good quality tools. Now, any developer can match my output easily. I feel no pride anymore in anything I build using AI.
The act of really thinking about an elegant solution makes me feel guilty because someone else has written a whole book of code by the time I'm done thinking. It sucks. I hate it and I don't know what to do. I have no energy left to work on side-projects anymore because my soul is sucked out of me every single day and I feel like anyone could just come in and sweep my idea in a few days of vibe coding.
No they can't. They can get a clanker to crap out code at a pace that matches or beats yours, but I assume your code is better, and know that you have a better understanding of your code than the clanker user has of the clanker code.
AI solutions are still extremely mediocre. Sure, if you tell it exactly what to build, it will build it way faster than any human could. LLMs still lack creativity to come up with the best solution. They’re trained on the average, so they’ll output the average. I constantly have to interrogate LLMs to get them to give me the best code. Taste is more important than ever.
Whats interesting is that this "We don't care about the code, we care about the outcomes" was common management speak for years before AI was on anyone's radar. In some ways it made sense, it was what distinguished your junior programmer from the senior engineer. One worried about the function, the other about how the systems would interact. But it also swept a whole bunch of technical debt and ticking time bombs under the rug. It allowed the ship it quickly and let it be someone elses problem in 2 years when no one can read this garbage, its slow as mud and will require a full rewrite to add one new feature.
AI and its use encouraged use, feels like gasoline poured on that existing mindset.
But with AI debugging and reviewing skills, this could be solved too I believe. It's not going to be write the code and review once anymore. We will run AI multiple passes reviewing it with various perspectives and we would have code as good as any senior engineer can design and build.
Nobody misses boilerplate. What's gone is that understanding used to be mandatory, you couldn't ship what you didn't understand. Now you can, so it became optional, and optional things atrophy.
The tiem that took experienced programmers to hand over their entire development process to AI was so remarkably fast, that it got me to believe that there are almost no programmers that enjoyed solving problems.
Once, my teacher in school told me that a software engineer needs to have some sadasim in it, because it will need to enjoy the pain of being stuck on a problem. I no longer believe this is true.
Most people are in it for the money, the status, the comfortable conditions, or maybe because this is the blue collar of the 21st century.
The people who are in it for the love of the game, are very little.
You also have to consider that the context and the problems you get are very important.
I do not like prompting coding agent at home, and still code as a hobby, but use coding agent extensively at work for various reasons: a feeling like better than good enough is not rewarded mainly.
In the grand scheme of things, programmer might just be an interim profession between when computer was invented and when AI starts programming everything. Future generation problem solvers would just have AI write the programs to solve problems, and probably only a small handful of people in the entire world would understand partially how it works under the hood.
This is a dumb take and a very weird attempt at gatekeeping. Of course programming is a profession. It can also be a hobby, but that doesn't mean that it isn't a profession. The world runs on computers and software development is lucrative.
I've been a developer for 8 years and in the first 7 years I've always heard from people that I was fast at building good quality tools. Now, any developer can match my output easily. I feel no pride anymore in anything I build using AI. The act of really thinking about an elegant solution makes me feel guilty because someone else has written a whole book of code by the time I'm done thinking. It sucks. I hate it and I don't know what to do. I have no energy left to work on side-projects anymore because my soul is sucked out of me every single day and I feel like anyone could just come in and sweep my idea in a few days of vibe coding.
> Now, any developer can match my output easily.
No they can't. They can get a clanker to crap out code at a pace that matches or beats yours, but I assume your code is better, and know that you have a better understanding of your code than the clanker user has of the clanker code.
AI solutions are still extremely mediocre. Sure, if you tell it exactly what to build, it will build it way faster than any human could. LLMs still lack creativity to come up with the best solution. They’re trained on the average, so they’ll output the average. I constantly have to interrogate LLMs to get them to give me the best code. Taste is more important than ever.
Whats interesting is that this "We don't care about the code, we care about the outcomes" was common management speak for years before AI was on anyone's radar. In some ways it made sense, it was what distinguished your junior programmer from the senior engineer. One worried about the function, the other about how the systems would interact. But it also swept a whole bunch of technical debt and ticking time bombs under the rug. It allowed the ship it quickly and let it be someone elses problem in 2 years when no one can read this garbage, its slow as mud and will require a full rewrite to add one new feature.
AI and its use encouraged use, feels like gasoline poured on that existing mindset.
But with AI debugging and reviewing skills, this could be solved too I believe. It's not going to be write the code and review once anymore. We will run AI multiple passes reviewing it with various perspectives and we would have code as good as any senior engineer can design and build.
Nobody misses boilerplate. What's gone is that understanding used to be mandatory, you couldn't ship what you didn't understand. Now you can, so it became optional, and optional things atrophy.
The tiem that took experienced programmers to hand over their entire development process to AI was so remarkably fast, that it got me to believe that there are almost no programmers that enjoyed solving problems.
Once, my teacher in school told me that a software engineer needs to have some sadasim in it, because it will need to enjoy the pain of being stuck on a problem. I no longer believe this is true.
Most people are in it for the money, the status, the comfortable conditions, or maybe because this is the blue collar of the 21st century.
The people who are in it for the love of the game, are very little.
You also have to consider that the context and the problems you get are very important.
I do not like prompting coding agent at home, and still code as a hobby, but use coding agent extensively at work for various reasons: a feeling like better than good enough is not rewarded mainly.
In the grand scheme of things, programmer might just be an interim profession between when computer was invented and when AI starts programming everything. Future generation problem solvers would just have AI write the programs to solve problems, and probably only a small handful of people in the entire world would understand partially how it works under the hood.
[flagged]
This is a dumb take and a very weird attempt at gatekeeping. Of course programming is a profession. It can also be a hobby, but that doesn't mean that it isn't a profession. The world runs on computers and software development is lucrative.