I use AI now for coding at work. I hate AI with a burning passion, but I use it for one thing: to replace the long hours of internet searching to find out how this-or-that API work in this or that language to do this-or-that task.
In other words, let’s say I want to use OpenCV in Python to recognize a square shape in an image. Before AI, I would look up the OpenCV generic documentation, the Python-flavored version of it, figure out what I needed to instantiate / initialize / convert… to do what I needed to do. I knew what I wanted to do, I just didn’t know how practically.
Now I just ask the AI server sitting next to me (yes, I refuse to do cloud-based AI, so my company bought me an expensive standalone AI-ready machine): “In Python, in OpenCV2, write a piece of code that takes a image as a numpy array of uint8’s and detects contours with 4 segments of equal lengh to within 10%”. The thing will piss 10 / 15 lines of code in seconds with the right OpenCV calls that would take me a hour to figure out myself.
In other words, I use AI as a supercharged search engine. Because I figured what a coding model really is is a condensate of technical stuff from the internet, both mainstream and obscure, with an ability to understand what I’m talking about.
It’s the opposite of “vibe coding”, in which somebody with little to zero coding skills asks a powerful AI to do their job (as in “Write an application that does xyz”). That’s bound to generate terrible buggy code, and it’s way beyond the capabilities of the models I can run locally on my little AI-ready machine. What I do is code exactly like I used to, but use the AI model to do the boilerplate and figure out the boring-ass stuff piecemeal, in tiny enough chunks that I can verify the code it wrote from start to finish and make sure it didn’t hallucinate stuff it didn’t actually know - which it does on a regular basis.
And it turns out, I don’t actually use AI that much when I do that, because it doesn’t take much for a prompt to take longer to write than writing the code yourself when you’re prompting simple, short things. The OpenCV example above is worth writing a prompt for because it’s quicker than perusing the extensive OpenCV documentation. I might even ask it to take some really long declaration and rework the format in some way that might take too long in vi. But for the rest? I do it myself really, because it’s plain quicker.
And so for me, the art of AI coding is figuring out when it’s quicker to do it myself and when it’s quicker to let the model do it and me reviewing and adapting what it did. And at least the way I do it, 95% of the time, it’s quicker when I do it myself 🙂
if you’re not constantly using the skills you’re offloading, are you not doing yourself a disservice?
I’d rather tackle these things myself and only if I get stuck would I want to press the big expensive “hint” button, and even then I’d first want to check documentation or forum posts
but what do I know? I’ve been “left behind” after all
Well, I look at it this way: my expertise as a coder, and in a sense my pride, is to produce good maintainable bug-free code.
When I code the traditional way, I constantly look things up I don’t know and I constantly “steal” from other people’s work. I take their knowledge and adapt it. I basically work to find the hints myself.
When I use AI, the AI companies have done the stealing for me and packaged it up into a model. The model contains the knowledge that I’d be looking for on the internet, and it can do the searching for me. It gives me the hints.
Similarly, I will point out that you’re using a compiler to generate machine language code. You’re not using an assembler. Why is that? Are you not doing yourself a disservice by not using your assembly-coding skills and delegating the job to the compiler?
Which way of getting the hints is better is as irrelevant as driving stick or driving automatic. And nobody bats an eyelid that modern cars don’t require you to adjust the ignition timing manually anymore. what’s important is driving well.
But I will say this: I’m fully aware that AI will take over my job eventually. At the moment, it’s too damn dumb to do any good without a real human coder doing heavy checking - something that is willfully ignored by bad engineers who think AI is trustworthy. But it will get better, and one day I won’t be needed and my job will disappear, just like taxi driver and truck driver will become a thing of the past when self-driving vehicles become overwhelmingly better than them.
That’s just how progress goes. At the moment, I enjoy the twilight of programming as a job - something I’ve done for 35 years - and I can still use AI as a super-fast search engine but call the code my own. But all that will be over in a few short years. Hopefully I’ll be retired before then 🙂
amusing that you think you will be left behind genuinely, while I think that my avoidance of LLM use may very well make me more skilled than you in the long run.
long run
Yours, not mine. I’m old enough that I look forward to never touching a fucking computer ever again in my life in a few years. Sadly, I have to dive into this AI aggravation that I didn’t need to finish my career. I make the best of it until I don’t have to. And then whether my job disappears or not is a matter of total indifference to me.
Fair point.
I personally hate LLMs, I think they are a cancer in our society and people don’t understand what they should offload to it. They obviously are going to use it wrong, they are going to lose their hard earned skills and only noticed it way too late and eventually it will be used as a way to create narratives to control the people since they trust it blindly.
That said, if you are using it as a search engine, review and understand what it spews and don’t trust it blindly, it seems to be a pretty sensible way to do some things.
Personally, if I have the choice, I will never use it for the reasons listed above, because they hallucinate too much for being useful to me and because I take pleasure in doing my job the old fashioned way but I can see some things being done faster with them, if combined with this approach.
they are going to lose their hard earned skills and only noticed it way too late
I hate AI too for the same reason. But in fairness, when was the last time you did any complex math in your head or on paper?
Like everybody else, I learn to add, subtract, multiply and divide on paper (do they still teach that in school now? I’m not even sure). And I learned cursive writing with fountain pens. And I did a lot of it when I was younger. But I reached for the slide rule as soon as the physics teacher told us we were allowed to use it, and then I reached for the calculator as soon as I could afford one. And I still do. I rarely calculate anything myself anymore, and I haven’t written more than 5 words on a Post-It note with a pen in decades.
AI is the same thing. It’s just going to make the old computer skills obsolete. Just like modern compilers made assembly obsolete to a large extent. I don’t even have a problem with it: like I said, it’s just progress.
My beef with AI is that it’s deployed by giant tech monopoly for the benefit of a few billionaires, rather than to benefit society as a whole. It’s gonna put everybody out of a job and unravel the very fabric of society so that a few super-rich fuckers can become even richer. And they’re deploying it to turbo-charge the surveillance capitalist system that made them rich in the first place, putting privacy and democracy in grave danger.
I have no problem with AI as a technology, and I have no problem with what it means for my job. My problem is with who controls AI and for what purpose.
I don’t do complex math in my head but I do all the “simple math” (money conversions, add, subtract, etc) in my head (unless they are huge numbers). I believe it’s important to keep sovereignty of our mind and not stop thinking and exercising it altogether.
It’s just going to make the old computer skills obsolete. Just like modern compilers made assembly obsolete to a large extent. I don’t even have a problem with it: like I said, it’s just progress.
I think less and less people will have these skills and maybe that will make this knowledge more valuable in the future. (reverse engineering, performance gains, etc)
It’s gonna put everybody out of a job and unravel the very fabric of society
I don’t know if this is accurate. Companies are starting to wake up and see that using LLMs is not as good and productive as they were led to believe. Some were even forced to rehire the people they laid off. Tokens are getting way too expensive. Things are slowly getting derailed.
they’re deploying it to turbo-charge the surveillance capitalist system that made them rich in the first place, putting privacy and democracy in grave danger.
100% agree with this. Furthermore, if most people become AI dependant and trust everything it says (as I’m seeing happening with young and some older people), you have the perfect tool for mass control, even better than social media…and we can still talk about the impact it has in the environment, which is another matter entirely.
I have no problem with AI as a technology
The technology is not at fault here, it never is. The problem always resides with the people and in this case with the values of each society. In a society were empathy is becoming a myth or a sign of weakness and where people only care about themselves, giving them a technology that works for them at a lower cost, will only reveal their true selfs.
