My company is strongly pushing AI. There are lot of experiments, demos, and effort from decently smart people about integrating it into our workflows. There are some impressive victories that have been made with AI tooling producing some things fast. I am not in denial about this. And the SE department is tracking improved productivity (as measured by # of tickets being done, I guess?)
The problem is I hate AI. I hate every fucking thing about it. Its primary purpose, regardless of what utility is gained, is spam. I think it’s obvious how google search results are spam, how spam songs and videos are being produced, etc. But even bad results from AI that have to be discarded, IMO, are spam.
And that isn’t even getting into all massive amounts of theft to train the data, or the immense amounts of electricity it takes to do training and inference, as well as run, all this crap. Nor the psychosis being inflicted onto people who emplace their trust into these systems. Nor the fact that these tools are being used to empower authoritarian regimes to track vulnerable populations, both here (in the USA) and abroad. And all this AI shit serves to enrich the worst tech moguls and to displace people like artists and people like myself, a programmer.
I’m literally being told at my job that I should view myself basically as an AI babysitter, and that AI has been unambiguously proven in the industry, so the time for wondering about it, experimenting with it, or opposing it is over. The only fault and flaw is my (i.e. any given SE’s) unwillingness to adapt and onboard.
Looking for advice from people who have had to navigate similar crap. Because I feel like I’m at a point where I must adapt or eventually get fired.
The problem is I hate AI. I hate every fucking thing about it. Its primary purpose, regardless of what utility is gained, is spam.
You are describing one type of AI, that being Generative AI. Even more specifically, Generative AI from publicly trained models, examples being ChatGPT, Claude, and Grok. If you hate those, don’t use those. This isn’t the only AI that exists.
We’re getting into data science here, but you can build and train Machine Learning models exclusively on your own data. So no theft/spam contamination here. If your needs are in the Generative AI space, you could even build and deploy your own Fine Tuned model from your own data on top of one of the public models, so it would have knowledge of your business or industry.
All AI incarnations are just tools. You don’t start with a tool. You start with a problem to solve, and you use a tool to assist or make it better. So the beginning of this journey is asking the question: “What problem are you trying to solve?”
Ask ChatGPT “How do I unionize my workplace to protect my job against AI obsessed management?”
+1 also look for “reverse centaur”, its a metaphor by cory doctorow which you may find interesting
AI is a tool, just like a hammer. You could use a rock, but that doesn’t give you the leverage that a hammer does.
AI is also a machine, it can get you to your destination faster, like a car or train.
Evil people have used hammers, cars, and trains to do evil and horrible things. These things can also be used for useless stupid things, like advertising.
But they can also be used for good, like an ambulance or to transport food. They also make us more efficient and can be used to save resources and effort. It depends on who uses it and how they use it.
You can’t control how other people may misuse these things, but you can control how much you know, how you use it, and what you use it for.
AI is a tool like a Gun, more like, specifically designed for a terrible purpose, used for a terrible purpose, and with no way to use it for good except by deconstructing it and finding a new use for it
One aspect that analogy doesn’t work for is hammers and cars weren’t built with the mass theft of intellectual property, they aren’t being leveraged to put people out of jobs, and they aren’t the driving force for building insane numbers of data centres that increase power bills for locals and ravage their water supply.
It’s not necessarily the pure usage of AI that I don’t like, as much as what has been and is being used to create it.
Cars have their own problems of course, and cause more issues with the direct use of them than what went into building them.
I read someone leave a different comment where they said something like “If human meat was the healthiest, least environmentally damaging, and cheapest food, they still wouldn’t eat it.” In this case AI doesn’t really match those benefits anyway
IP itself needs to be abolished, so that part isn’t as important. Further, cars did put people out of jobs that used to draw horse carriages and maintain them. The original commenter is correct with their analysis.
If IP is abolished, that would to me imply that use of AI should be free for everyone, as it’s based on everyones collective knowledge.
Sure, I don’t see a problem with that.
Regarding the idea of IP theft, I think it’s more complicated.
Let’s say you buy a book, learn from it, and use that knowledge to your benefit. We don’t think of this as stealing, even though you didn’t buy the knowledge, you only bought the text.
If you borrow a book, likewise the knowledge is yours, not just while you’ve borrowed the book. Even if you pirate or steal the book (please don’t), the knowledge is still yours regardless.
So I don’t think there should be an issue of an AI learning freely from content.
However, I would agree that when it reproduces a work without attribution that it becomes a problem. The problem that we as a society have even when AI is not involved is defining where that line is. Because there are cases where some kinds of reproduction are ok (parody, homage, etc.). We as a society do not have clear rules for these things because it is hard to define those rules. That’s our problem, not a problem with AI. Using AI just makes it too easy for someone to cross that line and therefore I find it risky to use it for the production of that kind of material.
But I do not think it’s an issue for me to ask the AI how to do some unusual thing in the terminal or to refactor a part of my code to work a little differently. There is no harm in asking it to create a GUI version of the cli program that I made.
As for putting people out of jobs, we may be at an inflection point in our productivity curve. Historically, these have caused short term job loss but ultimately lead to improvements generally once we’ve had time to adjust once people learn how to leverage these tools effectively. Likely, it will create more jobs in new areas.
Humans will always use more resources, especially energy. Until the last few decades the source of thst energy wasn’t concerning. Now it is and we need to find more ways to produce more energy cleanly. Arguing for less energy use is never going to work. We will always use more energy.
Likely, it will create more jobs in new areas.
Yeah, actually fixing the mess the AIs created. I don’t look forward to it.
Tech bros are mostly transhumanists. When interviewed by the New York Times Peter Thief couldn’t bring himself that end of the human race would be a bad thing.
Only thing these dudes care about is total power. Living forever and going down in history as god’s to the machines that replace us.The rest of the world can burn.
In thiels on words from the NYT Trans is only bad when you half ass it like transgender. That is only part trans so it is bad. When you do full trans like trans human, that is actually good.
These dudes have no humanity. That’s why they can be so cruel. And they are the ones that control these tools.
Combustion engine cars cause cancer, AI doesn’t. Checkmate, Atheist.
I am also encouraged to use AI at work and also hate it. I agree with your points. I just had to learn to live with it. I’ve realized that I’m not going to make it go away. All I can do is recognize it’s limited strengths and significant weaknesses and only use it for limited tasks where it shines. I still avoid using it as much as possible. I also think “improved productivity” is a myth but fortunately that’s not a metric I have to worry about.
My rules for myself, in case they help:
- Use it as a tool only for appropriate tasks.
- Learn it’s strengths and use it for those things and nothing else. You have to keep thinking and exploring and researching for yourself. Don’t let it “think” for you. It’s easy to let it make you a lazy thinker.
- Quality check everything it gives you. It will often get things flat wrong and you will have to spend time correcting it.
- Take lots of deep breaths.
I agree with all your points. The problem is that quality cheching AI outputs is something that only a few will do. The other day my son did a search with chat GPT. He was doing an analysis of his competitors within 20km radius from home. He took all the results for grated and true. Then i looked at the list and found many business names looked strange. When i asked for the links to the website, i found that some were in different countries. My son said “u cant trust this”. When i pointed it out to chatgpt, the dam thing replied “oh im sorry, i got it wrong”. Then you realise that these AI things are not accountable. So quality checking is fundamental. The accountability will always sit with the user. I’d like to see the day when managers take accountability of ai crap. That wont happen, do jobs for now are secure.
Use it as a tool only for appropriate tasks.
Which tasks do you use it for?
For my purposes I find it good for summarizing existing documents and categorizing information. It is also good at reformatting stuff I write for different comprehension levels. I never let it compose anything itself. If I use it to summarize web data, and I rarely do, I make it provide the URLs of all sources so I can double-check validity of the data.
Sounds good. It can also write corporate emails well. I’m just writing insults and harsh truths like I would want to throw against my conversation partners, and the LLM tones it down to some bland corpo speak.
I also think “improved productivity” is a myth
Stop thinking, start knowing: https://fortune.com/2025/07/20/ai-hampers-productivity-software-developers-productivity-study/
Thanks! I skimmed this and have it in my reading list for later. I wonder how this pans out across disciplines other than software development. I would imagine there’s a huge diversity of skills out there that would affect how well people can craft prompts and interpret responses.
Make a list of the tasks you hate doing or are repetitive, time consuming and not normally automateable. Then see if any of them are a good fit for an AI workflow.
ai is another abstraction of search. When integrated into software it can perform many things that individual plugins could before but it is one and done. If you had an intern that would prep work for you or look over what you are doing to identify errors you would likely consider yourself lucky but you know you have to evaluate everything they do. Its kinda the same. If you take away the energy usage I would be relatively fine with ai usage. You ask it something. It provides and answer along with references to where it got the information. You can read over the references and if everything looks good you saved yourself some time over a standard search. Personally my writing is way casual and full of bad grammar and misspellings unless I make drafts and review multiple times. ai being able to fix up what I write such that all the spelling and grammar is correct and maybe a bit more polite. thats great.
Treat it like an especially junior dev that just graduated University.
It knows simple boilerplate stuff pretty well, but never trust it implicitly.
simple boilerplate stuff pretty well
-
It does not. My coworkers did some demos on this and it generated random, unnecessary, bloated, shitty, boilerplate. And worse, “because AI told me to” is now used to cement bad practices at my company. Just because it generates 1000s of LoC doesn’t mean you actually need that.
-
If you really need “standard boilerplate”, we’ve had tools to generate deterministic code for a long time now. They’re called snippets or templates. Just setup a company git repo template for your ideal project or whatever and have people clone that. Plus, this template repo would be reproducible, fixable, and debuggable, instead of rolling the dice with AI.
-
People keep saying this. I don’t want to work with an especially junior dev. I’ve been doing that my while career and the only thing good about it is that they get better.
I do know that I am the minority on Lemmy. I believe, there is a space for AI in our life. Of course it is not the way big corporations is trying to shove down in our throats.
The ethical and moral problems of AI is not part of your question.If you decide to work for your company that forces you to use AI, then you either use AI or get a new job.
That’s how capitalism works.You don’t have to like it. You just have to accept the “must use AI” as part of the things they are paying you for.
There is no metric aside of: they are Paying you to do this way.
If you don’t want to do that way - the ball is in your court.Edit: Browsing Lemmy I just saw this post. Maybe it will help you:
https://lemmy.ml/post/40233766The linked article is interesting food for thought.
Same as everything else in life - like the bits that are useful to you and ignore the rest.
As for doing what you’re told at work, who said we had to like it provided it’s a reasonable request?
I’m at a point where I must adapt
What’s wrong with adapting? The one constant in life is that things change. This is a change and you’re not the only person who has faced their job changing - at least you still have it. Adapt or go raise goats.
I’m literally being told at my job that I should view myself basically as an AI babysitter
Feel you 100%.
I dunno why but my entire career everyone always talks like doing IT is simply a stepping stone to becoming a manager, so stupid. Like god forbid you’re not the lEaDeRsHiP type.
And now with the rise of “Agentic IDEs” it’s even fucking worse, I don’t want to be managing people let alone herding a pack ofblind catsautonomous agents.Unfortunately the only solution is to stop caring, Yes, really.
I know it hurts producing sub-par garbage when you know you’re capable of much more, but unfortunately there’s no other way.
If upper management doesn’t care about delivering quality products to their consumers anymore, you shouldn’t either. You’ll stress and burn yourself out meanwhile those responsible won’t lose a blink of sleep over it.
Do exactly what they want. Slop it all. Fuck it. Save your energy for what really matters.That or start looking for another job, but you might struggle to find one that isn’t doing the same shit.
My company does annual reviews. You have to write your own review, then they will read it over and then sit down to talk to you about it.
Last year, I just had ChatGPT write it for me based on all of my past conversations with it. Turned it in. The first question they asked me was, ‘Did you use AI to write this?’ Without hesitation, I said absolutely. They loved it so much, they had me show everyone else how to do it and made them redo theirs. I couldn’t frikin believe it. Everyone is still pissed they have to use ChatGPT this year, but the bosses love that corporate hogwash so much.
They’re about to receive a stack of AI-generated drivel so bad that I bet they have everyone go back to handwriting them.
That’s especially saddening because writing the review is specifically meant for you to contemplate what went well and perhaps what can go better next time. You would think managers would want you to reflect on that. For the benefit of the company, at minimum.
But with stories like yours it is becoming more clear that the only objective is to “use ai” or deliver ai-generated results. Why even bother caring or trying when management does not?
It definitely sounds like you’ve already decided to find AI useless regardless of what it can do, but on the chance that I can maybe change your mind a little bit…
I’m a huge AI skeptic myself, at least compared to my coworkers. Almost everything it spits out has been incorrect for me, except in very narrow use cases.
First, I find it useful to find links to actual documentation for tools/libraries/languages I am completely unfamiliar with. The examples and text it generates are usually poor, but it can do a decent job of finding webpages.
Second, I’ve found it good at guided code reviews. It is no substitute for a real human review, but adding an AI pass before you open a pull request can knock off some of the low hanging fruit.
It definitely sounds like you’ve already decided to find AI useless regardless of what it can do
On the contrary. I’ve seen impressive time-savers be accomplished with it, just as I’ve seen it fail at times. But it’s not about finding it useless. It’s about how I hate it, but also changing my mindset to effectively adopt it.
Second, I’ve found it good at guided code reviews. It is no substitute for a real human review, but adding an AI pass before you open a pull request can knock off some of the low hanging fruit.
We do this automatically, now, on pull requests
I remind my boss that giving AI full access to our codebase and access to environmemts, including prod, is the exact plot of the Silicon Valley episode where Gilfoyle gave Son of Anton access. His AI deleted the codebose after being asked to clean the bugs…deleting the entire codebase was the most efficient way of doing that.
If you don’t mind me asking, what do you do and what kind of AI? Maybe it’s the autism but I find LLMs are bit limited and useless but other use cases aren’t quite as bad Training image recognition into AI is a legitimately great use of it and extremely helpful. Already being used for such cases. Just installed a vision system on a few of my manufacturing lines. A bottling operation detects cap presence, as well as cross threads or un-torqued caps based on how the neck vs cap bottom angle and distance looks as it passes the camera. Checking 10,000 bottles a day as they scroll past would be a mind numbing task for a human. Other line is making fresnel lenses. Operators make the lenses, and are personally checking each lens for defects and power. Using a known background and training the AI to what distortion good lenses should create when presented is showing good progress at screening just as well as my operators. In this case it’s doing what the human eye can’t; determine magnification and defraction visually.
The AI in this case is, for all intents and purposes, using Copilot to write all the code. It is basically beginning to be promoted as being the first resort, rather than a supplement.
Many people think they’re 20% more productive with AI, but they’re actually 20% less productive.
https://fortune.com/2025/07/20/ai-hampers-productivity-software-developers-productivity-study/
