I’m in IT at an upper level and know painfully well what “AI” really is and that it’s not the disruptor people think it will be. However I feel like I can’t post it anywhere without being judged about it as almost every exec I know has bought into it hook, line and sinker. Even other people I talk to about the issues and limitations look at me like I’m completely weird “you’re in IT and you don’t embrace AI? wtf is wrong with you?”
So what do you all do? I don’t want to make things career limiting but I feel like I’m screaming in the dark seeing where things will really go. It reminds me a lot of the move to cloud and everyone going all in on it without knowing the real ramifications.
Just lay low, change topic and laugh from failures, it’s only approach that would keep you healthy instead of being frustrated. It’s not worth it, you can’t change people.
I can’t even hide my disdain for the marketers including the word “intelligence”.
I feel like I have to at work, 100p. Many of my coworkers use it, and I just can’t bear it myself. I don’t think I’ve ever used AI in a professional context.
Now, in a non-professional context, like a Twitch stream, hell yeah I use AI. They can say some funny shit!
I don’t embrace AI, but I don’t necessarily hate it. The problem with AI is that your average user is not going to understand it. This means that they will trust LLMs to do things that LLMs aren’t supposed to do. Like when people talked about how an Atari can beat ChatGPT at chess… Of course it can, the Atari program was purpose built to win at chess. ChatGPT was purpose built to aggregate text so you wouldn’t have to read multiple sources to get an answer (not always the right answer, but an answer) for your query.
I’m trying to get a job (and failing) after a layoff at the end of last year, and maybe 1/4 of the job postings I see mention you need to be familiar with “AI.” So, yeah, in interviews, when AI is brought up, I try to be pro-AI, but cautious. I even messed around with stuff like LangChain/LangGraph because I see so many job postings that require stuff like that, but the results are underwhelming. Now I’m seeing a lot of job postings that require experience in Azure’s cloud AI stuff, but I’m not paying to learn that. I have “real” ML experience, and it frustrates me that many of the responses I get back from applications just want me to glue shitty LLM tools together.
My problem is search has become so horrible escpically when looking up tech questions. AI is ok at this at least faster then wading through outdated search results looking for the one that uses the new code.
Aye, same boat here. I protested quite vocally, but politely, when our CEO wanted to stick copilot everywhere. Got reprimanded in private by my manager. CEO still stands firm on “you should use AI for everything!”
I tried, they didn’t want to hear it. I don’t really know what else I can, or should do. I just don’t use it, and have blocked it from my sight as well as I can.
I will still be honest and tell people why I am against it if someone asks, I won’t lie, but I’ll just keep quiet unprompted. Holding on to my “told you” for an appropriate moment I guess.
Except your manager now “tells you” that your productivity should be +50% and if it isn’t…
Nah, fortunately my manager is a good guy, CEO is just a pissy finance guy who got butthurt being disagreed with publicly and probably was on my manager’s ass about it. He mainly criticised me, that I should’ve come to him first to discuss my concerns, and he would’ve helped me bring it up to thinskin execs privately.
I disagree with him, but get his point. I mean, the CEO did ask for comments during the meeting, I had comments
¯\_(ツ)_/¯Kinda sounds like CEO wanted to fire you, and manager had to stick their neck out to keep you.
I mean, even if he did want to, he couldn’t. Labour laws in my country are fortunately quite robust. I’m pretty sure my manager didn’t score any extra points with him, though, that’s for sure.
I don’t really know what else I can, or should do.
Well, if they insist so much on using it, you should oblige.
Hey AI, make this email I have to send twice as long.
Hey AI, take notes about this meeting while I do something else.
Hey AI, reply to this company wide Teams post with the most empty, meaningless, long winded reply possible.Damn, you were right Mr. CEO, this AI is helping me write sooooo much better ;)
I’m skeptical about a lot of things about it in general, but I found it really helpful in one particular scenario: I’m learning a new framework and I type out something I think is right, but is not, I ask it why it’s not working and how to fix it. It pulls all the relevant files and errors from the IDE and tells me something. Sometimes I have to remind it to use a different version of the framework as it has a different approach slightly, but overall it helped me find and figure out quite a lot of issues. But my codebase is pretty small still, I don’t use it to engineer big changes.
i kinda feel like you could do better with the basics of good documentation, good tutorials, and good training. is AI really helping in the long run?
I mean, has it’s uses for sure, but the scope in which people are deploying these cookiecutter generalist LLMs and expecting bloody miracles is unrealistic. Plus, I like my privacy, AI’s are kind of an antithesis of that. Environmental cost is way too high as well.
It’s a tool, and it can be used successfully, but it’s so inefficient that I can’t justify it personally.
Ill do you one extra even. One of our customers, a lawfirm, has asked we remove AI from their environment over privacy concerns. We, as their IT, are still trying to push and sell licenses to our other customers and even use it ourselves.
All against the wishes of the actual workers themselves in my company. The execs and sales team is pushing it so hard and ignoring that a literal law firm is super against it.
Surely this will turn out well. I let that customer know that we’re still pushing AI, and management is very pro AI still. It might be a good idea to look for a different IT provider if that goes against your values.
Surely this will turn out well.
I’ve had the privilege to watch people who intentionally pissed off lawyers to try to make a quick extra buck.
Interestingly, the lawyers promptly pursued legal action, and everyone agreed to settle out of court for some reason…
It almost felt like fucking with someone on their home court is a stupid move.
It’s just not the disruptor it’s advertised to be (smart AI that is better than ten employees). The fact that it’s being shoved into everyone’s faces and threatening or actually replacing people’s jobs (inadequately or not) means it is a very large disruptor. Fixing the mess left if we ever get a chance to get back in control is going to be fun.
And it’s not just LLMs. The whole everything-in-a-cloud shit, the constant monitoring of clicks to determine productivity, the forced obsolescence of working equipment for new crap that doesn’t. We hit a peak in tech quality some time ago, the LLM factor is just another thing on the pile of junk.
Ironically enough, we got a company wide email a while back saying how they’d partnered with an Indian tech recruitment firm “because we can’t find enough talent at home”. So while all this talk of AI taking our jobs is clogging up the headlines, it’ll be simple outsourcing to lower wage countries that’ll do most of us in, still.
AI: Alotta Indians.
Time to dust off those C/C++ Legacy Codebase skills…
I retired early. Didn’t have the energy to deal with the shit show anymore. The last year of my career watching so many people sucking down the Copilot kool aid while producing worse output was exasperating.
For those of you still fighting the good fight with critical thinking skills, good luck!
I work in IT and we have to go through a whole process to get a licence to use it, so I’ve just not done that. Other people on my team have, it’s a ticking time bomb and I’m just sitting here watching them all ignore the obvious signs.
Just last week someone was complaining about the AI randomly adding files to our code, and when I mentioned yeah these AI keep deleting customer data and stuff, they just brushed it off. We’re in an industry that’s super regulated so this is gonna end with legal action I’m sure. I’ve also watched people use AI to automate basic tasks, and when I’ve reviewed what they’ve done it’s all completely wrong, to that point that if the wrong person sees it, we get a full audit and the entire team is placed under constant supervision with a lot of our access heavily restricted for months.
Meanwhile I’m just sitting here doing more and doing it better without AI. It’s not worth pushing back because I already cop enough flak for stuff other people do or don’t do. I hope the company gets destroyed in a massive lawsuit, it won’t, but god it would be awesome.
If you’re in a tightly regulated industry, take DETAILED notes and store them off-site for when the lawsuits inevitably hit and they go hunting for someone to pin the blame on. Make sure that it’s clear that a) you didn’t use the stuff that breached regulation, and b) you know who did.
Been working IT for over a decade now. No, this has never been an issue for me. I voice my hate for AI anywhere and everywhere I can. My boss knows, his boss knows, our CEO probably has an idea of where I stand.
I’m in the job hunt and man…the circle jerk around AI is just… inescapable.
I don’t get it. By the time you write “a good prompt” I will have written something usable and passable. The level of effort to debug whatever the prompt gave me vs my code is, let’s say, roughly the same.
So…what time am I saving?
The only thing this helps is people who can’t code. And we just assume that it’s passable code only because it works.
Hell I’m at a point of commit very wrong code to GitHub under an MIT license just to start poisoning Copilot.
You actually don’t have to worry about poisoning their models, they’re already well on their way without any of our input. Broken vibe coder projects are probably already filling github at an alarming rate.
The AI trained on vibe coding will train future AI to be even more enshittified.
Hell I’m at a point of commit very wrong code to GitHub under an MIT license just to start poisoning Copilot.
I’m way ahead of you. To…uh…poison the AI…not just because a bunch of my code is shit…
I’m a sr. devops engineer and I abhor this shit.
I don’t really know what to do. Considering a career change. I’m trying to use my company ERP to talk to a counselor about my feelings around all of this.
Our current state of things is one of blind compliance where people are using AI to generate slop that will be evaluated by a different AI. No one is really doing much anymore at all. It feels like things are over. As people “embrace” AI (as they’re told to do), they’re literally training their own eventual replacements. It doesn’t matter that things aren’t there yet, they’re starting to get replaced by the shit that’s in use now.
I’m trying to use my company ERP to talk to a counselor about my feelings around all of this.
I’m pretty confident ERP isn’t “erotic role play” here but I’m not sure what it is.
You never know, companies are always looking for new ways to fuck their employees.
Usually ERP means Enterprise Resource Planning but I think they meant EAP, Employee Assistance Program. EAPs are usually free short term counseling provided by your employer.
Fuck… yea, I meant EAP.
It doesn’t matter that things aren’t there yet, they’re starting to get replaced by the shit that’s in use now.
This is what I keep trying to convince people of, too, even others who hate AI. I come from customer facing support in cloud infrastructure in my previous role.
They upped the cost of high level enterprise support while making it so that non-paying support contracts couldn’t even talk to a human anymore. Made us feed every ticket into the AI. Of course none of this new profit or savings was passed on to our salaries and it was a move designed universally to piss customers off. We had a customer satisfaction feedback of over 95% positive (almost unheard of) and it was the only metric they really used for our performance given we were all remote.
Hell, they replaced the INTERNAL IT systems with it. Couldn’t open a generic support ticket for yourself or your workstation without first doing a runaround chatting with the AI.
It doesn’t matter in the slightest that it’s not good enough, that we know it can’t complete all the necessary tasks. It has the appearance of good enough so they’ll flog it for short term profits and call it a day. Not gonna dox myself, but this is a big company, you know it, have fun guessing. I wouldn’t be surprised at all to know others are following this pattern.
I am fairly certain I know the company.
Telk me it’s Broadcom without telling me it’s Broadcom.
“you’re in IT and you don’t embrace AI? wtf is wrong with you?”
This is always hilarious to me. “You’re an expert, unlike me, and you’re not whole-hog into this thing that’s in your field of expertise and not mine, while I am? What’s wrong with you?”
The arrogance of the capitalist class is beyond amazing.
