this post was submitted on 21 Dec 2025
151 points (98.1% liked)
Fuck AI
4969 readers
1748 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
No it wasn't, shitty bosses were and AI was their excuse.
I lead a lot of the AI work at my organization. Sr leadership is chomping at the bit to find an area to reduce headcount. It's just not there. Nobody can understand a solid enough ROI. Most all the companies say it saves time and can shave FTEs but there's no proof. If anything it added an entire section we had to fill expertise which has added headcount. Which data scientists for AI and engineers and architects who understand AI aren't cheap. Helluva lot more expensive than the jobs they're hoping to replace.
Hilariously the people LLMs could replace most cost effectively are probably management and higher. They’d probably do a better job too.
Let's look at the job requirements for executive level and see who has the advantage.
Biological relation to other C-suites....humans win.
Ability to spout an endless stream of nonsensical garbage... AI wins (it was close however).
Wasting valuable resources for little to no gain.... AI wins.
Ability to kiss ass.... Humans win.
Making correct conclusions based upon available information... AI wins (only wrong 60-70% of the time).
Ability to claim other people's work as their own... Too close to tell. Gonna have to give this one a tie.
Ability and desire to harm/kill other people for personal gain. Humans still win.
I did replace a bunch of tasks with “AI” this year, but none of those tasks are worth a full employee. They do save time so our employees can get more done.
Execs want me to build more AI tools, but frankly I’ve built all the ones which can be made reliable.
I pretty much use LLMs to manage jira for me.
Someone I work with uses the jira LLM to make the worst tickets I’ve ever seen though. They’re massive.
I've been asked by my manager to come up with a goal for next year on how to augment my job with GenAI.
Why not combine two hatreds at once - GenAI and Jira. Thanks for the idea!
I write Jira tickets with what needs to be achieved and why, and usually my preferred method of doing it and if there's any constraints. I usually don't much care exactly how it's done, as long as it works, but sometimes it needs to fit into the bigger picture in a way that might not be obvious. My team have different strengths, and I'm more than happy for them to do what they do best. Most of my tickets range from two to six sentences in length - some are longer if it's complicated, but most things aren't.
My managers don't think that's enough for a ticket, and have been using LLMs to boost them up to several pages. That obviously requires making up tonnes of shit and overspecifying shit that doesn't need specifying. We have to waste time verifying that we've not now got requirements that make no sense, and now have pages of test notes of things that don't need testing, which means tickets now take days rather than hours to complete.
No-one can read these multi-page monstrosities, and are using LLMs to compact them down to a few sentences again.
I can't believe that we're boiling the oceans for this shit.
Yeah, this is annoying. My Jira tickets are usually as long as "Implement new email notification system. Make sure opt-out is respected", I set a due date, assign myself, and move it to the active state. Once I've started work I post the PR.
If it doesn't fit on a sticky note then the task or documentation is misspecified.