this post was submitted on 23 Apr 2026
204 points (99.0% liked)

Fuck AI

6880 readers
504 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
top 35 comments
sorted by: hot top controversial new old
[–] one_old_coder@piefed.social 78 points 1 week ago (1 children)

you can turn off Copilot completely by following the instructions on this page.

Here is a direct link for the instructions, it's easy: https://www.libreoffice.org/download/

[–] Ludicrous0251@piefed.zip 13 points 1 week ago (3 children)

Disabling copilot seems to be a weekly action I have to take on my work computer. I wish Libre Office was a viable option...

[–] swagmoney@lemmy.ca 8 points 1 week ago (1 children)

is it not viable? it's my only word processor for school

[–] Murse@slrpnk.net 7 points 1 week ago (1 children)

Previous poster is probably talking about a company-owned computer. I'd be fired in a hurry if I installed anything on one of the computers at the hospital I work at... MS office is the only option for me at work.

[–] Steve@startrek.website 4 points 1 week ago (1 children)
[–] randamumaki@lemmy.blahaj.zone 2 points 1 week ago

There are portable versions available which you could dump on a USB stick if they haven't disabled USB use. https://www.libreoffice.org/download-other/

LibreOffice is only a "not viable option" if you have to work from a corporate computer and they won't allow you to install it.

[–] Naich@piefed.world 26 points 1 week ago

Reason #43,509 to avoid Microsoft stuff.

[–] Zephorah@discuss.online 19 points 1 week ago (1 children)

The latest spell checks are already trash. The older versions did better. Now, it wrongly corrects grammar, happily inserts the wrong your or their, and assumes wrong words based on context instead of just correcting the spelling.

[–] mrgoosmoos@lemmy.ca 8 points 1 week ago (1 children)

my phone is so bad at this now. I thought for the longest time it was just my fat fingers on a touchscreen, but after a while I started seeing it "correct" things live. just absolutely dumb shit, like you said - completely the wrong grammar or word for the context

[–] SolarMonkey@slrpnk.net 3 points 1 week ago

I got to witness first-hand exactly how much worse it’s gotten in the last few years all at once.

See, I’m stuck on an iPhone for the moment (thankfully a 14 which isn't compatible with the extra AI bullshit, or I’d probably have just dealt with the vulnerability..), and because of liquid glass and a bunch of other reasons, I declined to update my OS from 17.whatever. Then the critical vulnerability was found, and the only fix was to update. So I snapped forward to 26 or whatever it is up to now. Literal years of backslide in one day. It’s so fucking unusable. I can’t wait until that motorola that’s compatible with graphene is available.

I went from having a sort of OK autocorrect that was only wrong like 30% of the time, to just flat out disabling it because it was so so much worse than my own typos and fixing my own mistakes. I gave it a few days, but yeah, it would correct words that I’d typed out correctly to random bullshit, sometimes several words later so you don't notice it. It wouldn't fix obvious typos even when they came up in red underline. It got the wrong conjugation or contraction frequently. It was aggressive about changing stuff no matter how many times I typed it in the correct way.

It’s fucking bad, man. You are not imagining it. At all. It’s just that it crept up on you as it gradually got worse.

[–] fira@lemmy.today 16 points 1 week ago (2 children)

I downloaded libre office & never looked back

[–] phx@lemmy.world 2 points 1 week ago

I'm stuck with it at work and already the bullshit like AI summary etc is driving me nuts

[–] ikidd@lemmy.dbzer0.com 0 points 1 week ago (1 children)

Just wish I could print in Landscape from it. Apparently that's just plain not possible.

[–] DonPiano@feddit.org 2 points 1 week ago

I did that last week, I'm pretty sure I put the resulting sign on my door

[–] daggermoon@piefed.world 12 points 1 week ago

I removed Copilot and Edge from my Windows machine. Don't worry, my main system is Linux.

[–] jj4211@lemmy.world 10 points 1 week ago

Work requires that github copilot be on, and most of the time I ignore it.

Recently they made it so that if I so much as accidentally hit tab on one of their suggestions and undid it, it wanted to say 'Co-authored by copilot'. I undid it, but it was pretty annoying.

Now I'm all for transparency, but the mechanism is heavy handed, actually doesn't show this addition in the commit message you see before commit, is frankly a bit presumptious to assert it 'Co-authored', and no obvious way to tweak the bolted on attribution. It's more an advertisement than a 'transparency' measure in truth.

I'd probably go with some option to indicate nuance but without advertising the platform. 'May contain AI generated completions' or 'May contain prompt driven AI content'

[–] Bluegrass_Addict@lemmy.ca 8 points 1 week ago (2 children)

I refuse to review anything by AI if forced to used it... if it's forced, and it makes changes.... and it fucks up, that's their problem. not mine.

[–] UnspecificGravity@piefed.social 9 points 1 week ago (1 children)

Yep. My workplace just rolled out two janky AI compliance agents. They are wrong about half the time, and they give different results for the same content every time you run it through. You could spend a whole day just changing things back and forth per the agent, and I have.

[–] jj4211@lemmy.world 1 points 1 week ago (1 children)

Fun fact, in a demo at my company someone was trying to show off the code review by AI feature.

So they clicked the button and got a suggestion, and showed how easy it was to accept and start over.

In the next round, it suggested to undo the modification it just made, and the demonstrator said this was good, that upon further consideration it determined it's suggestion was wrong and accepted it and asked for a review.

Then it suggested the first change again...

[–] UnspecificGravity@piefed.social 1 points 1 week ago (1 children)

The second someone suggests that the agent learned from an interaction is when you know they are full of shit because that's not even how LLMs work.

[–] jj4211@lemmy.world 1 points 6 days ago

Yeah, had someone in my work say that. They gave it a 'college homework assignment' type problem to see if it worked, and it mostly worked but made a mistake. The next day, as an entirely separate chat session he repeated the experiment and it happened not to make the mistake and he assumed it learned from his previous day conversation. That he was the first ever person to post a very obvious intro to programming problem to the engine and taught it this.

But in this particular scenario, you didn't have to attribute learning during the interaction, it's just that the original human way looked close but wrong to a more usual pattern, and so the model wants to make it the usual pattern. Then with the usual pattern, suddenly it resembles a common mistake in context, and it wanted to put it back. So it just oscillated between 'looks like they meant to do something more usual' and 'looks like they made a mistake by applying a common pattern incorrectly'.

Of course, telling that their human generated initial code explicitly avoided the pitfall and the human still shrugged and hit 'accept' when the GenAI said to modify this code.

[–] Zephorah@discuss.online 4 points 1 week ago (1 children)

Basic spellcheck, love it. Anything else introduces unwanted fuckery into the verbiage.

[–] mrgoosmoos@lemmy.ca 3 points 1 week ago

basic spell check requires zero AI, tho

[–] Marshezezz@lemmy.blahaj.zone 8 points 1 week ago

It’s like they’re trying to drive away all their users

[–] randamumaki@lemmy.blahaj.zone 5 points 1 week ago

Fortunately LibreOffice exists and doesn't do this shit.

[–] mrgoosmoos@lemmy.ca 5 points 1 week ago

kind of surprised companies don't push back on this more

the amount of time my company loses because I have to figure out how to turn off AI shit that fucks with my work is non-zero, paid at overtime rate, and has real consequences on project timelines. I can't be the only one

it's like auto-save being on by default. that almost cost the company tens of thousands of dollars because somebody opened a customer document, made a change and saved it, and the rest of us assumed that's how it came from the customer. it was only caught because I noticed the version history and got curious

[–] D1re_W0lf@piefed.social 4 points 1 week ago

That approach hasn't exactly been universally loved. Critics – including folks over at Mozilla As if Mozilla is not trying to follow their footsteps integrating AI… 😒

[–] lechekaflan@lemmy.world 4 points 1 week ago

Goddamn enshittification.

[–] aesthelete@lemmy.world 4 points 1 week ago* (last edited 1 week ago) (1 children)

I wonder what we'll think when we look back at this era....five years, three years, or even a single year from now.

[–] titanicx@lemmy.zip 2 points 1 week ago

It will either be normal, they will look at us like we look at office workers from 20 years ago. Or they will cheer the fight we fought against LLM.

[–] Bluewing@lemmy.world 2 points 1 week ago

This is why I use AbiWord and Gnumeric for the basic office tasks I need. I don't even need LibreOffice anymore. It's so freeing, like running through a green pasture with pretty flowers growing in it.

[–] Aeri@lemmy.world 1 points 1 week ago (1 children)

It shocks me that people still use Microsoft office when libre office is free and doesn't have any of the enshittification

[–] bridgeburner@lemmy.world 1 points 1 week ago

Prolly a lot of enterprises where you as an employee have to live with the software your company provides and don't have the rights on your laptop to install anything else.