Anti-AI

127 readers

9 users here now

Fed up with seeing Generative AI trash everywhere? Us too. This is a community specifically for those who hate Generative etc AI and want to put a stop to it.

Rules:

Absolutely No LLM, ML etc regurgitated trash. If you do so that will result in an immediate ban.
No Doxing
No transspeciesphobia
No pluralphobia
No Islamophobia
No anti-Semitism (anti-zionism and anti-israel stuff is okay)
No racism
No queerphobia, transphobia
No misogyny
No ableism
No fatphobia
No eugenics
No queer identity erasure (e.g. bisexuality or nobinary identities)

#NoBot #NoAI #NoBridge

founded 5 months ago

MODERATORS

oftheair@lemmy.blahaj.zone

The AI is poisoning FOSS and my sanity (lemmy.blahaj.zone)

submitted 2 weeks ago* (last edited 2 weeks ago) by lyya@lemmy.blahaj.zone to c/AntiAI@lemmy.blahaj.zone

33 comments fedilink hide all child comments

Hi. I'm neurodivergent and I love computers. I probably emotionally depend on them a lot besides other interests like art. FOSS and Linux world felt like home. I do not use(as little as possible) any proprietary(or related to AI in any way) software. I feel raw disgust towards those. I feel like all stability in life is gone now that Linus has AI "brainrot". What can I even do? What can others do? "Fork it"? "Rewrite it"? Okay, I'll wait like 35 years. But being serious I feel lost. BSD family doesn't feel like a solution for me. Everything else is experimental at best. I'm certainly not touching Windows or Mac ever again. Should I consider running something like older Debian exclusively to at least "conserve" myself until solution comes up or things change? I almost never post, but I ask you, the digital void, for advice. I don't want to feel this way.

you are viewing a single comment's thread
view the rest of the comments

[+] oftheair@lemmy.blahaj.zone -6 points 2 weeks ago* (last edited 2 weeks ago) (1 children)

did he really drink the coolaid?

Yes

Edit: Though the Kool-aid thing was a misidentification, it should be "drank the Flavour Aid"

[–] applebusch@lemmy.blahaj.zone -1 points 2 weeks ago (1 children)

ugh gross hes using it for code review and finding bugs? thats like asking their review process to be worse.

[–] TheHonorableT@lemmy.blahaj.zone 4 points 1 week ago (1 children)

I mean, shouldn't a bug be patched regardless of who finds it? Some bugs are vulnerabilities, especially if a single AI code review can find dozens of them. If the kernel maintainers don't find them first, then security goes out the door in short order.

[–] applebusch@lemmy.blahaj.zone 2 points 1 week ago (1 children)

lets assume a bug exists, ignoring for the moment the ambiguity in what makes something a bug at all. you want to find the bug but you dont know if it exists or not. you use an ai tool to try and find it and theres a chance it finds it correctly, a chance it doesnt find it, and a chance it finds a bug that doesnt exist. you have no way of knowing it missed something without doing a proper review manually with tests against requirements and stuff, and you need to evaluate each bug it finds to see if its a real bug or not. then you need to decide what to do about it, which could potentially introduce new bugs.

the problem with using ai tools for this is not whether it can find bugs you would have missed. the issue is that people develop an intellectual laziness when they start to rely on these tools. they start to trust them and devote less effort to the activity the tool is replacing. in the case of code review and finding bugs, they might see that the ai says its fine and take the easy path of assuming its right. they might trust that a false bug is real without really thinking about it. worse still they might just apply the bug fixes that the tool recommends.

the insidious thing about software bugs is sometimes they are subtle and you dont figure out they exist until the software interacts with the larger system its embedded in. ai tools might correctly find easy local bugs, but it often struggles with context, which is required to find and fix subtler system level bugs, so some of the fixes for local bugs might create larger system level bugs that a person doing their due diligence and thinking about the software from a system level wouldnt.

i could go on but suffice to say the issues related to ai tools in software are subtle and insidious. its not obvious the ways they fail so developers will see them do what seems like really good work at a local level that ends up being hot garbage at a system level. theres a lot of nuance and the people defending ai for software seem to always be ignoring it, which is concerning.

[–] Blue_Morpho@lemmy.world 2 points 1 week ago* (last edited 1 week ago)

they start to rely on these tools

The entire history of computer development is based on relying on tools to be lazy!

so some of the fixes for local bugs might create larger system level bugs that a person doing their due diligence and thinking about the software from a system level wouldnt.

If people were perfect, the bug wouldn't have been there to begin with! Yet you expect the person to find an obvious bug that they didn't see but AI did find and also see the subtle bug that AI couldn't see.