Instagram story viewer> @chiefgyk3d> Post
21 hours ago
I wholeheartedly agree with Linus here. Especially the part about expecting people to understand what they are doing.

Using AI doesn’t absolve you of responsibility for your code. You still need to review it, test it, understand the permissions it asks for, and explain why it works. If your entire explanation is “the AI said so,” you haven’t done your job.

But dismissing someone’s work purely because AI was involved, without even reviewing it? What exactly are you evaluating at that point?

I use AI in my projects. I also challenge it, reject suggestions, and make it justify decisions. Sometimes it gets things hilariously wrong. Knowing enough to catch that is part of using the tool properly.

There are legitimate arguments about AI, including privacy, resource consumption, and corporate control. I’m not asking anyone to pretend those concerns don’t exist. But they don’t automatically tell you whether a particular piece of code is good, secure, or useful.

If you find a flaw in my work, show me. That gives me something I can fix. “You used AI” tells me nothing about whether the damn thing works.

#Linux #LinusTorvalds #OpenSource #AI #SoftwareDevelopment
18
Posted On: October 16th 2026, 12:25 am
I wholeheartedly agree with Linus here. Especially the part about expecting people to understand what they are doing.

Using AI doesn’t absolve you of responsibility for your code. You still need to review it, test it, understand the permissions it asks for, and explain why it works. If your entire explanation is “the AI said so,” you haven’t done your job.

But dismissing someone’s work purely because AI was involved, without even reviewing it? What exactly are you evaluating at that point?

I use AI in my projects. I also challenge it, reject suggestions, and make it justify decisions. Sometimes it gets things hilariously wrong. Knowing enough to catch that is part of using the tool properly.

There are legitimate arguments about AI, including privacy, resource consumption, and corporate control. I’m not asking anyone to pretend those concerns don’t exist. But they don’t automatically tell you whether a particular piece of code is good, secure, or useful.

If you find a flaw in my work, show me. That gives me something I can fix. “You used AI” tells me nothing about whether the damn thing works.

#Linux #LinusTorvalds #OpenSource #AI #SoftwareDevelopment
Comments (18)
freezen_cold 7 hours ago

freezen_cold

@celebha_pro739 you evidently are not a software engineer because that statement could not be more wrong.

zaaaaaaaaa.m 9 hours ago

zaaaaaaaaa.m

Wait until people understand that a compiler can be poisoned. And LLMs seem to be exactly that.

russell_2s2l 9 hours ago

russell_2s2l

Well I guess AI is as reliable and consistent as compilers now

voidwalker_13 11 hours ago

voidwalker_13

I mean, there's still difference between asking ai to assist making components and describing it as completely as possible and then reviewing it and "make me an cookbook app with account and import/export feature, make no mistake"

deadlyhandsomeman 13 hours ago

deadlyhandsomeman

Right! The concern we should have is simply,: " Is it ethically sourced". Are there any hidden costs to using ai. Other than that, ai is spectacular. Its magic

ulgenwanders 13 hours ago

ulgenwanders

I would like to introduce you to concept of determinism. An LLM is not a compiler.

giacomoaragnetti 15 hours ago

giacomoaragnetti

The only real thing is "Don't use AI to do stuff you're supposed to do to learn". If you do your assignment with AI you're actively going against your objective.

connorhagans 20 hours ago

connorhagans

“AI slop” is such a diluted term now because too many people call everything slop, it’s just a TikTok buzzword now

tom.wilcox404 20 hours ago

tom.wilcox404

People now just hate AI because of the propaganda they see on tikky tokky

charles_heimke 21 hours ago

More posts

back to up