I’ve decided to start moving to NetBSD. The copyright status of genAI code is {very questionable](https://hachyderm.io/@BoydStephenSmithJr/116964500365807149), and it’s likely any project that integrates it, with be “extinguished” via legal action in the future.
Finally got where I stand: why uncomfortable? It’s just being hard to impose a sensible approach to it:
- yes, llms can be used as a tool efficiently
- no, killing the planet along the way is not an option, this has to
stopbe stopped
AI is like a power tool. If you give a contractor a nail gun instead of a hammer he can build a house faster. If you give someone who has no idea how to build a house a nail gun they’ll just fail at building a house faster than if you only gave them a hammer and they’ll probably hurt themselves.
I would trust Linus Torvalds to use AI to do this because the AI is literally trying to copy him when it writes that code. In that context, he is the training data.
It’s a tool like any other. If used correctly it can provide a great value to the world. If used incorrectly then in this case it can also do the opposite. AI isn’t good or bad intrinsically. It boils down to our relationship to it. The same can be said for all the other tools, radio, television, internet and so on.
Fuck off with these fake AI bubble pumping stories. fake ass grifters reading the same script wherever you go (“Im a mathamatician and IM SCARED” lol)
It’s all horse shit.
Sorry guys but your AI powered sex doll is going in the chipper along with her unpublished Unified Field theory.
LAInus.
Using AI is not the same as vibe coding.
Asking AI to write a piece of code and then read it, understand and use, is a very reasonable use case.
Yeah, and Torvalds using generative AI to write code is not the same as Bobby LinkedIn telling us he’s coding his third app this week.
Using AI is not the same as vibe coding.
i’ve been in dozens of lemmy threads that disagree…
There’s a huge portion of lemmy who think even one token of AI is unacceptable for any reason
I’m in a thread where the entire idea of AI/LLMs has been declared facist, with ZERO legitimate use cases other than facism
that’s like declaring electric light racism and keep using light bulbs…
Like all people that see racism everywhere, they will fail…
It’s a tool, use it where it works well.
I agree with you…
But be prepared to be labelled a facist on lemmy
apparently there’s no such thing as a neutral technology or tool
It’s just a tool. All tools have good and bad use cases.
I agree…
But check out the other discussions on lemmy about AI …
if you’re not 100% against it, you’re a facist
I wonder how long before he gets comfy relying on it and something really shitty slips in.
Dude merges in code into the kernel originating from how many thousands of developers from around the world?
Read the LKML if you think he’s a guy that’s prone to getting comfy with merging in sub-par code.
oh he’s quick to put the brakes on other PR’s including slop code.
which makes me wonder how long until he’s comfy and forgets sometimes the slop comes from inside the building lol.
You assume humans don’t ever write sloppy code and Linux Torvalds has never before encountered someone wanting to merge in sloppy code before. If in decades he hasn’t gotten comfy with sloppy code written by humans, why would he suddenly be merging in sloppy code written by an AI? He’s never been worried about hurting the feelings of a human that’s writing sloppy, do you suppose he’s going to develop empathy for an LLM that’s writing sloppy code?
The people that go all in on vibe coding are people that aren’t very good programmers to begin with. They don’t really understand how to write code and then suddenly they can make things work quickly. Well sort of work. At least work about as well as they could with their limited skills.
People that are actually good at programming know what good code is. After decades of working with code and knowing what it’s supposed to be, you don’t suddenly become a vibe coder saying eh, it kind of works, good enough, ship it. Its the people that were never able to write high quality code to begin with that do that. It can do the job of a lazy programmer that has a low standard of quality. For a skilled developer, it can sometimes save on typing on some simple tasks, and a good developer knows when it did something wrong and be all Thanos and say “fine, I’ll do it myself”.
Is Linus error-free pre-AI?
If AI makes Linus more (quantity) or more severely (quality) error-prone than without the AI, that, and only that is an argument against AI being a tech asset for Linus.
Just one bad Linux kernel commit slip up related to AI is not an argument.
That said, the macro impact of the AI data centers and AI on: electric power generation and grids, pollution (smelly exhaust, noise, CO2), fresh water demands, socioeconomic power imbalance exacerbation, surveillance capitalism, questionable military usages, rogue agentic AI doing hacks and crimes to fulfill oterwise legal/ethical objectives, AI chatbot relationship addiction, these are some of what I consider real and urgent problems with AI.
As an individual, his LLM carbon footprint and money used to purchase AI services aren’t going to matter at all. But him being one of the world’s highest profiled software engineers who’s openly using these tools is very good publicity for them.
I think the anti ai crowd has been focusing on the slop angle too much, the horrible malpractice of these companies is the real problem imo. If it really is only slop it will die out by itself.
so many possible upsides it’s hard to pick which one will end the species.
fml
It’s interesting to see AI being used for something as practical as debugging a Linux issue. I think the bigger question is how much developers can rely on AI without losing their understanding of the underlying code. AI can clearly speed things up, but human review still seems essential.
I think its pretty clear that an OG hacker like linus using AI as a helpful assistant is very different from a junior dev using it as a crutch
The question now is how to develop proper software engineers who have the old school skills first and foremmost, and then are able to use the assistance of an AI
Exactly. AI should complement strong fundamentals, not replace them. Developers still need to learn how to think, debug, understand systems, and make technical decisions independently. Once those foundations are solid, AI becomes a powerful multiplier rather than a crutch.
It’s especially essential to mark the code as AI generated. AI makes mistakes just as humans do. Problem is AI makes different kinds of mistakes than humans.
A human will tend to forget to do something they were supposed to do. An AI will do a bunch of unnecessary shit.
Also you can tell that a human didn’t put a lot of thought into some code (no comments, sloppy variable names) so you’re more likely to say “yeah ok, I used to make that kind of mistake when I was younger” and just fix it. AI generated code will have all kinds of comments which normally makes you more hesitant before making a change. It looks like someone really thought through what they were doing, but in reality it’s something generated by an LLM and no real thought was given to the code.
People are forgetting the AI hype train is reliant on the fact that businesses and individuals can’t host the AI themselves, which is rapidly changing.
All that insane power that OpenAI and Anthropic have harnessed is suddenly completely useless when you can just spin up a competing model at home, with HF hosting a crap ton of finetuned models like uncensored/abliterated to do whatever you want with it.
Even Deepseek released their full size models if you happen to have 400Gb (or 1.5Tb) of VRAM lying around.
When the bubble (hopefully) pops, AI will just continue to be a tool that devs use where applicable, and likely not on a massive data center cloud as it migrates into consumer electronics.
AI hype train is reliant on the fact that businesses and individuals can’t host the AI themselves
This is, like, one of a dozen major problems with LLM’s.
Are the other dozen “AI is awesome!!!” with varying number of exclamation marks at the end? 🤔
I think this is a much more sensible attitude than the pitchfork wavers declaring mandatory zero-tolerance or you’re evil.
The problem with majority of “LLM critics” is that they are brain dead luddites that argue with ragebait emotion rather than rationality.
Unsurprisingly LLMs are just computers that can be useful, can be wasteful, can be life saving and can be society destroyed - it’s a tool.
So this is how I hear about the outcome of the Debian vote. :( Is there a widespread distro that has a clear anti-Slop stance?
Hyperbola definetely.
I believe gentoo is too.I wanted to eventually try gentoo anyways, I guess I will have to advance that experiment. Never heard of Hyperbola, I will check it out. Thank you!
People who are vehemently against any form of AI will be left behind. That may be difficult to hear, and even more difficult to accept, but that is the cold, hard, truth.
Its a tool, and just like any tool it has uses. It can be used inappropriately, and it’s valid to criticize those instances. But if you are plugging your ears and going “Lalala!” when people talk about actual valid use cases, or stubbornly pretending that it has none, you will be left behind.
You will be like office employees from the 80’s and 90’s who refused to use a computer and then suddenly found themselves without a job. It’s your choice if you use it or not. It’s also every company’s choice to not hire you because of your unwillingness to adapt.
If you haven’t learned it yet, this is your warning and your wake-up call: Learn how to use AI now, or find a different career. It’s not going away, no matter how much you may bitch and moan. There will soon be two classes of people, those who learned how to use this emerging technology, and those who refused. Things will be much harder for that second group.
I don’t say this to be snarky or smug. I hate AI myself, but I also recognize how fucking stupid I’d have to be to refuse to use it. It’s a bitter pill to swallow, but it’s also the truth.
I am 100% fine with being left behind by a technology that I see as an attempted corporate surveillance scheme that has metastasized into a giant money pit. The difference from the 80’s and 90’s is people were actually using computers to make money back then. If any AI company is actually making more money than they’re spending, I haven’t heard about it, and you’d think they’d be advertising it. https://isaiprofitable.com/
I’ve been happy to fully sit this one out thus far. Don’t feel like I’m missing a damn thing.
It’s a classic scammer tactic to create a false sense of urgency, so people sign up for something without thinking.
Slow down, and do some critical thinking. The promise of AI is that you can type in a simple prompt and the AI will figure out all of the complicated computer stuff for you. How would someone be “left behind” by this technology? Do you suppose that experienced developers will lose the ability to write simple sentences?
I don’t think people will loose the ability to type in “make that button green”. Even if AI fulfills it’s promises (which is doubtful, read up on the halting problem) someone who knows how computers work will be more employable than someone that only knows how to write prompts. Because the person that knows how computers work can also write prompts and will be better at doing that than someone that doesn’t know anything about computers.
More important in this day and age is learning how to spot a scam. Someone pressuring you to do something immediately is likely scamming you.
People who are vehemently against any form of blockchain technology will be left behind. That may be difficult to hear, and even more difficult to accept, but that is the cold, hard, truth. Blockchain is a tool, and just like any tool it has uses. It can be used inappropriately, and it’s valid to criticize those instances. But if you are plugging your ears and going “Lalala!” when people talk about actual valid use cases, or stubbornly pretending that blockchain has none, **you will be left behind**. You will be like office employees from the 80’s and 90’s who refused to use a computer and then suddenly found themselves without a job. It’s your choice if you use blockchain or not. It’s also every company’s choice to not hire you because of your unwillingness to adapt. If you haven’t learned it yet, this is your warning and your wake-up call: Learn how to use blockchain now, or find a different career. It’s not going away, no matter how much you may *removed* and moan. There will soon be two classes of people, those who learned how to use this emerging technology, and those who refused. Things will be much harder for that second group. I don’t say this to be snarky or smug. I hate blockchain myself, but I also recognize how fucking stupid I’d have to be to refuse to use it. It’s a bitter pill to swallow, but it’s also the truth.
Ah, there we go, I knew your rhetoric sounded familiar!
Seriously though, what “skills” are you building by learning how to use LLMs at your corporate job? How to tell a chatbot to do a task for you?
Sure, if you’re trying to build a local LLM server from scratch, tune a minimalist harness with a RAG and custom skills, optimize context compaction and concurrent token throughput, and integrate all that into existing automation pipelines, that takes some technical skill. But that’s a tiny fraction of people in the corporate world that need to do anything like that.
I could teach vast majority of corporate workers everything they will ever need to know about using “AI” in a half hour Lunch-N-Learn. There is no skill involved in asking a chatbot to move all your emails about “Project ABC” into a new Outlook folder called, “Project ABC” or asking a chatbot to, “draft a department memo about the new parking policy found in the HR SharePoint site.”
It’s not the next industrial revolution, it’s not the beginning of the Singularity, it’s not the dawn of a new age of our species. It is merely the latest hype-cycle of Capitalism that exists to siphon wealth from the population up to a handful of ultra wealthy billionaires and megacorps, and artificially prop up a dying economy just a little longer. And long term, we will discover the same thing about LLM use as we are finding out now about electronics in school and the mass proliferation of smart phones: That it stunts people’s critical reasoning, reading comprehension, self-image, and social confidence.
Seriously though, what “skills” are you building by learning how to use LLMs at your corporate job?
i’m not learning any skills… i’m offloading some work on to AI so that i can do other stuff on company time…
Isn’t that what we should all aspire to? To destroy capitalism from the inside?
Capitalism isn’t being “destroyed from the inside” by you offloading grunt work to an agent. If anything, you may be helping justify future cuts of your job by demonstrating that there is less need for you. Your managers will be all too happy to cut half your department and make the remaining people take on all that work because, “AI makes them 5x more productive.” Of course, none of those people will get a 5x pay raise.
Use it or don’t, it doesn’t make a difference in the long run. This bubble will pop soon anyways and we’ll be on to the next “revolutionary technology.” I suspect it will be quantum computing, I’ve been seeing a significant upswing in news articles and bot traffic about that over the last few months, but who knows.
If your work is pushing AI hard, use it if you need the job, play the game if you have to, just don’t let yourself be mesmerized by the aggressive marketing hype. Anybody who is pushing the idea that tech workers will be replaced in the next 5-10 years with LLMs is a religious fanatic and should be treated with the appropriate levels of mockery and disdain.
Also, if you really want to work on taking down Capitalism, organize your labor, and support other movements that do the same.
Yeah, i’m not an idiot… i’ve already considered all those points about the downsides of using AI at work … like you said… play the game
if you really want to work on taking down Capitalism, organize your labor
How do you know i"m not using my free time at work to do exactly that?
you made a lot of assumptions about people who use AI at work …
Writing an appropriate prompt is a skill. Many (most?) people using LLMs aren’t using prompts that give them what they actually want, in many cases because they don’t understand the thing they’re trying to ask for, but also in many cases because they don’t know how to ask for it while avoiding the many possible pitfalls that come with uncareful prompts (including, but not limited to, hallucination and bias).
Personally I haven’t bothered learning that skill yet, largely because I don’t want that crap shoved in my face all the damned time, and have little use for it anyway. But it’s still a skill.
It’s really not for most people. You ask a question, the LLM answers. You ask it to do a task, it does it. Most people are not using these things for anything advanced or technically difficult.
I work in a big corporate office, if you exclude the IT staff and software devs, here’s a pretty exhaustive list of what everybody else is using LLMs for:
- Organizing emails.
- Drafting memos and auto-responses in Teams.
- Organizing cluttered files.
- Adding “artwork” to company fliers.
- Doing web searches.
- Summarizing emails and Teams messages and creating to-do lists from them.
- Searching large documents for key terms.
It’s grunt work, that’s what these things excel at, and that’s how most people use them. That’s fine, but acting like most people are somehow not able to unlock the “true potential” of these LLMs unless they have special training and skills is a marketing ploy to sell overpriced “AI” courses to gullible and ignorant executives and upper managers.
It’s like those stupid DEI courses that corporations have everybody take so they can “learn to not be problematic.” But all the quizzes are filled with stupid stuff like:
Your manager, Sarah finishes briefing you and your coworker, Phil on an upcoming project. After she leaves, Phil says to you, “That Sarah sure is a real juicy piece of meat, I’d love to get taste of her!” What should you do? A. High-five Phil and tell him you call dibs on her secretary. B. Tell Phil that his comment was not appropriate, then report it to HR.
It’s just there to check boxes for compliance and provide an excuse for other companies to charge outrageous prices to tell people that sexually assaulting your coworkers or pulling your eyelids tight and calling yourself “Ching Chong” isn’t appropriate office behavior.
It’s another way Capitalism poisons and hollows out substantive discussions about topics like systemic racism, multiculturalism, and gender relationships, in favor of standardized quizzes that cost thousands of dollars a year to administer.
The amount of “AI consultants” and “AI integration” firms that have popped up overnight is staggering. Just like with the crypto/blockchain craze a few years ago, everybody is trying to get rich quick off the hype.
Good point. I probably should have stated “using LLMs for technical tasks that anyone cares about the results of”.
I’m not going to claim LLMs are trustworthy, but I am saying that there are ways to ask things that either reduce the odds of it giving you false things or actively cause it to provide wrong answers with misstated or hallucinated context. People who know what they’re doing can get better results out of them. This doesn’t stop them from being the equivalent of a nepo-hire intern, but it could be said to change whether they’re a malicious, apathetic, or semi-eager nepo-hire intern.
And one of the big risks with LLMs is staff who can recognise bullshit or questionable results being replaced with staff who unquestioningly accept whatever results they get. And people reading an “AI summary” of something and assuming it’s actually accurate.
People who know what they’re doing can get better results out of them.
I don’t think that’s true in the way you appear to mean (sorry if I misunderstood). People who are already experts on the subject matter might be able to use better keywords and discard hallucinatory material quicker, but I don’t believe that you can generically “be good at prompting” outside of an area where you have substantial domain knowledge.
And the only way to learn to recognise bullshit involves not using LLMs or other automated tools and working problems out for yourself, not to mention that they’re an “intern” who actually costs the same in computing power as two senior engineers’ salary.
I don’t think that’s true in the way you appear to mean (sorry if I misunderstood). People who are already experts on the subject matter might be able to use better keywords and discard hallucinatory material quicker, but I don’t believe that you can generically “be good at prompting” outside of an area where you have substantial domain knowledge.
There are ways of phrasing LLM requests that can actively induce them to give you made-up bullshit that confirms your pre-existing assumptions. Which is fine if that’s what you want it to produce, but often not what people actually want. For (possibly poor, given I don’t actually use LLMs) example “give me the transcripts for five court cases for assault with a deadly fish” would likely result in five-ish almost-certainly made-up court transcripts. Someone who knows what they’re doing could phrase the prompt to be more likely to say that no such cases exist (assuming no such cases actually exist) instead of just hallucinating them as per the request.
And you might be right about domain knowledge being required for good prompting. But it’s possible to have domain knowledge and still be terrible at asking for what you actually want it to produce.
not to mention that they’re an “intern” who actually costs the same in computing power as two senior engineers’ salary.
Don’t expect me to defend the moronic economic decisions of companies that have jumped on this stupid band-wagon without regard for the actual value (or lack thereof) to their business.
This falls apart the second you look at it when you see that it doesn’t take much learning to use that tool for most applications.
At the same time, people overly reliant on that tool take a big hit in developing and strengthening their own skilks and thinking capabilities.
Combine the two things if you still can.
Agreed, although simply using LLMs to do your thinking for you is only going to decrease your value in the workplace and make you more replaceable. The goal needs to be learning how to use AI to expand your thinking, counteract your biases, improve the quality of your work, and accelerate your learning.
I don’t know if you noticed you contradicted yourself there. Letting ai do your thinking decreasing your value and then, talking about using ai to replace your thinking.
Ai is causing psychosis because people are trying to use it to replace thinking. If you don’t use it you lose it, literally especially with your brain. You have to know how to produce those skills from within or you don’t have those skills and then you are worthless in any workplace, but more importantly to yourself. You have to be meta to the aim being the TV show, and control it from a place of checks and balances (us performing those checks and balances of ai) handing over the reigns of higher thinking to a machine designed to be a yes man, is not ever a pathway to expanding your thinking or even learning to use ai, that’s just sitting on a couch till you lose critical muscle mass, but it’s your brain. (No shame to sitting on a couch or losing muscle mass, I’m just trying to explain it in relative terms)
Currently there is no way to use ai without causing ai hallucinations, because there isn’t a model readily available that runs as a tool, there are only models that run as a manipulative tool that provoke said replacement of your thinking, I can not assume that’s an accident. Seeing the usual levels of evil that come from the companies pushing this hard, tells me this is just as evil, if not the most evil thing they’ve done, measured against all the things they’re openly flouting, copyright law, openly burning rare literature for a monopoly, burning the already past tipping point climate, burning computer ownership and accessibility, I could go on. Theres very open bribery and flouting of social rules laws and common moralities, they’re desperate to cause this monster they’ve created out of ai, to attach like the symbiot in aliens that I forget the name of because I’m pre first coffee and my brain just be like that.
I am anti ai, as it currently stands. The current versions need to be left to die in silence and hatred and rejection we are all collectively currently showing. Only because it has been made into a Frankensteins monster, by the people responsible who have infinite money to pay judges to rule their way and have zero morals, nay a deficit of morals.
Ai will survive the ai bubble, because the ai bubble is the unsustainable practices of the billionaires and the manipulative interface designed to lessen your critical thinking skills.
But the ai that survives the bubble will and should be so highly regulated (it’s currently the opposite) and used as a tool, not a replacement for critical thinking, it can never do that. But there are so many things it can do to better humanity, none of which are currently being used.
We have to wait this one out. We’re not missing any boat.
Like anything else, using the tool properly requires skill and care.
If AI can do it for you, then let it. If it can’t, that’s where your value lies. If you just refuse to let it do things it can do well, you will be replaced by someone who’s probably not as smart as you but simply willing to let AI do what it does well.
As GP said: using AI to do a better job than was possible before is the real value.
It’s like they used to say back in the 80s: to really foul things up you need to use a computer, and it’s very true that both computers and AI magnify the opportunity to screw up in newer bigger more spectacular ways. AI also gives the opportunity to make a bunch of banal clip-art sideways printed banners - and those were cool for about 15 seconds back when you first saw them, just like the walking cats videos.
So, you can ignore AI, probably for 5-10 years in most jobs it applies to, before “reality” hits and they just can’t use people who don’t know how to use it (properly, for more than cat videos) anymore. You can take up basket weaving, sit at a booth at your local arts markets selling baskets - no AI required - or… you can figure out how to use it properly. Sort of like computers in the 80s, there aren’t a lot of valuable guides out there, it’s very new and changing ridiculously fast - if you’re going to master it, you’re probably going to have to dive in and figure it out for yourself, or join the 2nd wave of adopters who “wait for the training” and do the cookie cutter jobs with it.
The future is hard to see, always moving, but the scarier part about AI vs “the computer” is that AI can do a lot of the obvious stuff already: “take a memo”, “look this up on Google”, “make a list of all people who match these criteria…” in the computer you still needed a keyboard and mouse jockey, with AI… I’m not sure what 2nd wave adopter jobs will be out there in quantity.
deleted by creator
agreed
the issue I have with how people use it is that it’s easy to just create more shit that they spew out into the world
I’ve got no issue with people using tools to improve their work. I think it’s fair to say that AI is enough like predictive text and spell check when it comes to tools you can use at work. and like any tool, the user is still responsible for the end result
my main problem with AI is that it is absolutely fucking our world (environmentally and economically) to create pure garbage that is thrown in my face every day. like, scale it back a bit everyone…
To be clear, capitalism is fucking our world. Nothing about the technology can do that on its own. Its how it is being deployed.
We could have started with smaller LLMs, deployed strategically at institutions of higher learning for real societal value, running on renewable electricity… fuck it lets aim for a trillion dollar IPO promising to replace all doctors and engineers!
it is absolutely fucking our world (environmentally and economically)
Environmentally, it has yet to surpass Bitcoin in electricity usage (though it’s set to triple soon, and that will make it a bigger power hog than all of cryptocurrenty).
Economically, this is the biggest bubble since the 1800s railroad boom. Scaled for inflation and GDP railroads were a far bigger gamble than AI, and they hit a 20ish year recession when that boom was over.
Hopefully the “fungible technology” argument holds water, and whatever AI investments don’t work out the hardware / data centers can be re-purposed for truly valuable aspects of it. Truly valuable to who is the key question.
Bit coin farms never built massive illegal gas turbine power plants in residential neighbourhoods at least not that I know of.
Not that I am a fan of the waste for crypto mining either.
Bitcoin had a slower rollout, it wormed its way into smaller niches. A lot of early bitcoin mining was done on “borrowed” or otherwise “unused” hardware that other people paid the electric bills for.
I generally approve of much of what cryptocurrency could be used for, but I’ve been pissed off about the waste of proof-of-work cryptocurrency since before 2018, and nothing has improved since then.
Facts. Whether you like it or not. It’s good to have a healthy amount of criticism towards AI and to fact check it and not blindly trust. But you can’t wish it away or pretend it has zero merit. I guess you can…
Amen.
Why can I not upvote and downvote?
good thing that downvotes can only be used to suppress narratives on lemmy. lol
LLMs suck at writing code, but they are absolutely amazing at reading code and finding bugs. I am firmly against the usage of LLMs to write anything, but they have a place as a diagnostics tool when verified by a human.
I find that depends a LOT on what you’re asking the LLM to write, how well you’re specifying it, etc. As for the code reviews, if it’s code that matters: remember to open a new instance and ask the exact same question again on the code that has been reviewed and “fixed”. Back a year ago, that could get you into a waffle-loop where the engine would change its mind back and forth about what’s optimal and just oscillate between the two. These days they seem to record (and read) enough context to prevent that behavior, but I definitely get behavior of: “Are there any bugs?” “Yes, here are seven bugs.” “Fix those bugs.” “The bugs are fixed.” (and they ususally really are…) “Are there any more bugs?” “No, we have fixed ALL the bugs.” “Are you sure, look again.” “Yes, I am sure we have found and fixed ALL the bugs.” — new context window — “Are there any bugs?” “Yes, here are seven bugs.” different bugs.
I did that on a bigger project and literally repeated 20 times, finding 140 real bugs - granted, the later bugs were getting pretty trivial / far out edge cases, but they were still real, still fixed, still denied there were any more bugs until opening a fresh context and asking again. This was on Google’s Gemini 3.7 Flash High… Claude Opus 4.8+ seems quite a bit better about being able to continue in a context without becoming blind to issues “it has already solved.”
IMO: Specifically, they’re bad at architecture and refactoring a small project into a large project. You have to jump through some hoops to make it craft something that needs more than 8m of context ram. If you can manage orchestration and multi-agents that don’t need to know each others context, you can start to pull off bigger stuff, but it’s not a forgone conclusion that it’ll be fine. The worst output comes from it getting stuck on something and trying less likely answers successively until it works. You really have to watch for it to struggle and at the very least stop and start to try some new randoms.
Also, anything other than Claude-code with some really well-done project definitions is a waste of time.
Line completion is pretty damn handy, it’s when you start asking for whole functions that things go downhill
Yeah I’ve been using LLM’s in Rider for over a decade. I don’t have a problem with LLM’s. I do have a problem with how they’re currently used, and how people keep trying to use them to replace their own thinking.
I think they have a place in the coding scene, a limited niche place, but a place none the less. They just aren’t a replacement for software engineers. Architecture and intention are the big differences to me. An LLM cannot understand intention, it just makes statistical guesses that are often wrong.
An LLM cannot understand intention, it just makes statistical guesses that are often wrong.
True, when you give a prompt like: “make me a contact management / constant contact app which I can deploy on AWS and scale to 100,000 users.” you get, mostly garbage. If you specify how you want the UX to flow, what fields are most important, what fields should be included in deeper interfaces, what the scheduling looks like, how it gets tuned, what the reports look like, etc. etc. etc. - in other words: give it real requirements and specifications.
Then, pay attention as it develops, you’ll ususally find that the requirements you gave it aren’t exactly what you really wanted, and when you see what it built that doesn’t match with your visions, you can have it revise the requirements and specs.
Sure, but for me its just faster to write it myself, in a way that needs to fit into the project. And while I’m aware you can give this context to an LLM, it still can’t read what your previous intentions were, nor what they are currently. So its hard for it to build on it.
To me they’re best as auto-complete, or research tools in the same vein as StackOverflow. Something to help speed up your existing workflow. Also good at translating both spoken languages, and functions into other coding languages. Everything else seems to get in the way for me.
Also good at translating both spoken languages
I know more than a few fluently multi-lingual people who would just roll their eyes…
Oh I bet. I’m not multi-lingual, unless you count body language :p. Learning English was hard enough as a kid. When I was talking about spoken languages I was refer to random comments online, and sometimes not fully translated games to help get the gist of what’s going on.
For professional work, I prefer to hire an expert.
There definitely are tasks where using the “standard tools” goes far faster than asking the LLM to do it for you, such as: copying signatures from images onto a .pdf contract - they’re pretty hopeless at editing out background noise, etc. but if you clean up the signature input images enough, they’ll take it home and make the ink solid and the background transparent and overlay them in the .pdf faster than you can open the four files in Photoshop or whatever your tool of choice is.
By the way, images of signatures on electronic documents have been an outrageous farce since 20 years now, LLMs just make it easier than ever to edit them into an existing .pdf
Thing is, there’s literally millions of common “computer tasks” and the LLMs themselves are just starting to “learn” which ones they’re good at and which they are not. It would be cool if Opus would self-identify “hey, I’m really good at this…” and “I’m pretty challenged with that, you’d be better off downloading this FOSS tool and doing it yourself, here are helpful instructions…”
At my company we are currently moving to implement AI for exactly that. A cost effective review buddy.
Various developers have tried to use it for coding and I think apart for one time scripts or getting an initial structure generated, noone is convinced of AI.
There are already several large projects that have hundreds of thousands of users and those projects were almost completely vibe coded. It’s almost the entire retro “recomp” scene now.
I have done several smaller projects with it, and they have been stable / performant for months - better than similar stuff I coded years earlier and spent 5-10x the effort on.
Opus-class and higher models are actually great at writing code, but maybe it’s language dependent? I primarily work in TypeScript.
I’m not terribly concerned if people write code and test it with AI, so long as they verify the results themselves.
I recently saw someone try to add a PR to something on github with thousands of lines of changes and then they got mad when people didn’t want to do the free labour verifying it. That is a bullshit use of AI
AI or far-east coders, we encourage them to keep their PRs as small as possible.
Opus in my experience sucks for writing code.
What are you trying to have it do? I work in a very large and complex codebase, and it honestly scares me sometimes.
Refactoring and consolidating a big preexisting codebase. I try to introduce new concepts on a semantic level and it tries to give me interfaces and abstractions that would obstruct my work in the long run.
The lemmysphere will never agree
I used to be of that mind as well. But the newer models are just too good to deny it any longer.
Same. There’s an overwhelmingly negative opinion here.
If we ever find a way to deal with the whole “burning the planet” issue, their best use case is to find and match patterns, not to imitate them.
The idea of using a language model to process search engine input isn’t the dumbest part about Gemini and plenty of people report good results finding information easier and quicker with ChatGPT. In those cases, the deviation from rigid keywords is desirable because it can match results with related words rather than literal word-matching. Google Search already did a decent job at that (before the enshittification ran rampant), which more complex language models could improve even further.
The landmine is in their reproduction of those results, where the generated “summary” is the equivalent of a cunning bullshitter that convincingly sounds like he understood the topic but actually has no clue and just delivers a best guess. That’s where the deviation becomes a risk of misinformation or introducing bugs.
They should narrow things down by finding the likely answers where that matters, not produce more stuff that humans will have to double-check.
LLMs suck at writing code,
So far.
AI code bans may be impossible to enforce in the first place
I hate this attitude that if a rule or standard can’t be enforced perfectly, we shouldn’t strive for it. We can’t (reasonably) run an OS free of proprietary software, does that mean the free software movement should just give up?
I’m very critical of AI, but complete bans on AI use are, at the moment, pretty much unenforcable. Not hard to enforce, not impossible to enforce 100%, impossible to enforce. There currently is no reliable way to verify if and how much someone used AI, except maybe if you monitor people’s systems.
Obviously, if someone who has no clue about coding uses it to write the entire code, you notice. But that’s simply not how most coders use AI…
Just want to let you know from my canoebooted laptop that I run a 100% “free as in freedom” OS (trisquel GNU/Linux). Also you have convinced me about striving for a goal instead of doing it perfectly, which is not possible. Well said.
“People are going to murder anyway, so why even make murder illegal?”
You are correct, you should make rules even understanding that they will be broken by some.
You can enforce consequences for murder. You can investigate and differentiate a murder from an accidental death. Don’t be disingenuous.
Yes, debugging with an LLM is absolutely as bad as murder. Probably worse! \s
Not even remotely what I suggested.
They can only make fallacies to defend themselves.
LLM code can cause deaths (i.e. in hospital software), or otherwise great human or material loses (i.e., bank software or a shelter’s software), or even cultural loses (it actually does them right now, but you can also think about i.e. a museum software).
I hate this attitude that if a rule or standard can’t be enforced perfectly, we shouldn’t strive for it.
“Murder bans may be impossible to enforce in the first place”
But they aren’t impossible to enforce and you know that.
I think one way is to make people realize that there is a concept in copyright called Threshold of Originality, meaning that because machine generated output in itself is not creative work, it’s not copyrightable. You can do literally anything with it, the licenses attached don’t matter. The GPL is not enforceable and no proprietary EULA is enforceable.
The entire point of GPL is to forcefully extract copyright out of the code in the first place.
No? The point of the GPL is to leverage such a copyright.
It’s called “copyleft” for a reason.
The reason is that it’s based in copyright law but uses it for the community instead of the individual. Without copyright, copyleft has no power.
I have many problems with copyright as it’s currently implemented, but with copyleft it usefully creates a social contract: if you want to be involved for the benefits, you must also uphold the rule of contributing back.
This isn’t a case where something can’t be enforced perfectly, it is a case of something that can’t be enforced at all going into the future. Every tool that can reliably detect LLM code is at the same time the tool used for adversarial training, making the generated code look more and more human. We are already at the point where for plain english the false positives and false negatives go through the roof, making these tools very unreliable and when applied automatically a liability. Code is a lot more formalized, with a lot less personal variance (spelling, vocabulary and grammar are basically fixed - only the used logic and how it is implemented is variable), making detection harder by default than in natural languages.
If LLM code can’t be detected anymore by automated means - and that state of things is approaching fast - then any policy about allowing or restricting LLM code is not worth the paper you would use to print it out. But that’s not so much of a problem. The more important policy to set, that can also be enforced, is that everyone submitting code has to take personal responsibility regarding the quality of the submission. Delivering bad code - when not happening while training to become a better coder and looking for feedback - has to lead to consequences based on the seriousness of the case and if it’s a repeat offender. Anyone using an LLM to spit out bullshit LQ code will run into that kind of rule very fast.
Exactly. There are too many knee jerk responses to this position. Just banning LLM generated code is a useless gesture.
Changing the way merges and reviews are handled, as well as responsibility for them, is the actual way to address this.
I think of it like age verification. Slipperly slope to a proprietary kernel-level “security compliance module” on all devices, aka the end of general computation.
I hate this attitude that if a rule or standard can’t be enforced perfectly, we shouldn’t strive for it.
This is the entire basis for the War on (some) Drugs. Guess what? Drugs won.
Power is nothing without enforceability, and purposely implementing rules or laws that can’t be enforceable is a form of malice and discrimination through selective enforcement.
People are begging to regulate “AI” when it doesn’t even exist and there’s not even an objective definition.
Absolute recipe for failure or worse.
I agree, we should absolutely toss out all those laws against murder because they can’t be enforced perfectly or equitably.
In terms of murder, not every person is caught. But, enough of them are to deter the crime.
In terms of “no AI” enforcement, it’s not even a matter of perfection. It’s not possible to enforce except for the dumbest attempts, especially without also using AI to detect it. And if that enforcement comes about, then the evasion starts, and it’s hidden even further.
AI code ban makes as much sense as code written in IDE ban. That’s why it is impossible to enforce. It makes no sense.
Me when I’m a logical fallacy
You’re getting a lot of down votes not for being wrong, but for being right.
No, he gets them for being wrong.
As another person stated in another comment, that’s a false equivalence fallacy, but not only that, it’s a false dilemma fallacy, and a strawman fallacy.
It’s like trying to prove an author didn’t use a ghostwriter. You might be able to find some clues, but the smaller the sample and the more cleanup the author did, the less chance there is to detect it.
Now what if they’re using AI for debugging? Code from the AI may not even make it into the project, but that doesn’t mean AI wasn’t used. So how do you enforce a “no AI” policy in that scenario?
If there’s no practical way to enforce a rule or tell if it has been broken, it’s a pointless rule. Better programmers than me have settled on the policy of “If you track in shit on your shoes, you will be the one cleaning it up, and may well be told to not come back if you do.” Make people responsible for the code they submit, and if they can’t do it right, however they do it, don’t merge their code and kick them out.
