You must log in or # to comment.
I don’t even use it that much. Sometimes as a last resort when searching yields no answers, but then usually it just answers something not quite right so… wouldn’t miss it.
I spent most of yesterday correcting someone’s AI transcribed copy of our surplus sheet because it was so wrong
LLMs are explosively imprecise, statistically luke-warm, grey goo extrusion spinsters, which is why all the logos for ‘AI’ companies are variations of arseholes.
This message doesn’t really land well at all in the way this comic is made. Another post does a much better job of undermining the comment by comparing to a meth addict.
I don’t use any of it now, so…it’d be pretty easy, actually.
I love shit like this.
Like. I work for a landscaping company. There is zero application of AI/LLMs/whatever to any aspect of my work. None of that shit is going to help me move piles of dirt, or plant saplings.
I use Claude Code daily at work and yes, I am VERY willing to give it up if this LLM and generative bubble completely blew up tomorrow. I’d be more than happy to see it disappear.
I am VERY willing to give it up if this LLM and generative bubble completely blew up tomorrow
The best part is that when the bubble blows we’ll still be left with dozens of LLMs that can run locally on modest hardware with results not far behind the “frontier” bullshit, and nobody will be shoving AI down our throats every waking moment.
I guess nobody did any work a few years back when AI wasn’t being shoved down our throats. Also imagine being a carpenter or an electrician and being asked if you can get through your entire workday without AI. That guy is such an idiot is makes my head hurt.
Your second point was where my mind immediately went. I have never used AI at work and have no way to use AI at work even if I wanted to, because I don’t use a computer for work. Even if I was a massive AI supporter, I’d still see exactly zero difference at work.
I can’t exactly remember the one time I used AI (probably two years ago or of curiosity) I can remember the last time I worked without AI (every day since then).
Exactly. I’m a professional, but I use AI in my work now. It makes things easier and faster (by digging though logs for 24 hours and whatnot), but I did that before.
Sure I’d be able to do my job just the same, but getting paged at 2am, I’d have to spend 45 mins figuring out what to do (clicking links, waiting for page loads, parsing 2000 log lines, etc) while AI does it in just 5-10mins. I get to focus on my actual work instead of the bullshit oncall alarms.
I hate the environmental impact, but those that know how to use it and program it are actually benefiting from it.
The environmental impact can be discussed. Maybe training Kimi 2.7 is like sending a rocket to the moon. But using it locally can be optimized and then is as much as running your GPU. Arguably the impact is similar to gaming then.
(I know this is an unpopular opinion because it reads as defending oligarchs. Sorry!)
Seriously. What were these people doing like two years ago and what are they doing now that is different? Everyone in my office that uses a lo of AI is just doing it because they are lazy and don’t care that their work product is shit because it was shit anyway.
Seriously - the AI bubble is very recent. Pretending like no one can possibly remember how they functioned before AI was being used everywhere is asinine.
He’s telling on himself. He doesn’t actually know how to do anything and he never did. He’s terrified of losing a.i. because there’s nothing for him to go back to
These addicts can’t remember how to function without LLMs. They’ve destroyed their brains.
I work as a software engineer and I don’t use any of those things… so yea… if I can do it, most people, the overwhelming majority of people can do it.
I envy you. The company i work at fully embraced ai …
My boss said it out loud yesterday: you know guys, i don’t really depend on you anymore.
He must have wet dreams about developing software for a customer, without having to pay some stupid developers for it.
My boss said it out loud yesterday: you know guys, i don’t really depend on you anymore.
As a poor soul who’s trying to refactor a fully vibe-coded project (in prod BTW) into something half-decent: hahahahahahahah
Sounds like he and AI deserve each other, and he should probably just start a new company with just him and AI to do all the work without those inefficient employees. In fact, he should ask AI if it’s a good idea. If there’s anything he’s unsure about, AI could reassure him to go for it.
Sis fr, I’m a freelancer music producer and I’ve seen my demand fall into the ground, a horrific free fall to the darkest of the bottomless gaps… All that while they just use and praise music that sounds all like if it were just chops from the same giant shitty song.
I’m forced to for work but I could happily go back if it all disappeared.
In fact I wish it would.
If by “forced” you mean they reprimand you for not spending enough tokens, do I have a project for you https://github.com/dtnewman/burn-baby-burn.git
But if it disappeared, how will all those middle managements write all those hundreds of TPS reports that nobody reads?
Or send automated posts of gibberish and feel productive?!?!?
The recipient of the report probably uses A.I. to summarise it. So all the useless reports are actually an indirect and inefficient communication mechanism between different A.I. instances. Maybe we should cut out the middle man(ager)
Has it’s ups and downs honestly, it helps me automate some annoying tasks, but yes it’s expendable, people seem to forget that’s at the end of the day it’s just an another tool.
Ah, yes. Just another tool that is in the hands of people who don’t want to pay humans. Who are stealing art and novels from actual people for “training” purposes. A tool that is chewing up natural resources and straining power grids. It is a weapon being used by those in power to destroy human creation. Not a tool.
I honestly somehow forgot America and isnotReal exist.
for context I have always only used Chinese open source models and yes I got your point there, it’s a valid point.
Let the bubble burst because of the Us Ai is more harm than good to the world, I’m not even that reliant on it.
Did you completely forget about local models?
Yeah, someone can run a local model and then train or fine-tune it with other people’s work, then make it spit a sloppy collage of all it and say is something new and that they did it so it’s of their property.
You don’t train nor fine tune local models, there’s no need. And for me they’re just helping write regular expressions most of the time.
You can actually fine-tune or even train or re-train local models.
Not everyone is doing that but a everyday bigger amount of people is doing so. And whoever that do that is because they want a a model tailored to their needs, that may be too hiperspecific.
Most of these AI-bros do use local AI so they don’t have to pay for it. And in the same line, if they’re “AI ‘artists’” (although the correct name would be AI-sloppers), they’re gonna train or fine tune these models with whatever they want to, without any permission, then they’re gonna sell the resultant collages (literally, collages) on internet while robbing other artists work.
I have suffered that (I’m music producer and I have listened to plenty of literal fragments of music I have manually produced inside and output of some “AI ‘musicians’”, and the worst part is those songs were WIPs I showed to some friends in a private group in Discord). My best friend too (he’s drawer and, again, he has seen literal fragments of his drawings, but also of drawings from artists he follows, and again, these drawings were only shared in a private group).
Can you link me to local fine tuning or retraining? I’d love to lean a new technique
Don’t threaten me with a good time.
I’ve only actually used AI like 10 times I don’t think it was actually particularly helpful in even half of those. It mostly just gave me wrong information it made up and I then had to go back and find the correct answers myself.
I, too, try it every once in a while just to see if I’m missing anyþing. So far, I haven’t gotten any results which change my opinion.
I saw someone else (on Lemmy) saying þey þink anti-AI people are stuck wiþ perceptions from Ancient History, like opinions formed 6 mos ago, and really, AI is good. It’s not accurate; most of us probably encounter some form of AI every day just because it’s being shoved down our þroats, by search engines if noþing else. Þe code is always bad, or wrong, or boþ, and þe search results are lies often enough þat you can’t trust it.
No like, seriously, this comment was really unreadable this time, please stop using thorn
Skill issue
Yeah, the skill of the AI
If we’re going to discuss AI, it might as well be relevant and not from one personal anecdote from someone who actively hates the technology.
We can all agree that’s not a fair take
It’s plenty relevant. The times I mentioned were actively trying to use it, putting in effort to rephrase prompts and whatnot until I hit the cap on the conversation and it still couldn’t give me what I needed, but even looking at the crap google spits out when I search it’s always half wrong made up slop. How anyone can trust this shit is beyond me.
You only “trust” it if you’re turning your brain off.
Anyone who wants to keep their grades or job is reading everything that comes out and applying their human taste, and making serious edits and rewrites
So then what the fuck is the point?
I didn’t have to write those expressions, and I avoided all the effort of going to read the docs and relearn something I very rarely need. Plus the latest models usually nail it on the first try these days.
I suppose they just can’t imagine that there are people who don’t use it.
They must think that everyone is addicted to it as they are.
I used to use ChatGPT and Copilot, but then I stopped because they sucked too bad to justify the price. I spun up an Ollama instance on my server but I almost never use that either because it sucks really bad. I integrated it into my Home Assistant instance and it sucks so bad that when I ask it to simply turn off a light, it only succeeds about 1 in 5 times.
‘You’re just prompting it wrong.’
-The people who say LLMs have achieved AGI
It might not even be a prompt problem, but an overkill to a simple job. Sounds like the breaking point might be voice recognition - which I really thought would be much further along by now with all the other advancements. It seems figuring out what unique voices are saying is one of the hardest things to do well, and then add in background noises. Probably why the drive-thru AI attempts were terrible, working that job as a human can be difficult.
This is a cry for help.
some1 intervene b4 it too late.
