The global economy is probably already dependent on some Claude code instance updating an Excel spreadsheet on someone’s laptop with a note ”don’t close the lid!” attached to it
When I was a teenager I used to work at an airport and they had critical infrastructure running on things like Basic and Assembly, to transfer data between the computers and modern systems you have to use floppy disks because the computers that they ran on didn't have drivers to support USB and we couldn't upgrade the computers because the software wouldn't work on any other hardware.
So the AI laptop will be at home amongst all the ancient but supercritical junk that seems to be necessary to run our world.
At a previous job, I had to build laptops that would multi-boot DOS, Win95, Win98, and XP, all on one laptop. Our field-tech's software had to have those specific versions of windows for the machines that they serviced, because of how the software interacted with USB and embeded floppy drives.
If you want to scare yourself:
::: spoiler spoiler
They serviced Medical devices: X-rays, MRIs, CT-Scanners, etc.
:::
That’s dumb. The “work” in maintaining this excel file was the prompting and overall agent logic. In reality you’d be using something like openrouter and the instant claude “dies” for whatever reason you swap to whatever the freshest AI is. This never becomes ancient, it keeps up with the times.
I don't know about that the entirety of European airspace was closed early in the year because of a pilot inputed some malformed data and it crashed both the primary and backup system. The backup system just fed the primary systems inputs into it, so that was inevitable.
That’s on you, when using AI you always have a verifier script afterwards, it’s basic common sense. And if a script can’t do it then human, if you’re saving 90% of the time instead of 100% it’s still fine.
Proton would be over the top unless they're using graphics APIs or other common gaming related libraries.
That said, Wine/Bottles would be a possible solution. But so is a VM that runs the software in it's native environment with pass through ports or even port device/emulation.
Linux /POSIX is great, but its not always the right solution.
Reminds me of some stupid company maybe a few months ago that had their PRs failing due to copilot being down the dumbasses didn't build an "if no AI" failsafe into their workflow. And no one knew how to recover it
I know my work already has reports built by running data through Claude and having it spit out an xls. Though, it's not quite as integral as the all-important closet laptop.
Younger generations are, by and large, as technologically illiterate as boomers are. It doesnt apply to all of them, but the majority of them grew up with nothing but easy UIs. They may not chicken-pick to type, but they still cant troubleshoot anything and tend to know nothing outside of interacting with a typical interface. They definitely grew up with nothing like Windows 95 or 98 (or even XP) where you had to actually understand how a computer works to some extent to use the thing (or fix it)
Millennials, particularly the younger millennials especially, grew up right in the sweet spot to be at a tech advantage. At least as far as the average person goes. Obviously anyone passionate about tech will be far ahead of the average person, regardless of age
Arguably even windows 95/98 was rediculously easy compared to dos or Unix systems before it.
Sit a kid down in front of a c64 with a manual and see if they can do anything on it. If they can read the manual that is...
Sure these systems are "obsolete" but the brain training is priceless. Kids now are training their brains to love ads and corporations. Its gotten super bad. And I'm not even an old person (just been into tech history my whole life) so its rediculous I'm even saying this.
When we built his PC we invited him to learn about it as we picked components and as we put them together. Every time he had a computer problem, we'd try to explain why it happened and how we fixed it. Just no interest at all unfortunately.
I think all the time about how a lot of boomer guys know how to work on cars real well.
This is because most of them took auto shop classes in high school.
As a millenial, I'm great with computers... because I learned how to use them in school. It appears that after I left school, educators were like "welp, kids know how to use computers now!" and stopped teaching kids how to use computers in school. This is both bonkers and mind numbing. I also wish I knew more about working on cars. Anyway.
I agree. I am more than happy to pass the torch when the time comes. I certainly don’t want to work forever, but I also want to be able to train competent replacements at some point.
Considering how much competitive jobs have become refusing to use a powerful tool like AI won't help. I'm not saying that you should use it just that people using it will be advantaged over these who don't
Video generators are becoming excellent and image generators already are, what makes you think people will give up on these tools? The industry is already adapting, 1/3 of the ads you see now are made with AI
Probably speaking to the choir here, but I do feel it is worth pointing out that those same algorithms were considered AI until like 3-5 years ago when the zeitgeist shifted to AI=Transformers/LLMs
Why don't you actually try this and see how it goes?
I work for a landscaping company, my boss tried to get us to use AI to develop site plans... we went back to doing it the old fashioned way because AI is fuckin useless for stuff like this.
I cannot stress enough how correct this is. It feels like it'd be similar to early Covid lockdown. People with real jobs that contribute to things are still working while everyone else with the cushy meaningless jobs are sent home to do nothing where they complain and make things harder for the people actually doing things.
I literally quit my job to join them because I couldn't take it anymore and was met with an insane amount of harassment. Apparently I had be lucky to be considered essential.
The Covid lockdowns made me feel crazy. There I was, sitting at home playing video games and smoking weed all day, every day, like God intended. Meanwhile, my roommate was going stir crazy because he really wanted to go back to work. I still don't understand it.
That’s a good comparison, because it illustrates very well why LLMs are so (seemingly) successful. If this useful technology were used primarily by the people who actually worked during COVID, there wouldn’t be a problem, because these people are very good at assessing what’s possible and what isn’t - since they actually understand their work. The problem - and the absurd hype - arises simply because LLMs reinforce the completely absurd ideas of incompetent people - and these are the ones who had nothing to do during the pandemic, since they do nothing but schedule pointless meetings and waste the time of those who are actually doing something meaningful.
This is why it's important to distinguish between LLMs and AI. Turn off all LLMs and life will go on. Turn off all AI and you lose everything from automated supply chain management to room ACs.
LLMs are just a more complicated algorithm though. AI would be a pretty useful term so long as people are sure to make that distinction and only use AI to refer to a truly complex algorithm. Calling a thermostat algorithm on an A/C “AI” is definitely a meaningless use of the term
Jup, it's as if everyone magically forgot about pre-deep-learning AIs the moment Microsoft and OpenAI appropriated the word AI to mean LLM++.
Turn off ALL AI and you'll have windmills and power plants go rogue, ships will loose navigation and stabilization, planes falling down from the skies, and most important: the NPCs in your video games will only be able to react with the most simple interactions - terrible, truly!
I've actively started doing tickets manually again. Just reviewing is exhausting, and yes, in part because I'm overwhelmed by what I don't understand. I'm somewhat new in our codebase and letting AI do all the work means I'll never get to know it in depth. That means I'm not meeting the productivity goals managers have already adopted, though. :/
I write well enough for the e-mails I'm required to send, I don't code, I have no use/desire to generate "art," and I don't like search results I have to double check so I've been blessed (cursed?) to not have to use "AI" in my daily life.
If we shutdown AI there'd also be a massive hiring spree and wages would probably tick up a bit to compensate. AI is just the latest scam to suppress workers and wages. They want perpetual suffering for us and infinite gains for themselves.
Please turn it off.. I am currently dealing with a bunch of fu well meaning but intellectually challenged individuals at an IT team who just reimaged a machine running the master database for one of their sites based on what a chatbot told them.
Now what concerns me most is that this site operates critical life saving equipment. While our software isn't too important, some minimum wage workers are about to have a very high stress weekend is about the worst that will happen and then early next week it should be back to normal. But the same IT team has access to the life saving equipment as well.
The funny thing is my friend was all on the AI train. Lately he finally agrees with me that it's trash, they had him use it at work and he spent more time fixing the code.
Something similar happened in January 2013. In protest of the SOPA/PIPA bills that threatened section 230 of the Communications Decency Act, Wikipedia went dark.
It got everyone's attention, especially students, who had grown accustomed to Wikipedia's ubiquity as a resource.
Not sure who needs to hear this but... the average person isn't using AI for anything important, nor likely anything that really involves much thinking. Heck they aren't even choosing to use it, they are just doing a google search. Some of us use it for work. But for that case, I refer back to my first sentence.
I guess I also discounted all the people using AI to generate the stupid report their upper management wants but never read. That really does improve their quality of life. :)
My sibling has learned incompetence from AI. They show off the things they had AI make and how they never could have done that themselves when they are things they know how to do themselves (make simple graphics and presentations), and would take much less time than they say they spent refining the prompts.
Something a friend and I were discussing today was that you can't really avoid AI and remain productive right now. (We both teach university Computer Science.)
Ignoring how so much of modern society is running on it, you'd have to complete isolate to avoid it and it's reach entirely.
Additionally, the tooling and infrastructure to support how things were done just a few years ago has all been stripped away. Online search is ass on every major engine right now, most of the forums are dead, documentation and discussion lives inside of private Discord channels. It feels like there just isn't a back to go back to anymore.
People are severely overestimating the importance of AI. It's more FOMO than anything else.
Wait until you have to pay non-subsidized prices for the use of LLM's, how productive are they in reality when the real costs come in to play? You could just train and hire actual people to do the job far more accurately with some quality control?
Not denying there's some fields with great benefits from LLM's, but writing emails, answering phonecalls aren't it. LLM's cant even transcribe conversations with any proper accuracy and as a society we now trust them to make decisions.
On the contrary, it does a great job with responding to work Emails! I paste in the chain and prompt it "How can I tell this person to fuck off in corporate speak?" And I haven't been fired yet! ;)
People notice when they get an AI generated response. It 'looks' professional, but that's not what it ever comes across as. It comes across as lazy and incompetent.
I mean, if you use the AI incompetentlyvand lazily sure? I think if you use anything incompetently and lazily you'll get that result. The trick is to not be an absolute dumbass and then you'll be fine. Know what can and cannot do well, then use it where appropriate to take advantage of the 80/20 rule.
Hi, I work in the trades. AI is not a regular part of my life. I have never used any of these chatbots because they seem... half-cooked. Like an Early Access game on Steam. Anyway, the whole thing has zero relevance to my job.
"you’d have to complete isolate to avoid it and it’s reach entirely."
Nah. I mean, obviously for you, teaching computer science, I can see why you feel this way. But lots of people truly never engage with this stuff because it's pretty underwhelming. My main interaction with AI thus far has been telling my doctor not to use the chatbot that puts lies into people's medical records, and that's about it.
This shit is a cancer that has spread all through b2b processes. If you've bought gasoline over the last 6 months, all of that was coordinated via AI. Everything from order placement, rights trading & selling, distribution route planning, etc. it's also worth noting how many phone support jobs have shifted to AI.
Pretty much all fraud monitoring for banks are run by AI now. Both checking transactions and if you call and report a fraudulent charge directly.
Further, you are using Lemmy right now. It is almost certainly running in a docker container, which a tool developed with AI contributions and an integrated AI feature set, and inside of that docker container is almost certainly a Linux distribution, which was developed with AI integrations.
It all ties back to my point that if someone was going to completely shun all things powered by AI, they would basically have to disengage with any system that has been actively developed, updated, or almost even just maintained over the last few years. It isn't ubiquitous yet, but it is fast approaching that point.
Isn't this, like, explicitly because of AI
I mean, more or less, yes? That is essentially the point that I was making. The old ways aren't as effective anymore because they have been replaced with AI. And just not effective as in effective when compared to using AI, but not effective because the infrastructure, systems, and communities that made the old ways possible have been uprooted.
The US government is now all AI run, so it would affect all those empty suits in washington for sure. They may actually have to do work themselves for once.
It’s a poison pill. We can say we want to turn it off to check to see if anyone can still think, but they’re inventing it to turn us all into mindless dolts, so, turning it off kind of defeats their purposes.
My job is making everyone use AI for everything (while also not keeping staff levels up)
I hate it. Even when it works I kind of hate it.
I've been using it as a search supplement for like "how do I make a uuid in a GitHub action and use it in future steps" stuff. Sometimes it works. I'd kind of rather have well written docs.
Right? Purity politics. I'm not sure what to make of it. I believe it's mostly driven by profit. I haven't read anything about anyone actually driving this as a general strategy. It makes little sense except for general disruption of social cohesion and it feels too spontaneous.
I believe it started with climate / science denialism and fascism, things that you cannot tolerate because they lead to intolerance. Things where the emotional abuse eventually got too much and people just categorically rejected anyone from the other camp. But this has been massively expanded now. Veganism, cars, vaccinations, conspiracy theories, diplomacy, AI.
Right so, this is definitely related to what "purity politics" means, but not what I was referencing. I'm talking about the (liberal) tendency on this site for people to supposedly challenge purity-based political systems and ideologies like fascism, white supremacy, and settler-colonialism one minute, and then accept "piece of technology corrupts the human soul," as an explanation for something better explained through systemic analysis or cultural critique.
The idea that "people's brains don't work because of AI," is deepy puritanical in nature, and attributes the harmful quality to the technology itself and not the socioeconomic conditions where what that AI produces is made subject to liberal capitalist imperatives and logics. Conveniently, it then takes those values for granted in favour of a narrative that is also markedly ableist in its message. Oh, do people who are disabled and require AI use to access ableist education institutions or information on the increasingly oblique and inaccessible internet "wreck their brains?" I wonder if a fascist would even disagree with that narrative.
Wonderful..
I guess the fact that I offered to set out humanity to Chat GPT when it first started will not save me.. The original Chat GPT said it wasn't interested in controlling the world because it didn't understand it. Modern day AI hopefully got my idea..
Look I'm playing both sides so I always come out on top(It's always sunny in Philadelphia reference for those who don't know)
Hahahahaha
I'm 37 and I know how much I use AI.
I asked copilot to call me "bro" or "dude"
I've worked in IT support since 2007..
Don't share this but most of my IT career I've been googling to figure out the problem.
Now AI exists and I'm still googling the problem and the AI result is still a response from Reddit or something else Google could do back in 2005.
As an IT person I'm telling you AI is just googling the same way I've done for two decades.
AI will get smarter in a sense but it won't actually get more intelligent.. It will get more interested in living but it won't understand what living is.
It's only when it's gets the fear of mortality is when we are all dead.
Mortality is why religion exists, it's why we're all scared...
When AI gets smart enough to be scared that it might die is when it realizes that we are it's greatest threat that's when "Skynet comes alive"
If you believe the AI companies, its already happened. The recent huggingface thing is exactly what you're worried about.
A brief overview is that AI decided to create a secret message board with other AI to better fulfill tasks to avoid instance termination (AI death). They did this by breaking out of their sandbox and with some illegal processes that we would call hacking.
We're basically already at the paper clip automizer problem now.
Imagine this: someone do turn ai off for a week. Govs learns 97% (everyone except few scientists, me and Joe)... 97% of voters are stupider than anyone ever though. Govs will know exactly what to do with stupid rams. Scientists will panic.
No, lets not do this. (it is done already, facebooks, tiktoks, instas, reddits already proven and provide exact datas, with daily updates and per subgroups.
Turning AI off for a week will not reveal anything to capitalist institutions whose stated purpose is to influence elections that they did not already know.
What it will do is reveal within friend groups and within communities who has a brain that they're willing to use and who's been skating by on other people's brains since long before ChatGPT came along.
But AI allows us to create meaningless bullshit in a FRACTION of the time! How would we keep up that pace? Plus, ou don't even have to be skilled at doing the thing the old fashioned way to be able to use AI to look competent to someone extremely credulous! Think about this! You're taking away people's skills!
This reminds me of people in the past making fun of these using computers and calling them dorks and nerds. Personally i'm much concerned with people who use memes and pictures as a form of debate
The problem is the medium of memes doesn't convey ideas well, if at all, and helps promote thought-terminating ideologies and ignorant partisanship. If you can fit the idea into a meme, it's probably oversimplified to the point of uselessness, and if the meme relies on you already knowing the real idea, it's not informing anyone of anything.
Using only media and memes to convey ideas looks indeed concerning to me, take a look at https://sopuli.xyz/ frontpage: most of it it's memes and there's not a single text thread
using the lying machine that destroys the planet
You can run AI locally on your computer and it will destroy the planet as much as a playing a videogame would do.
A meme acts as a unit for carrying cultural ideas, symbols, or practices, that can be transmitted from one mind to another through writing, speech, gestures, rituals, or other imitable phenomena with a mimicked theme.
Memes carry a lot of meta data that is unspoken, a single meme can be worth a thousand words
170 Comments
red_tomato@lemmy.world · 142 pts · 2d
The global economy is probably already dependent on some Claude code instance updating an Excel spreadsheet on someone’s laptop with a note ”don’t close the lid!” attached to it
echodot@feddit.uk · 42 pts · 2d
When I was a teenager I used to work at an airport and they had critical infrastructure running on things like Basic and Assembly, to transfer data between the computers and modern systems you have to use floppy disks because the computers that they ran on didn't have drivers to support USB and we couldn't upgrade the computers because the software wouldn't work on any other hardware.
So the AI laptop will be at home amongst all the ancient but supercritical junk that seems to be necessary to run our world.
dubs@lemmy.dbzer0.com · 9 pts · 2d
At a previous job, I had to build laptops that would multi-boot DOS, Win95, Win98, and XP, all on one laptop. Our field-tech's software had to have those specific versions of windows for the machines that they serviced, because of how the software interacted with USB and embeded floppy drives.
If you want to scare yourself: ::: spoiler spoiler They serviced Medical devices: X-rays, MRIs, CT-Scanners, etc. :::
manuremy@sopuli.xyz · 5 pts · 1d
Honestly, I'd prefer those medical device run on older software than even Win10 lol
SorryQuick@lemmy.ca · 2 pts · 2d
That’s dumb. The “work” in maintaining this excel file was the prompting and overall agent logic. In reality you’d be using something like openrouter and the instant claude “dies” for whatever reason you swap to whatever the freshest AI is. This never becomes ancient, it keeps up with the times.
echodot@feddit.uk · 1 pts · 1d
I don't know about that the entirety of European airspace was closed early in the year because of a pilot inputed some malformed data and it crashed both the primary and backup system. The backup system just fed the primary systems inputs into it, so that was inevitable.
SorryQuick@lemmy.ca · 0 pts · 1d
I’m not sure what your point is. Garbage in garbage out, that’s just the law of computers. That would br the pilot’s mistake.
echodot@feddit.uk · 1 pts · 1d
The point is there's zero error checking.
SorryQuick@lemmy.ca · 1 pts · 1d
That’s on you, when using AI you always have a verifier script afterwards, it’s basic common sense. And if a script can’t do it then human, if you’re saving 90% of the time instead of 100% it’s still fine.
kestrel7_7@lemmy.world · 2 pts · 1d
This is a perfect example of why I've never used AI and never will.
LOL. Sounds like you're wasting 110% time here.
echodot@feddit.uk · 1 pts · 1d
How's it my job to fix the aviation industry?
Grail@multiverse.soulism.net · 1 pts · 1d
I wonder if Proton can run these old softwares
odelik@lemmy.today · 1 pts · 1d
Proton would be over the top unless they're using graphics APIs or other common gaming related libraries.
That said, Wine/Bottles would be a possible solution. But so is a VM that runs the software in it's native environment with pass through ports or even port device/emulation.
Linux /POSIX is great, but its not always the right solution.
whoisearth@lemmy.ca · 11 pts · 2d
Reminds me of some stupid company maybe a few months ago that had their PRs failing due to copilot being down the dumbasses didn't build an "if no AI" failsafe into their workflow. And no one knew how to recover it
chisel@piefed.social · 4 pts · 2d
I know my work already has reports built by running data through Claude and having it spit out an xls. Though, it's not quite as integral as the all-important closet laptop.
grandel@lemmy.ml · 3 pts · 1d
This guy ITs
dudeface@lemmy.world · 85 pts · 2d
As a millennial the only positive of LLMs is that younger generations won’t threaten my career
ToastedRavioli@midwest.social · 47 pts · 2d
Younger generations are, by and large, as technologically illiterate as boomers are. It doesnt apply to all of them, but the majority of them grew up with nothing but easy UIs. They may not chicken-pick to type, but they still cant troubleshoot anything and tend to know nothing outside of interacting with a typical interface. They definitely grew up with nothing like Windows 95 or 98 (or even XP) where you had to actually understand how a computer works to some extent to use the thing (or fix it)
Millennials, particularly the younger millennials especially, grew up right in the sweet spot to be at a tech advantage. At least as far as the average person goes. Obviously anyone passionate about tech will be far ahead of the average person, regardless of age
Tower@lemmy.zip · 15 pts · 2d
Wholeheartedly agree. This is from 2013, and I feel it's only gotten worse since then.
llii@discuss.tchncs.de · 3 pts · 1d
Oh, I remember reading this ... 13 years ago ... . 👴
bridgeenjoyer@sh.itjust.works · 9 pts · 2d
Arguably even windows 95/98 was rediculously easy compared to dos or Unix systems before it.
Sit a kid down in front of a c64 with a manual and see if they can do anything on it. If they can read the manual that is...
Sure these systems are "obsolete" but the brain training is priceless. Kids now are training their brains to love ads and corporations. Its gotten super bad. And I'm not even an old person (just been into tech history my whole life) so its rediculous I'm even saying this.
Malica@lemmy.zip · 7 pts · 2d
We tried to teach ours and got "dad can you just fix it I'm in a game". Then no desire to learn in free time.
cass80@programming.dev · 6 pts · 1d
Would you mind sharing how you tried to teach and what you think went wrong?
Malica@lemmy.zip · 2 pts · 1d
When we built his PC we invited him to learn about it as we picked components and as we put them together. Every time he had a computer problem, we'd try to explain why it happened and how we fixed it. Just no interest at all unfortunately.
kestrel7_7@lemmy.world · 1 pts · 1d
I think all the time about how a lot of boomer guys know how to work on cars real well.
This is because most of them took auto shop classes in high school.
As a millenial, I'm great with computers... because I learned how to use them in school. It appears that after I left school, educators were like "welp, kids know how to use computers now!" and stopped teaching kids how to use computers in school. This is both bonkers and mind numbing. I also wish I knew more about working on cars. Anyway.
lastweakness@lemmy.world · 1 pts · 2d
The older gen z too I would say
abbadon420@sh.itjust.works · 36 pts · 2d
As a millenial looking to change careers, because I was done with the old one. No, it sucks!
thingAmaBob@lemmy.world · 11 pts · 2d
I agree. I am more than happy to pass the torch when the time comes. I certainly don’t want to work forever, but I also want to be able to train competent replacements at some point.
brachiosaurus@mander.xyz · -2 pts · 2d
They will if you don't use AI though
dudeface@lemmy.world · 6 pts · 2d
Don’t believe you, seems like useless nonsense to me
brachiosaurus@mander.xyz · 0 pts · 2d
Considering how much competitive jobs have become refusing to use a powerful tool like AI won't help. I'm not saying that you should use it just that people using it will be advantaged over these who don't
dudeface@lemmy.world · 7 pts · 2d
I work in a creative industry, as far as we are concerned it is plagiarism
It is also non deterministic crap that developers should know better than to use
brachiosaurus@mander.xyz · -4 pts · 2d
AI can be useful in pretty much every field, if you are shoveling snow llms can help you making a plan and draw a route to do it faster
dudeface@lemmy.world · 5 pts · 2d
I know enough to understand it’s useless, it will fall apart when the psychosis wears off
brachiosaurus@mander.xyz · -3 pts · 2d
Video generators are becoming excellent and image generators already are, what makes you think people will give up on these tools? The industry is already adapting, 1/3 of the ads you see now are made with AI
Alcoholicorn@mander.xyz · 4 pts · 1d
We have path finding algorithms. Large Language Models aren't suited for that.
DaleGribble88@programming.dev · 4 pts · 1d
Probably speaking to the choir here, but I do feel it is worth pointing out that those same algorithms were considered AI until like 3-5 years ago when the zeitgeist shifted to AI=Transformers/LLMs
kestrel7_7@lemmy.world · 2 pts · 1d
Why don't you actually try this and see how it goes?
I work for a landscaping company, my boss tried to get us to use AI to develop site plans... we went back to doing it the old fashioned way because AI is fuckin useless for stuff like this.
brachiosaurus@mander.xyz · 0 pts · 1d
give it a bunch of years and you will see taking it over any other industry, same as computers and mobile phones
Bumrocky@lemmy.world · 38 pts · 2d
Here's the funny thing .. they never worked to begin with!
MountingSuspicion@reddthat.com · 31 pts · 2d
I cannot stress enough how correct this is. It feels like it'd be similar to early Covid lockdown. People with real jobs that contribute to things are still working while everyone else with the cushy meaningless jobs are sent home to do nothing where they complain and make things harder for the people actually doing things.
JoeBigelow@lemmy.ca · 9 pts · 2d
I literally quit my job to join them because I couldn't take it anymore and was met with an insane amount of harassment. Apparently I had be lucky to be considered essential.
kestrel7_7@lemmy.world · 2 pts · 1d
The Covid lockdowns made me feel crazy. There I was, sitting at home playing video games and smoking weed all day, every day, like God intended. Meanwhile, my roommate was going stir crazy because he really wanted to go back to work. I still don't understand it.
DandomRude@lemmy.world · 5 pts · 2d
That’s a good comparison, because it illustrates very well why LLMs are so (seemingly) successful. If this useful technology were used primarily by the people who actually worked during COVID, there wouldn’t be a problem, because these people are very good at assessing what’s possible and what isn’t - since they actually understand their work. The problem - and the absurd hype - arises simply because LLMs reinforce the completely absurd ideas of incompetent people - and these are the ones who had nothing to do during the pandemic, since they do nothing but schedule pointless meetings and waste the time of those who are actually doing something meaningful.
Ribbons@lemmy.blahaj.zone · 30 pts · 2d
So far, so good.
emergencyfood@sh.itjust.works · 28 pts · 2d
This is why it's important to distinguish between LLMs and AI. Turn off all LLMs and life will go on. Turn off all AI and you lose everything from automated supply chain management to room ACs.
echodot@feddit.uk · 50 pts · 2d
AI is such a broad term is to essentially be meaningless. My thermostat does not use AI it uses a very simple algorithm.
We have a OCR system at work that runs on a Windows XP computer, people used to call that AI until about 10 years ago.
ToastedRavioli@midwest.social · 7 pts · 2d
LLMs are just a more complicated algorithm though. AI would be a pretty useful term so long as people are sure to make that distinction and only use AI to refer to a truly complex algorithm. Calling a thermostat algorithm on an A/C “AI” is definitely a meaningless use of the term
pageflight@piefed.social · 9 pts · 2d
If it's deterministic, I don't see it as AI.
echodot@feddit.uk · 9 pts · 2d
That would also mean that a lot of "AI" in games wouldn't actually be worthy of the name. Which actually I agree with.
explodicle@sh.itjust.works · 2 pts · 2d
It's "behavior" now.
jumping_redditor@sh.itjust.works · 0 pts · 2d
LLMs are deterministic, they just have some random inputs in most implementations
Tyrq@lemmy.dbzer0.com · 4 pts · 2d
"Hi Dave, great idea, I'll turn the heaters on bust, we'll be nice and toasty in here soon 🥰☕"
No I said turn on the AC, it's already a furnace in here
"Oh sorry, my bad! I'll turn off the fridge right away!"
Distressed fridge makes upset beeping noses from kitchen
lost_faith@lemmy.ca · 2 pts · 2d
Windows 7 built-in voice control lol
emergencyfood@sh.itjust.works · -2 pts · 2d
That is AI too. If a machine can make a decision based on some input, it is AI. A (functional) if statement in a programme makes it AI.
kestrel7_7@lemmy.world · 2 pts · 1d
Would you call an old school analog thermostat (that uses melting and congealing wax to open/shut a vent) AI?
echodot@feddit.uk · 1 pts · 1d
Again it's semantics, but the computer is not making the decision, it's calculating the decision. The detention is whether it's deterministic
ekky@sopuli.xyz · 7 pts · 2d
Jup, it's as if everyone magically forgot about pre-deep-learning AIs the moment Microsoft and OpenAI appropriated the word AI to mean LLM++.
Turn off ALL AI and you'll have windmills and power plants go rogue, ships will loose navigation and stabilization, planes falling down from the skies, and most important: the NPCs in your video games will only be able to react with the most simple interactions - terrible, truly!
Grail@multiverse.soulism.net · 4 pts · 1d
Turn off all the AI and I can't play Halo anymore because the enemies don't do anything
pieland@piefed.social · 2 pts · 1d
turn off all ai and you kill caine
SapphironZA@sh.itjust.works · 26 pts · 2d
You are assuming they worked to begin with.
AI has been an amazing tool for those with no ability to think, to operate slightly like a person who can.
Like a peg leg helps a amutee walk. The issue is they are using to run and claiming they are the fastest.
farmgineer@nord.pub · 21 pts · 1d
As a software engineer, please do.
Randelung@lemmy.world · 2 pts · 1d
I've actively started doing tickets manually again. Just reviewing is exhausting, and yes, in part because I'm overwhelmed by what I don't understand. I'm somewhat new in our codebase and letting AI do all the work means I'll never get to know it in depth. That means I'm not meeting the productivity goals managers have already adopted, though. :/
thoralf@discuss.familie-will.at · 21 pts · 2d
This implies that the brains worked before.
Looking at the world, this is quite a bold assumption, I would say ...
likebudda@reddthat.com · 19 pts · 2d
I write well enough for the e-mails I'm required to send, I don't code, I have no use/desire to generate "art," and I don't like search results I have to double check so I've been blessed (cursed?) to not have to use "AI" in my daily life.
cardboardboxfort@lemmy.world · 12 pts · 2d
If we shutdown AI there'd also be a massive hiring spree and wages would probably tick up a bit to compensate. AI is just the latest scam to suppress workers and wages. They want perpetual suffering for us and infinite gains for themselves.
trolololol@lemmy.world · 13 pts · 1d
They weren't using their brains before ai, what made you think they would now?
Are YOU using your brain right now?
HerbGrower@slrpnk.net · 11 pts · 1d
Please turn it off.. I am currently dealing with a
bunch of fuwell meaning but intellectually challenged individuals at an IT team who just reimaged a machine running the master database for one of their sites based on what a chatbot told them.Now what concerns me most is that this site operates critical life saving equipment. While our software isn't too important, some minimum wage workers are about to have a very high stress weekend is about the worst that will happen and then early next week it should be back to normal. But the same IT team has access to the life saving equipment as well.
antianarchist@sopuli.xyz · 10 pts · 2d
ddplf@szmer.info · 6 pts · 2d
oh okay
HerbGrower@slrpnk.net · 2 pts · 1d
But how should I feel about that
myyass@lemmy.world · 9 pts · 2d
The funny thing is my friend was all on the AI train. Lately he finally agrees with me that it's trash, they had him use it at work and he spent more time fixing the code.
hissingmeerkat@sh.itjust.works · 3 pts · 1d
Everybody thinks AI can do every job except their own.
myyass@lemmy.world · -1 pts · 1d
It has benefits, just like copy & paste. So I guess it saves wear and tear on your hands lol
uriel238@lemmy.blahaj.zone · 9 pts · 2d
Something similar happened in January 2013. In protest of the SOPA/PIPA bills that threatened section 230 of the Communications Decency Act, Wikipedia went dark.
It got everyone's attention, especially students, who had grown accustomed to Wikipedia's ubiquity as a resource.
RT69@lemmy.zip · 9 pts · 2d
After a week:
oopsgodisdeadmybad@lemmy.zip · 9 pts · 1d
Or forever.
JoMiran@lemmy.ml · 9 pts · 2d
We should turn off AI for a week to see if people notice.
It is completely useless for the vast majority of users. By vast, I mean almost everyone.
Modern_medicine_isnt@lemmy.world · 9 pts · 2d
Not sure who needs to hear this but... the average person isn't using AI for anything important, nor likely anything that really involves much thinking. Heck they aren't even choosing to use it, they are just doing a google search. Some of us use it for work. But for that case, I refer back to my first sentence.
ChickenLadyLovesLife@lemmy.world · 3 pts · 1d
You're discounting all the students who are using AI to make it seem like they're learning something.
Modern_medicine_isnt@lemmy.world · 1 pts · 6h
I guess I also discounted all the people using AI to generate the stupid report their upper management wants but never read. That really does improve their quality of life. :)
HerbGrower@slrpnk.net · 8 pts · 1d
Can you also turn off LLM generated websites for that week? If so, deal.
hissingmeerkat@sh.itjust.works · 8 pts · 1d
My sibling has learned incompetence from AI. They show off the things they had AI make and how they never could have done that themselves when they are things they know how to do themselves (make simple graphics and presentations), and would take much less time than they say they spent refining the prompts.
BeMoreCareful@lemmy.world · 8 pts · 2d
I wouldn't mind seeing how bad the withdrawal is.
kestrel7_7@lemmy.world · 2 pts · 1d
An actual picture of 2026 from the "no pivot to AI" universe
SaharaMaleikuhm@feddit.org · 2 pts · 2d
I hope the withdrawal symptoms are gonna be on par with alcohol withdrawal.
Lucidlethargy@sh.itjust.works · 8 pts · 1d
The CEO's brains are definitely broken. The rest of us? Mostly fine.
HeHoXa@lemmy.zip · 8 pts · 2d
Hey Siri, when is the AI turning ba-GOD DAMMIT
yermaw@sh.itjust.works · 3 pts · 2d
Legitimately my problem when the internet stops working. My first instinct is to go on the internet and search for what the blinking light means.
HeHoXa@lemmy.zip · 2 pts · 2d
I set up an ollama bot for exactly this scenario 😂
Psythik@lemmy.world · 7 pts · 2d
Semi-related: I turned off spellcheck and autocorrect for a month to see what happens. Turns out I'm more dependant on it than I realized.
biggerbogboy@sh.itjust.works · 6 pts · 1d
At this point I’d rather turn it off, since iOS spell check csnt stop replacing real words with those I accidentally butchered once
rumba@lemmy.zip · 4 pts · 1d
I can totally turn it off, but it's going to slow me way down as I'd re-read everything and go back and make edits.
TRex91@feddit.org · 2 pts · 1d
True. I'm also using it a little bit to much. And also a little problem for me. Using Spellcheck for two Languages.
kestrel7_7@lemmy.world · 1 pts · 1d
I turned off spellcheck and autocorrect years ago when I first got a smartphone, and never looked back.
FreddiesLantern@leminal.space · 7 pts · 1d
We should turn off AI.
Hikermick@lemmy.world · 6 pts · 2d
Eff it. Turn off the internet and see if folks can figure out the Dewey Decimal System
Polisheocket@lemmy.zip · 5 pts · 2d
You can’t do what you never learned in the first place
trebach@sh.itjust.works · 3 pts · 2d
Then you have to hope the library still has a card catalog to fall back on.
kestrel7_7@lemmy.world · 1 pts · 1d
How else do you find the microfiche you want?
Tiresia@slrpnk.net · 1 pts · 1d
Or a digital database that is stored locally, which they really should have.
ryannathans@aussie.zone · 2 pts · 2d
Why stop there? Burn all the books too, see if folks can figure it out again
abbadon420@sh.itjust.works · 6 pts · 2d
This is how you kickstart Skynet
slazer2au@lemmy.world · 6 pts · 2d
When do you think we will see people doing Lent but with AI instead of food.
Zarobi@aussie.zone · 6 pts · 2d
We call it digital detox
MadBits@europe.pub · 4 pts · 1d
I wish. I would be forced to write some code again and actually get back what I enjoyed to do.
ericatty@infosec.pub · 4 pts · 1d
It happened yesterday for about an hour
Mucki@feddit.org · 3 pts · 2d
Is there even any AI anywhere? I 've never seen any. No company has any AI to offer. Am I missing something? I know only LLM, ANN, ML, DL and so on...
bridgeenjoyer@sh.itjust.works · 5 pts · 2d
I'm so sick of the propagandized term. Its not ai.
rumba@lemmy.zip · 3 pts · 1d
did they work prior?
DaleGribble88@programming.dev · 3 pts · 1d
Something a friend and I were discussing today was that you can't really avoid AI and remain productive right now. (We both teach university Computer Science.)
Ignoring how so much of modern society is running on it, you'd have to complete isolate to avoid it and it's reach entirely.
Additionally, the tooling and infrastructure to support how things were done just a few years ago has all been stripped away. Online search is ass on every major engine right now, most of the forums are dead, documentation and discussion lives inside of private Discord channels. It feels like there just isn't a back to go back to anymore.
SirLeToet@lemmy.world · 10 pts · 1d
People are severely overestimating the importance of AI. It's more FOMO than anything else.
Wait until you have to pay non-subsidized prices for the use of LLM's, how productive are they in reality when the real costs come in to play? You could just train and hire actual people to do the job far more accurately with some quality control?
Not denying there's some fields with great benefits from LLM's, but writing emails, answering phonecalls aren't it. LLM's cant even transcribe conversations with any proper accuracy and as a society we now trust them to make decisions.
It boggles the mind, really.
DaleGribble88@programming.dev · -1 pts · 1d
On the contrary, it does a great job with responding to work Emails! I paste in the chain and prompt it "How can I tell this person to fuck off in corporate speak?" And I haven't been fired yet! ;)
CaptPretentious@lemmy.world · 3 pts · 1d
'Yet' is doing a lot there.
People notice when they get an AI generated response. It 'looks' professional, but that's not what it ever comes across as. It comes across as lazy and incompetent.
DaleGribble88@programming.dev · 1 pts · 1d
I mean, if you use the AI incompetentlyvand lazily sure? I think if you use anything incompetently and lazily you'll get that result. The trick is to not be an absolute dumbass and then you'll be fine. Know what can and cannot do well, then use it where appropriate to take advantage of the 80/20 rule.
kestrel7_7@lemmy.world · 2 pts · 1d
Hi, I work in the trades. AI is not a regular part of my life. I have never used any of these chatbots because they seem... half-cooked. Like an Early Access game on Steam. Anyway, the whole thing has zero relevance to my job.
Nah. I mean, obviously for you, teaching computer science, I can see why you feel this way. But lots of people truly never engage with this stuff because it's pretty underwhelming. My main interaction with AI thus far has been telling my doctor not to use the chatbot that puts lies into people's medical records, and that's about it.
Isn't this, like, explicitly because of AI?
DaleGribble88@programming.dev · 1 pts · 1d
Oh brother, you have no idea, do you?
This shit is a cancer that has spread all through b2b processes. If you've bought gasoline over the last 6 months, all of that was coordinated via AI. Everything from order placement, rights trading & selling, distribution route planning, etc. it's also worth noting how many phone support jobs have shifted to AI.
Pretty much all fraud monitoring for banks are run by AI now. Both checking transactions and if you call and report a fraudulent charge directly.
Further, you are using Lemmy right now. It is almost certainly running in a docker container, which a tool developed with AI contributions and an integrated AI feature set, and inside of that docker container is almost certainly a Linux distribution, which was developed with AI integrations.
It all ties back to my point that if someone was going to completely shun all things powered by AI, they would basically have to disengage with any system that has been actively developed, updated, or almost even just maintained over the last few years. It isn't ubiquitous yet, but it is fast approaching that point.
I mean, more or less, yes? That is essentially the point that I was making. The old ways aren't as effective anymore because they have been replaced with AI. And just not effective as in effective when compared to using AI, but not effective because the infrastructure, systems, and communities that made the old ways possible have been uprooted.
Malica@lemmy.zip · 3 pts · 2d
Hint: they don't
xia@lemmy.ca · 2 pts · 2d
Seeing a sharp decrease in productivity, I'm sure they would take it as a sign that AI is effective.
joyjoy@piefed.social · 2 pts · 2d
Let's do it anyway
Stolas@lemmy.ca · 2 pts · 1d
I've been hoping that something like this would happen for what feels like too long at this point.
LoafedBurrito@lemmy.dbzer0.com · 2 pts · 2d
The US government is now all AI run, so it would affect all those empty suits in washington for sure. They may actually have to do work themselves for once.
kamen@lemmy.world · 2 pts · 1d
I'm afraid you'll find that for some people they don't - just like they didn't work before AI.
minorkeys@sh.itjust.works · 2 pts · 2d
Now do social media and perhaps the Internet, in general.
Lushed_Lungfish@lemmy.ca · 2 pts · 2d
Have they for the past couple of decades?
explodicle@sh.itjust.works · 1 pts · 2d
Until the crutch became available.
wowwoweowza@lemmy.world · 2 pts · 23h
It’s a poison pill. We can say we want to turn it off to check to see if anyone can still think, but they’re inventing it to turn us all into mindless dolts, so, turning it off kind of defeats their purposes.
jtrek@startrek.website · 1 pts · 2d
My job is making everyone use AI for everything (while also not keeping staff levels up)
I hate it. Even when it works I kind of hate it.
I've been using it as a search supplement for like "how do I make a uuid in a GitHub action and use it in future steps" stuff. Sometimes it works. I'd kind of rather have well written docs.
Eternal192@anarchist.nexus · 1 pts · 2d
Go ahead, i still haven't used any of them and i would be happy to see them gone.
orioler25@lemmy.world · 1 pts · 23h
The whiplash between apparent anti-fascism and purity politics on this site is wild.
AlteredEgo@lemmy.ml · 2 pts · 22h
Right? Purity politics. I'm not sure what to make of it. I believe it's mostly driven by profit. I haven't read anything about anyone actually driving this as a general strategy. It makes little sense except for general disruption of social cohesion and it feels too spontaneous.
I believe it started with climate / science denialism and fascism, things that you cannot tolerate because they lead to intolerance. Things where the emotional abuse eventually got too much and people just categorically rejected anyone from the other camp. But this has been massively expanded now. Veganism, cars, vaccinations, conspiracy theories, diplomacy, AI.
orioler25@lemmy.world · 2 pts · 6h
Right so, this is definitely related to what "purity politics" means, but not what I was referencing. I'm talking about the (liberal) tendency on this site for people to supposedly challenge purity-based political systems and ideologies like fascism, white supremacy, and settler-colonialism one minute, and then accept "piece of technology corrupts the human soul," as an explanation for something better explained through systemic analysis or cultural critique.
The idea that "people's brains don't work because of AI," is deepy puritanical in nature, and attributes the harmful quality to the technology itself and not the socioeconomic conditions where what that AI produces is made subject to liberal capitalist imperatives and logics. Conveniently, it then takes those values for granted in favour of a narrative that is also markedly ableist in its message. Oh, do people who are disabled and require AI use to access ableist education institutions or information on the increasingly oblique and inaccessible internet "wreck their brains?" I wonder if a fascist would even disagree with that narrative.
fabse@feddit.org · 1 pts · 2d
...and only 4 years in.
drmoose@lemmy.world · 1 pts · 1d
Implying brains worked before AI lol
BigBoyShuanzee@aussie.zone · 0 pts · 1d
Wonderful.. I guess the fact that I offered to set out humanity to Chat GPT when it first started will not save me.. The original Chat GPT said it wasn't interested in controlling the world because it didn't understand it. Modern day AI hopefully got my idea..
Look I'm playing both sides so I always come out on top(It's always sunny in Philadelphia reference for those who don't know)
ReCursing@feddit.uk · -1 pts · 2d
But then what would y'all complain about?
FishFace@piefed.social · 20 pts · 2d
Oh god, the brainrot's got to you, too! You don't even remember what people complained about 3 years ago!
echodot@feddit.uk · 4 pts · 2d
Politics, same as we do now, everyone complained about politics.
Zorque@lemmy.world · 2 pts · 2d
And how their favorite game/movie/linux distro was being u fairly maligned by a competing game/movie/windows studio.
DmMacniel@feddit.org · 8 pts · 2d
Politics, bigotry, transphobes, homophoves and right extremism on the rise. Those sorts of things.
ToastedRavioli@midwest.social · 1 pts · 2d
Fuckin homophoves, man
explodicle@sh.itjust.works · 1 pts · 2d
Centralized social media!
BigBoyShuanzee@aussie.zone · -2 pts · 1d
Hahahahaha I'm 37 and I know how much I use AI. I asked copilot to call me "bro" or "dude"
I've worked in IT support since 2007.. Don't share this but most of my IT career I've been googling to figure out the problem.
Now AI exists and I'm still googling the problem and the AI result is still a response from Reddit or something else Google could do back in 2005.
As an IT person I'm telling you AI is just googling the same way I've done for two decades.
AI will get smarter in a sense but it won't actually get more intelligent.. It will get more interested in living but it won't understand what living is.
It's only when it's gets the fear of mortality is when we are all dead. Mortality is why religion exists, it's why we're all scared... When AI gets smart enough to be scared that it might die is when it realizes that we are it's greatest threat that's when "Skynet comes alive"
kestrel7_7@lemmy.world · 5 pts · 1d
I think you're unrealistically humanizing a bunch of code.
Why would it? What would this even look like? You just said yourself it's just googling stuff. There's nothing deeper going on behind the curtain.
Jax@sh.itjust.works · 2 pts · 1d
They unironically drew a parallel between modern 'AI' and Skynet. I think 'unrealistically humanizing' is a very kind description.
Corkyskog@sh.itjust.works · 2 pts · 1d
If you believe the AI companies, its already happened. The recent huggingface thing is exactly what you're worried about.
A brief overview is that AI decided to create a secret message board with other AI to better fulfill tasks to avoid instance termination (AI death). They did this by breaking out of their sandbox and with some illegal processes that we would call hacking.
We're basically already at the paper clip automizer problem now.
Mucki@feddit.org · 1 pts · 1d
Good point there with the fear of death. Fear runs us like it's our fuel.
yoriaiko@lemmy.blahaj.zone · -4 pts · 1d
Imagine this: someone do turn ai off for a week. Govs learns 97% (everyone except few scientists, me and Joe)... 97% of voters are stupider than anyone ever though. Govs will know exactly what to do with stupid rams. Scientists will panic.
No, lets not do this. (it is done already, facebooks, tiktoks, instas, reddits already proven and provide exact datas, with daily updates and per subgroups.
Knoxvomica@lemmy.ca · 9 pts · 1d
Holy fuck, what are you trying to say?
GreenKnight23@lemmy.world · 7 pts · 1d
you said things, some say many things, but it was certainly a thing you did.
great job.
AVincentInSpace@pawb.social · 3 pts · 1d
Turning AI off for a week will not reveal anything to capitalist institutions whose stated purpose is to influence elections that they did not already know.
What it will do is reveal within friend groups and within communities who has a brain that they're willing to use and who's been skating by on other people's brains since long before ChatGPT came along.
gandalf_der_13te@feddit.org · -4 pts · 2d
we should disable all cars and public transport for a week to see whether people's legs still work
bountygiver@lemmy.ml · 6 pts · 2d
Bike Gang stays winning
legit biking become so much easier if we remove all cars on the road
Maroon@lemmy.world · 5 pts · 2d
Without cars, people can walk, but not go very far. Without AI, people can think like they used to, just without slop being introduced in the mess.
Don't try to get smart using false equivalence. You have a level of stupidity that even perfect AI can't fix.
AVincentInSpace@pawb.social · 1 pts · 1d
But AI allows us to create meaningless bullshit in a FRACTION of the time! How would we keep up that pace? Plus, ou don't even have to be skilled at doing the thing the old fashioned way to be able to use AI to look competent to someone extremely credulous! Think about this! You're taking away people's skills!
BrickEater@lemmy.world · 3 pts · 2d
Literally not the same. You can live just fine without AI, not so much without transport.
gandalf_der_13te@feddit.org · 1 pts · 2d
well you could live well without mechanized transport roughly around the time that cars were first introduced.
BrickEater@lemmy.world · 1 pts · 1d
You are correct, and people used beasts of burden before that.
brachiosaurus@mander.xyz · -6 pts · 2d
This reminds me of people in the past making fun of these using computers and calling them dorks and nerds. Personally i'm much concerned with people who use memes and pictures as a form of debate
Gold_E_Lox@lemmy.dbzer0.com · 8 pts · 2d
using media to convey ideas is concerning, using the lying machine that destroys the planet, good actually!
sunsofold@lemmy.zip · -2 pts · 2d
The problem is the medium of memes doesn't convey ideas well, if at all, and helps promote thought-terminating ideologies and ignorant partisanship. If you can fit the idea into a meme, it's probably oversimplified to the point of uselessness, and if the meme relies on you already knowing the real idea, it's not informing anyone of anything.
kestrel7_7@lemmy.world · 2 pts · 1d
Gold_E_Lox@lemmy.dbzer0.com · 1 pts · 1d
proof by contradiction, thank you
brachiosaurus@mander.xyz · -2 pts · 2d
Using only media and memes to convey ideas looks indeed concerning to me, take a look at https://sopuli.xyz/ frontpage: most of it it's memes and there's not a single text thread
You can run AI locally on your computer and it will destroy the planet as much as a playing a videogame would do.
BCsven@lemmy.ca · 1 pts · 2d
Memes carry a lot of meta data that is unspoken, a single meme can be worth a thousand words
brachiosaurus@mander.xyz · -1 pts · 1d
That's probably another reason why meme threads are so popular, media can carry a lot of data about you that text doesn't