I noticed almost everybody here agrees that all AI is bad, and anything produced by AI is slop. and refuse to use any software made in part or in whole with AI or help of AI? Please mention your background and usage of AI with your answer.
You must log in or register to comment.
I work in software. Just in day to day:
- lazy people are now using AI to slop out meaningless drivel at me rather than taking five minutes to articulate an actual thought
- shit is frequently broken because people surrender their thinking to the LLMs, don’t understand what the agents are doing, and don’t check the work output
Broadly:
- the industry is built on theft and whitewashing of human effort
- fraudulent valuations and circular dealing
- the frontier AI companies are all run by lying sociopaths
- and yet the global economy now hinges on a handful of AI or AI-adjacent companies
- many corporate types are now convinced they don’t need real human labor anymore, and treat the working class accordingly
- AI is an incredible force multiplier for scamming/spamming
- hyperscaler data center build out is mathematically impossible, environmentally ruinous, and harmful to your health
- LLMs are literally making us stupider
It’s unethical on so many fronts I’d assume you’re a psychopath if you don’t understand.
I’ve been a software engineer for over thirty years, with a bachelor’s degree in Software Engineering, and a Master’s in Machine Learning. I love AI, and can’t help but point out how silly all the hate is - even though it gets me nothing but downvotes. People say it’s unethical, but what IS ethical? Certainly not the rare earth electronics built in polluting factories we use AI on, nor the likely dirty electricity used to power them. LLM are wonderfully powerful, useful things and I think they should be embraced - at least until Capitalism itself falls, and our civilizations along with it.
Most lemmy users live in countries which are run for a rich few, rather than the general public. Technology, including AI, is also used to benefit the rich, often at the cost of the rest of the population. Unfortunately this leads to people hating the technology rather than the political system.
There was a recent survey where they pointed out how people in English-speaking countries are most concerned about AI (because it is used against the people there) while people in the Global South are most positive about it (because it is used to benefit them). Here’s the summary chart:-
yeah no, this is weird framing. i’m south korean myself and from what i feel it’s not much (if any) different from what i see around the western internet
Hmm yes South Korea isn’t a Global South country. It’s closer to Western countries in terms of politics and economic situation.
Btw is there an age-based or rural-urban division in perception of AI?
rural-urban division: don’t think so. even rural areas are pretty urbanized(?) and there isn’t that much of a cultural? difference overall
age-based: i think yes, not much different from other countries. older people not being able to tell ai image/videos for example.
I’m a graphic designer.
The entire point of AI as it exists in the western imperialist world is to cut out workers, labor rights, and the means of production by automating jobs, capturing processing power to replace brain power, and essentially to funnel money from consumers to those who control AI. Those companies are using our materials, our work, our capital, and our attention to train their models, and NONE of it is meant to make our material conditions better. What do we get from it now? Text that sounds like a middle manager, and that’s factually correct as often as one. Images that look like shit, and that are making people dumber. Cars that disable their AI at the last possible moment specifically to avoid responsibility, and that doesn’t work well-enough to be road legal under ideal circumstances. Search engine answers that show nothing but that AI has automated being confidently wrong on the internet. Entire towns that no longer have water or electricity, but do have constant jet noise, diesel pollution and markedly higher temperatures.
We aren’t even at a stage where anyone but Nvidia is actually getting anything out of it. Budgets aren’t being saved because companies are using AI, they’re just spending the money on AI instead of for actual work. It’s a cult. It’s operating as a cult, and for some reason, no one is treating it as if it’s a cult except “Rationalists”, who are balls deep in it because it’s a cult.
As I understand people in design see the worst part. I think that is because design has a huge artistic part. But when other people do it, they do not care about art, they just want to to pass some information even if it makes my eyes bleed and make my cognitive load spike thru the roof (I would be exhausted by the end of the day looking at all the slop). But truth is, those kind of people you mention wanted this situation even before AI and GenAI was a thing, they just did not have a way to do it. Now they have it. Most Big companies do not reason in the way we real humans do. They want to make money, so they build a system that makes them money, then they keep tweaking it to make even more money. They don’t care about the actual product or how it is made. They care about making money.
That is why we have the situation you describe.
Honestly, the only parts I’ve felt as if AI has impacted me is in hiring. Designers tend to sell themselves based on their portfolios, and most of the decently-paying companies seem to now be using AI to sort through résumés, so regardless of how good your work is, you need to use arbitrarily-selected keywords or you won’t even get an interview. The other way is that HR departments list being “proficient with leading AI platforms” with no idea how that might be important to the job, and from what I can tell, resistance to that is a red flag.
Because it got enshittified and politicized by Tech Bros.
If you wanna tell me about the usefulness of LLMs, you are preaching to the choir. I was playing with GTP-J 6B finetunes in 2022. I am quantizing some weights as I type.
However, you can’t look at me with a straight face and tell me everything is okay in this space. It’s a nightmare, a giant scam that is going to implode.
As Lemmy is a refuge for people who see how caustic mainstream social media is, folks here are particularly sensitive to the issues surrounding “AI” too.
In other words, Lemmy literally selects for enshittification sensitivity.
There are a lot of software developers on Lemmy.
AI is being pushed as a software development tool. It tends to remove the fun part of our jobs (coding, debugging, optimizing) and relegates us to the shitty parts (code review).
We don’t like that.
Personally I find I’m able to focus on high level design problems which I enjoy a lot. I still enjoy crafting elegant solutions to unique problems, but at least in my current job, most code is trivial and feels tedious to write by hand. Also, LLMs generate better code than some of my coworkers. Most of the slop I’ve had to deal with this year is in 20+ y/o codebases, and none of it was generated by LLMs.
I’m disabled and used to make book covers for indie authors. Now everybody uses AI for that and I can’t find another job.
Several legal, environmental and social reasons I can tell
I agree with you on the environmental and social reasons. The impact here is important. But what about the second part of the question, the usefulness and correct productive usage?
For many people, the cons outweigh the pros, and therefore they can’t even consider them as pros at all.
Yes, this is what I start to understand here from the comments. Most people commenting here signify they care more about the resource usage and impact it has on the world and us rather than its actual output.
What other way would you look at it after weighing the pros and cons ?
It is antisocial and nihilistic technology that makes my hobbies worse and less accessable
It’s because we are adults, so we already have better skills than a general LLM. We learned to read, write, and research in school. While I personally didn’t go to art school, many people did, and they developed skills that allow them to make better art than a general use art AI.
To us, AI is just wasting our time, because it does what we can already do but worse.
Children are in an unfortunate trap, where they need to spend years producing work that is worse than the LLM to eventually be able to develop the skillset to surpass the LLM. Buying into the idea that it makes your work faster or easier is a trap. You’ll be permanently dependent on it.
we already have better skills than a general LLM
Can you read a thousand pages of tech documentation and find every reference that could be relevant to the issue you’re having in minutes? Then no, you don’t always have ‘better skills than a general LLM’. Humans are better at some things. Computers are better at some things. Use the right tool for the job.
It’s funny you sent this just now. I just spent like 45 minutes trying to get Co-Counsel to make a list of all of the cases in a brief I’m working on. It got confused and couldn’t. I would’ve been done by now if I just used my brain. /shrug
Are you using AI to help you with legal stuff? I don’t think you should use AI for that since it would have data on rules from all sorts of places, and might not be up to date on laws in your place. Or you could train an AI on exclusively the data in your area, but then (1) you’d need hundreds of models, and (2) they would be much less powerful since the training data is a lot less.
CoCounsel is a professional legal service offered by WestLaw. WestLaw is one of the two major legal information companies. They were the ones who transitioned the law from books and brought it online in the 90s. Their dataset is exclusively their databases, which I would guess contain between 90-99% of all American Law ever written.
CoCounsel does exactly what you said in your earlier comment. It scans thousands of legal documents to find exactly what you’re looking for. If I want to find a case that talks about both the 4th amendment during a jurisdictional issue, it can instantly find a case with a a direct quote.
In this case, I asked it to do a multi-step process: find ~50 cases referenced in a specific document, alphabetize them, and add a two sentence summary below each one. That was too much for it to handle at once. It got confused.
In the US, each state can have its own laws, right? How does the AI handle this?
Also, how does it understand grammar? Wouldn’t it first need to be trained on a large corpus of text?
I agree with you generally. However can you clarify please when you say “does what we can already do but worse”. Can you please clarify ? What is your use of AI? Software development? Engineering? Architecture? Document editing? General usage and research? Paid and advanced models vs free or simple ones?
Because how you use it with give you a varying set of results.
Not OP and sorry you’re getting downvoted when you’re asking genuine questions and not trolling. Mileage varies with AI. Most people aren’t paying for AI services, meaning that they only work with really dumb models. The people creating slop videos, vibe coding usable and useful apps, using AI to do legal stuff are paying AI companies for tokens or they have expensive GPUs with them. Many companies use AI for work and train general models for specific tasks in-house. They aren’t using ChatGPT available to regular people. That’s why AI content can range from slop videos to really good software patches.
The experience varies a lot depending how you use it. That is for sure. I don’t k ow why some insist otherwise. However in this thread, most thoughtful answers disregard this because they consider a bigger issue with ethical and environmental impact it has.
If the people building and hosting models won’t build them from ethically sourced data and won’t host them without disrupting or negatively impacting the communities they are hosted in (noise pollution, power and water rates increasing, layoffs, diesel generators, etc) then I won’t use them and would advocate against their use.
I don’t expect the tech to go away but you can make a hammer from stolen resources and use it to harm individuals and communities or you can build it from ethically sourced material and use it to build things people need. If there’s an ethically sourced model quantized to run on commodity hardware locally without contributing to further wealth inequality, theft, and community disruption I won’t have the same moral objections to using it as I would for any of the current corporate owned llms.
I agree with you on most part and value your reasonable response. However, before we can get AI to run ethically on commodity hardware, I guess most of us will not be here, or will be very tired of it. We may need a major break thru in hardware to be able to run it and make productive use of it. I am continually in search of ways to run it locally and retain the same level of productivity to using it over the cloud. But truth is current state is still not ready for self hosting LLMs.
So I respect your stand point but be prepared that a change for the better is going to take a lot of time.
you can self host and train your own AI on a mini pc if you have enough ram.
You need RAM and GPU which are not cheap, and also lits of disk space to store the model. Thus is no longer a mini PC.
Well, if you say so. People are really emotional about AI, so I’m not going to argue with you or try to educate. Have a nice day.
Maybe you csn give me si.e recommendations I could discover?
anything produced by AI is slop
except for me I’m special, when I tell chapGTP to make cool shit it’s because I’m a genius with good taste
We all contributed to the training data, now they’re selling us our own work back to us?
I know you will not like my response and would not agree to it. I am not defending big companies, but having or hosting the data is the easy and cheapest part. Actually ingesting and training it is the hardest and most resource consuming. If it was that easy I would have done it at home and not have to rely on a cloud service, but since I tried and realized I would need at least a $2000 system to get anything useful I chose to pay a 5$ monthly fee. Compared to what it needs to run something that doesn’t just give a good output but a productive output would have cost me magnitudes more than $2000. So they are not just freeloading on our work. The real issue I would agree on and defend would be resource consumption like electricity and water.
Something you’re not acknowledging is that this technology stems from surveillance-capitalism and the surveillance-state. Google et al were financed to solve the needs of the state. Google search for consumers was a by-product from the state’s need and desire to make sense of their mass-surveillance programs.
Surveillance-capitalism operates under complete deception. Pokemon Go players didn’t know they were collecting data for autonomous-murder-drones.
Solving Google’s Captchas was unpaid labour: you were creating data for Google’s OCR program and now you’re creating data for autonomous cars or whatever their investors want you to toil on next.
Out of all this theft, unpaid labour and intrusive surveillance what does the working class get? Housing costs are increasing. Food costs are increasing. Poverty is increasing. What is AI going to do for us? It’s not building houses. It’s not harvesting crops. It’s not giving the working class any material benefit, but is further degrading living conditions. Enjoy the endless drone of a data-center! Enjoy breathing the fumes of gas-turbines! Enjoy talking to chatbots and never a human!
AI is deployed as a barrier between the working class and businesses. AI is deployed to decrease labour costs. AI is deployed to kill. The fact you can create text, images, videos and audios that mimic reality only means more surveillance on the working class because digital-signals can be faked absolutely.
You can’t eat what AI creates–it is completely immaterial, but it is being used to commit genocide. I’m sure that warms the cockles of the hearts in all of us.
It’s legitimately no different than cryptocurrency and NFTs, but this is the techbro scam that finally stuck after flinging shit against the wall over and over.
The training data sources are built destructively, without any consent from the artists, writers, programmers, anyone whose data is pulled into these models. The information these things spit out usually has so many inaccuracies - search engines are fucked, software is built without understanding a damn thing it actually does, poor translations of documentation are made public, the list goes on.
On top of that, data centers are one of the biggest environmental nightmares we’ve ever seen. Why are we pushing these so hard in the middle of a goddamn climate crisis?
There’s genuinely not a single thing it actually does well, and the only reason it genuinely took off is because investors in the same 5 companies are patting each other on the back.
It’s also worth mentioning how chatbots built off this stuff are designed to always be extraordinarily friendly and confirming to the user. It’s such a massive factor contributing to AI psychosis, to the point that there people who legitimately believe these lines of code are sentient and conscious. Meanwhile we’ve already seen just how easily these closed source models can be tweaked to spit out pure propaganda, making them an extremely valuable political tool - it’s no wonder fascists love this shit so much.
