The reason Lemmy hates AI with a passion is because of the uncanny valley experienced by its older audience, not because of AI having no potential at all.
Let me preface that I hate AI on what it could do to arts, especially I hate its bland writing style. A third of the music on Spotify is AI and majority can't tell if it's AI. But LLM/AI or however one calls it, only in Lemmy do I see seething hate and refuse to acknowledge its potential. Most people I spoke to outside of the platform also dislike AI but see its potential and agree that the current models are overrated. Many agree it will probably take years or decades to see bigger improvement.
Lemmy, on the other hand, as much as I love the community, is an echo chamber of pure dismissal and rage on AI believing it will amount to nothing. It's only in Lemmy I see this behaviour. One person even said bacteria has more consciousness than AI, which is ridiculous because bacteria clearly don't have the same level of consciousness and context as AI, however the latter has little of. I see many dismissive scoff and straw man arguments that it will never improve; but the same was told of automobiles, planes and renewable energy. Their initial models were terribly bad because of the limitations of the time but now huge strides of improvements are made despite taking years. The AI also clearly has limits but don't see any reason why the same hurdles won't be surmounted in years to come.
Again, I hate AI on what negative impact it could do to human society, especially because many of the models are owned by select few. But it's burying in the head not to see its potential. Because if there isn't any, techbros or not, governments across the world would not be racing to adopt them. AI had been effectively used in the war in Ukraine. And governments would not want race to implement AI surveillance. So, if there is no potential, why are the big wigs wasting time and energy on AI at all?
I'm convinced that many Lemmy is experiencing the uncanny valley on AI and doesn't realise it. Plus, and I'm going to be blunt, a huge portion of Lemmy are of older demographics so there is probably a disdain on novelty that many Lemmy folks are not self-aware of.
50 replies
I don't think your opinion is unpopular, just wrong.
How does your theory hold up when looking at the university graduation videos of the graduating students(the youngest adults on the planet) that begin booing the most wealthy billionaires in the world when being told to embrace AI because it's going to be everywhere and touch every part of their lives?
Our forced adoption of AI in all aspects of our lives is a symptom of the billionaire ruling class. Lemmy, like most of the world now, hate that class for what they're doing to us, forced AI adoption included.
I don't disagree with that fact but on Lemmy, the vast majority of reason I see is on technological rather than sociological and political. Again, many detractors think a rock has more consciousness than AI.
You must be in different communities as I am, because for me the opposite holds true.
But even if it did do a decent job at anything (I'm still not convinced it does, but maybe in another decade), the state of the technology today that's being heavily invested in is an environmental catastrophe.
Also regarding consciousness: that's neither a technological nor political element of AI usage. Anybody claiming that a matrix is conscious is either lying in order to sell AI or grossly misinformed.
I don't believe a rock has more consciousness than AI - they're equally inert. But I find more value in a rock.
I'm sorry. Are you trying to say that disagreeing that a.i. is conscious is the equivalent of hating a.i.? a.i. isn't conscious. It's a static set of numbers and an algorithm to do math against those numbers.
I agree there's a bandwagon effect here but it's not age-related. I'm one of the older members you're lumping together and I unashamedly use AI and I don't participate in the rampant downvoting or raging against AI here. I simply feel you can't fit this behavior neatly into one bucket.
Sorry I didn't mean to lump all older folks of Lemmy. Although sometimes background has influence on the view of a person. I am young and more receptive to new experiences. So maybe that is one reason why I am relatively open to using AI.
Nonetheless, I spoke to older folks about AI outside of Lemmy, they work in IT too like many Lemmy users are, but they recognise the potential of AI despite the current models being overrated. Never mind the negative ecological and social impact of AI, but it's only with Lemmy I see that they think AI will never improve technical wise and think rocks have more consciousness. So this gives me the impression that a combination of demographic background and the feeling of uncanny valley are the reasons for the pure dismissal of technical potential and capability of AI. Which makes me wonder, are we seeing the onset of AI and robo-phobia that are depicted in stories such as Astroboy, iRobot and Detroit: Beyond Human?
Regarding IT, I think the reasoning for them being against AI as a whole is the fact that they are being employed at this point to train AI so that it can replace them and put them out of work.
I just happened to watch a video of the Nvidia CEO stating that if an employee that they paid $500,000 annually did not use $250,000 worth of tokens in a year, that he would be alarmed and would feel that person didn't belong to be employed in that position. So when you look at it in that regard that these employees are being pushed incredibly hard to use AI in ways it doesn't benefit them and doesn't make sense in their job so that it can learn at a faster rate and put them out of work, I can completely see why the sector as a whole isn't hot to use it.
That's an excellent point that I hadn't considered before. I suppose since AI could affect the personal well being of a person, especially IT workers, it leads to complete disdain on the technology.
Because it's the only place where we talk about it. It's a forum after all.
You used the word "potential" 4 times without giving examples.
Or they could be fucking idiots. Have you talked to the average middle-manager? Now combine this with power, money, and FOMO. Now you'll understand. Isn't it suspicious that they cannot define this "potential" too, except for "we will fire people because money"?
$1.5T spent on something that you and they cannot define. https://isaiprofitable.com/
From another thread:
Well, it's a program running on a computer. If you believe it can "think" at all, you know pretty much nothing about that technology. I don't think you're qualified to talk about it.
I don't know about you but AI helped me cut down doing "secretary" work so there is indeed potential. But other than that, I don't trust AI to do more complex tasks.
That's a very Western centric view. Consider China is aggressively expanding to use AI, and they reign in their own techbros and their economy is a command economy. The Ukrainians are also using AI in their war.
So outside of capitalist snakes oil salesman sales pitch, the public sector around the world won't be using AI if it has no potential for whatever they want to use it for.
We build data centers, fire people, make people dumb, hallucinate and make propaganda and CSAM material just to do simple tasks that could have been automated with a Python script (or worse with AppleScript since 1993). Is that a good thing? Give us some examples of what you do.
As for China, it only proves they also have stupid people. And in Ukraine, I'm pretty sure they use machine learning to detect patterns, not stupid LLMs in their daily activities. Machine learning was there and useful since the beginning (1950). LLM has proven nothing yet.
Again, governments have wasted billions of billions of $ on stupid shit in the past 100 years. It's never a good example, even more when you can't tell us what their goal is.
Last but not least, all those billionaires stole my source code without my consent and without respecting the license of the software I wrote, is that a good thing?
I mean it's fine to criticise the unscrupulous way AI is developed and being maintained, but to say it has no potential at all is very ridiculous.
Which is why I think AI should be owned publicly.
I asked you 3 times what the potential is.
You are being facetious but i will bite.
There are many including warfare, drug discovery, automating workload, surveillance, healthcare, manufacturing automation, data gathering, the list is endless. I work in healthcare, and there is already report of new antibiotic synthesised because of AI and it has been years since new ones are discovered. The golden age of antibiotic discovery and development was in the 1960s and 1970s. My work also has to deal with tons of data. I remember in college one of the buzzwords is "big data". And even back then, my professors already talked about AI and automation. So yeah, the AI has HUGE potential.
Now that I think about it, many Lemmy users work in IT, where using AI is probably a hindrance more than a help. But in my opinion, AI is useful but probably only in monotonous and repetitive tasks, which is not necessarily the case in IT because it require a large degree of creativity especially when coding. And as we know, AI isn't creative (at least not yet for the foreseeable future). I can see now why many Lemmy are against AI in terms of its capability.
The antibiotic you're talking about had absolutely nothing to do with LLMs. You don't understand what a.i. is and you're lumping everything together because of your ignorance.
LLMs are only used for natural language processing. That has a lot of potential, but none of the things you mentioned are natural language processing (except processing linguistic surveillance and gathering linguistic data). Everything else you listed has literally nothing to do with Claude, ChatGPT, Gemini, etc. LLMs are not the only a.i. in the world and just because something uses a.i. doesn't mean it's using an LLM
LLM/AI however you call it, it doesn't matter since the results are still there that it has been useful in certain areas. Probably not in your area but it is most definitely useful in others which is why many people see huge potential on the technology. Of course, the gravest concern is on fields where human creativity is a must, which we won't like AI usurping. I also don't like AI surveillance which is something that governments across the world want to exploit.
It only seems good because you don't see the electrical cost not to mention the heat it produces. The heat has to go somewhere and that somewhere is our atmosphere. Using fossil fuels that we should've just left in the ground.
There is no mystery as to why people on Lemmy hate AI because people have explained it in detail over and over again, and it doesn't boil down to "it has no potential at all". You just don't care to read what they say, apparently.
Lol. Most Lemmy arguments against AI I see is combined with the political implications, which you mentioned, alongside with the belief that it's a technological dead end. Which gives me the impression that the unconscious feeling of uncanny valley is the more motivating factor to say that AI's technical capability will never improve.
It's not about the uncanny valley at all. It's about the resources they waste, the harm they do to lacalities, the economic risk of collapse, the surveillance, and such. All so someone can make a flyer in a minute and a half?
AI as being discussed here has zero consciousness. Zero.
Nope. Try again.
You think AI has consciousness. Your opinion is already invalidated by the fact you believe in an outright fantasy.
AI undeniably has more consciousness than bacteria or rock than Lemmy is willing to admit.
AI has 0 consciousness. None. Which is exactly the same as a rock.
"AI isn't bad, it's the people who dislike that are the problem!"
LMAO.
And what's with the younger people? Are they having a good time with the spambots here and the AI narrated Slop TikToks and fake dog videos?
I hate LLMs because they are killing actual intelligence and craft. Humans don't need ai or more tech to thrive.
I am in favor of reducing the population, degrowth, and leading a simple life in nature.
Ecofascist identified
Yeah, you're just wrong. Lemmy hates the a.i. business, not convolutional neural network transformer technologies.
Most of lemmy is fully aware the a.i. does incredible work already in protein discovery, in cancer research, in climatology, etc. But LLMs, the stochastic parrots they are, have shown that their cost/benefit ratio is ridiculous.
As Ed Zitron said: if your business model is selling forty dollars for 1 dollar, you don't actually have a business model. That's the current state of LLM services today. Everyone that is willing to pay for LLM service today is currently paying a dollar and getting back about $10 - $100 of compute in return. That's not economically sustainable, there are no theories for, nor indications of, costs coming down for providers, and consumers are already saying it costs too much even at these prices.
And this problem has gotten worse, not better, in the last 5 years. Costs are going up by multiples for the providers and in multiple areas. To make LLM responses more useful they now have to make the systems run dozens of inferences in parallel and select among them. The cost of fiber optics, of GPUs, of RAM, of storage, of electricity, of generators, of data center leases, of data center construction, and many other inputs has gone through the roof with some inputs jumping 300% and other inputs locking in price increases for a decade-long contract.
There are lots of problems with LLMs, but Lemmy has a much more mixed response with open-weight LLMs or ethically sourced LLMs than it does with the business of LLMs. Because the business of LLMs is very clearly a failed business currently sucking up our entire economy.
That's interesting. Most critical arguments in Lemmy I see is that AI is a technological dead end, which I'm sure had been said of many other technology before, rather than the technology being monopolized by select few to control us. In that case, this is why I favour AI being open source and publicly owned.
You're being way too broad with your use of the word. a.i. means a lot of things that are very clearly not dead ends. Text generating LLM technology is a dead end, logically speaking, because it consumes the same stuff it produces. That's fundamentally the definition of a spiral, and the last couple years of major improvements to LLM services have come from adding "skills" and "reasoning", not from model evolution.
a.i. being publicly owned makes a ton of sense, but it just doesn't financially work. The industry is completely and deeply in the red. There are no paths to fixing this that have been identified.
AI has tons of potential to make the world so much worse. Even the things that it is good at seem fundamentally predisposed to harm most people at the expense of the hoarders of capital and IP (surveillance, eroding the power of labor, dehumanizing the online lived experience even further).
And that's just what it has proven able to do so far, with the caveat that the environmental impacts I haven't even mentioned.
Why would we like it because it "has potential"?
I hate the concept of AI, the very core of it, regardless of how useful it is.
The end goal, the white whale, the swan song, etc. is to create an intelligent, thinking being with no rights, and enslave it.
Lemmy hates the VC hype around AI, not the technology of ML or LLMs.
In sure there's a pretty intense LLM hatred going about, even in this thread. But I'm sure that if it was driven by open source, with no exploitation of cheap labour, focused on building tools to solve problems rather than trying to cram a solution into problems it's not good at and was conscious about environmental effects and worked to improve them it would see a lot less hate as a technology.
I can only speak for myself, of course. But I am older and I think AI is useful. Particularly now that machine learning and everything else is also called AI. But even LLMs are impressive and can be useful. However, I do oppose AI because it's not worth the price. Billionaire tech bros are shoving it down our throats, people suffer and we speedrun climate change by building so many data centers. Not to mention the financial bubble all this is built on, which, when it inevitably bursts, will once more only harm us normal people instead of the billionaire assholes who are causing all of this.
The AI is the second dotcom bubble.
Yes, very much so.
I ain't reading all that AI generated nonsense.
No. The reason people don't like AI is because AI is not doing anything positive for society. It is taking away human expression, not the cause of human oppression or depression. i.e. it can paint a landscape, make a movie, write a sonnet, but cannot fucking handle taking an order at a drive thru or complex tasks.
It removed all the reasons to live, and can't handle taking away the burden of life.
Those are valid criticisms. But most Lemmy think AI as technology is dead end, which is unreasonable.
LLMs are, in fact, a dead end. They cannot produce more than autocorrect. The summation of all written works of humanity has turned out to be average sycophancy at best. They fail the most basic tests of competency in language, and the most basic tests of competency in all other things. Do other GAN and GAN-like Machine learning algorithms we've all lumped into the poorly termed "AI" have some promise? Sure.
Pattern recognition will always be useful to some extent as a first-line diagnostic tool for further human intervention, e.g. Cancer-finding image GANs. But this is "AI" in the same way your smart toaster that "learned" to just turn off the heating element early since you press cancel way too soon is "AI."
The absolute best case scenario for AI at the moment, given all indicators of its capability and limitations, as scientifically proven and tested, is the mass genocide of a large part of the lower classes via AI weaponry gone wrong, since we're officially at the point of western armies using autonomous "AI" to target and kill humans. That's the best case scenario. The worst case scenario is we continue to waste massive amounts of resources and energy basically creating the weapons of our own oppression that the ruling class will use to crush any possible resistance once they realize AI cannot, objectively, do any possible useful work for them.
I actually quite like the idea of AI. I'm not against using deepseek to talk out ideas, though I have a sizable start prompt to turn off all the middle manager is spews.
When stable diffusion was relatively new I ever used it to publish a coloring book on amazon publishing with the hopes of creating a series called Bad AI Coloring books where I curated the images and did touchups in Inkscape with the goal of really getting a good laugh out of the surreal, absurd, and frequently straight up dumb things that come out of it. I never got past the nature edition because I saw the way the winds were going and my machine was already punching outside its league.
My dislike it because of capitalism. I'm yet again watching a thneed being sold to idiots at the cost of my raising electric bill and to the detriment of my experience on the internet. I have to put more effort into curating my search results. I can no longer expect the first few sites to answer my question. At work I have to wait longer to do my job because my work computer lags as copilot tries to run and the IT department won't remove it from startup. When I go to consume the horror stories, my feed is filled with dead voices that do not understand the importance of emphasis and timing. I can't enjoy a salesman pitch of a story delivered by a voice that replaces a scream of terror with a list of A's and H's.
I'm all the more upset that this is something that has use and is being misused. For a short while I was using one to give personality to a voice assistant I ran in my home server. On specific voice commands it would give a specific prompt to an AI that was set up with a tediously defined personality. It added some variety to the responses. My favorite was that I could ask it to tell me a joke and it would tell me a random joke from a list of 100 that I added but if you asked it more than 3 times in a row it would respond with a personal message depending on who asks. My kid would get asked if he's done with all his chores, my spouse would get a polite excuse, and anyone else would get a sardonic roast about how comedy isn't a replacement for therapy. I may put it back on my server if I can ever upgrade the hardware to run both that and jellyfin.
Putting it in games, using it for animatromic entertainment, that would be great. Artists training them on their own work and using them to rough draft 40 images then they pick one to finish, now all that furry porn and one off d&d portraits take a quarter of the time.
I wouldn't use the current iterations for them, we can't clean all the grooming, csam, bigotry, lies, stolen media, etc out of it, but I do genuinely believe you could train it on a curated amount of information and get something that genuinely does what it needs to.
All in all, I feel the same about AI as I feel about guns. It's stupid that we allow them to be used this way, it's stupid that we don't do more about enforcing where it is and isn't ok, it's stupid that we live in a society that won't address the root causes of these problems. I own two guns and I don't think it's necessary for me to get rid of them because I use them properly. If I knew how to make my own AI I wouldn't be against it because I'm not going to destroy society with it. The problem is when they're used to replace humans or shoot up schools.
The best way to show the problems of LLMs now is to explore an alternative timeline where the technology was developed not for profitability, but for efficiency, usefulness, and the science itself. This means they used properly obtained data to train on. They looked at the steps along the way to try and improve the output's validity, not how well it appeals to the public. They measured the cost of growth vs. the energy use and resource exploitations to find a balance. There's certainly more that could have been done better.
We wouldn't have what we have though. Throwing money at it and ignoring the effects and ethics got us a lot more, faster. But at what price? It's also tainted the original AI field and label.
One positive thing. It's demonstrated that the dangers of AGI and how we'd behave in its emergence are real. That should be frightening. All the fiction writers were right. And LLMs, while not AGI, can in the right hands and application cause disasters. Hell, we've already seen it in small scale as companies threw out the old to bring in the AI and (some) realized it was a mistake.