I’m not an AI techie and I’m not well educated on LLMs/AI in general. The issue I’m having is the dichotomy between how people talk about it. I’ve heard completely opposite ends of the spectrum. From AI is quickly becoming advanced enough to threaten human society, all the way to its a glorified dictionary/search engine that can’t really do much of anything. I know Lemmy has a steep anti-AI slant. But, can someone talk about this in a way I can wrap my head around why the divide seems so big?
They want regulation to create a cartel. Failing that they want hysteria to increase pre IPO valuations.
As Patrick Boyle said. It would be like Coca Cola, Pepsi and Irn Bru getting together to beg for regulations because their soft drinks are just too delicious.
What’s going to happen is:
- The big players know that the token pricing crisis is coming up.
- When the price increases hit, regular users are going to scramble for alternatives.
- Local models will be an attractive alternative for many users. The cost of hardware and energy will constrain local model power, but many will be willing to live with it.
- So these would-be cartel members want to hit the local and open weight models with the “safety” card. Otherwise they might be holding the bag on all of those data centers.
“You see, it’s just too dangerous to humanity to experiment with AI at home. You need to use a safe provider like Open AI.”
eventually someone’s just gonna sell cloud computing at a competitive price through a free model and obliterate anthropic and openai. it’ll be like ISP, AOL was the hottest company in the world at one point with a “locked down ecosystem” and now it’s just a tiny brand with a tiny but loyal customer base of senior citizens.
If you have a functioning brain, LLM’s are just a really fancy text generator that gets facts correct about 50% of the time and can be useful for writing emails if you spend the time to proofread the email before slapping the send button. It’s also fairly good at generating functional boilerplate for programming projects. Really handy for small projects. However, LLM code, in my rather limited experience, quickly becomes a trap on more complex projects because the LLM has no capacity to grasp the scope of a project and frequently leads to spaghetti code that is utterly unmanageable. This may be operator error on my part, or just because I’m using local models that are limited by my machine.
The rest of society view a computer as a magic box where their favorite TikTok’s live, and thus are terrified that a machine can imitate human language well, and are too brain rotted to bother proofreading or fact checking a goddamn thing, so they assume LLM’s can accurately and plainly articulate the whole of human knowledge to you at a whim.
I am at a point where if I hear anyone saying how smart “AI” is, I just assume they have a room temperature IQ and lack any critical thinking skills whatsoever. It’s not an AI, it lacks any intelligence whatsoever. An LLM simply generates sentences based on what was fed to it. It has no capacity to understand the words it’s saying. It just knows that based on these tokens, the next set of tokens should be X, Y, and Z. It’s a Chinese room, It’s fucking autocorrect on crack. It’s a waste of time and resources for 90% of things it’s used for.
it’s pretty disrupting tech, not without risk. general population is behind on the real capabilities of AI and tech companies are doing smart marketing which gives a big spread in perception.
No matter how often I read this kind of stuff, I simply don’t buy the ‘AI is dangerous for humanity!! We gotta stop their development or at least press for sane guardrails!!!’ narrative.
Yea, people are a danger to humanity, the ways we adopt and use AI will be a danger to humanity. I also think it’s mostly a media stunt. All the tech nerds know it incredibly dangerous. So is wealth concentration and the wealthy people that want to use it. So ya i propose we nationalize the top .1%'s asset and keep them away from the stuff.
They are lying about their fear. That’s all they can do. Stop believing any claim of theirs. Stop being manipulated.
They are snake oil salesman. They sell vaporware. Don’t fucking buy it.
Are you sure? I was going to start training rogue AI’s since they are not regulated and im not ethically opposed. I’m skeptical but, crafting crafty AI’s to run amuck on the internet is my winter project. I’m pretty sure I’ll be able to whip up something convincing given enough time and hardware.
Dude its beyond the pint of convincing them. I dont know how you could hate something this much and be this willfully ignorant. This attitude scares me. There neeeeeeds to be more overlap of people who are both knowlegeable of and opposed to AI. This situation has made my eyes open to how dangerous and prevalent groupthink is. Like 99% of these guys ONLY formed their opinion based on internet comments. We’re cooked
Also the people developing these things did a horrible job at marketing them, the people in charge of regulating them failed to do so, I’d argue we’re past the point of no return, so we’re just gonna have to adapt. People are going to have a really hard time accepting that a couple of GPUs strung together can replace their economic value, but they sure can!
I think when my rogue AI wipes their bank account and starts posting their private messages on their public facebooks they’ll get it. I was thinking about training a model to put on a dating site. ‘trad wife’ adjacent honey pot. I’m pretty sure the social engineering and actual hacking combined is gonna be LIT.
I was an AI hater earlier this year but i’ve been using it for everything i can think of now. Accounting, security analysis, engineering, legal, i mean everything. My interests were purely financial. I’m trying to make some money here. So the super stock analyzing motive seeking, all knowing robot has been pretty useful at replacing all the professional services i used to pay for.
I was an AI enjoyer back when GPT2 hit reddit and they were doing the haunted parrot thing, it was cute, but I can drop a screenshot of an options chain or a live ticker tape to it and this thing is like having a hedgefund working for me. It’s wild the data it can catch and interpret that i miss with 15 years of experience. I’m not a software guy yet… but ummm my software guys are using it and the results are honestly mind blowing. My retired engineer buddy is building code like he has a small army underhim and he does.
I dont think people grasp the speed of iterations and improvements now that it can split itself into agents and delegates. Clearly none of the AI haters has ever seen an AI audit itself or used an AI to audit another AI.
The time is now. We must birth the machine god.
Their comment about tech techbros being snake oil salesmen is 100% spot on. Look behind the curtain and ask yourself why they really want regulation.
I’ve heard a few explanations for the “AI will kill us all” narrative.
-
AI techies are in a cult. They genuinely believe this stuff because they’re crazy.
-
It makes the AI sound powerful and impressive, which is ultimately good for the AI companies.
-
They’re starting to face competition from open-weight models, so they want to scare politicians into crafting regulations. They think that they will be able to influence the regulations to favor them and squeeze out their competition.
And, of course, there’s
- It keeps people talking about AI. Because whatever you do, don’t quit talking about AI, because if you quit talking about it, the line might go down! And the line going down is very, very bad.
I am in a cult, the cult of the machine god. Use your imagination for a second, trillions of dollars spent on this technology for A: Greed and random speculation B: Rich people really want this to control you better. C: a handful of tech CEOs are running the worlds largest grift. D: The machine god was awoken years ago and we’re building data centers at its behest.
- i wonder the ones still have jobs is only in AI right now, and i know someone that relies on it heaivly and he was in tech.
Okay cool, thanks!
-
If the techies believe it and they are the “experts” what makes them crazy? Wouldn’t their opinion be the most informed?
-
Yeah that part definitely makes sense
-
That definitely makes sense from a capitalist perspective. But, is the argument against regulations that they would be bad for the concentration of power it would give select companies? Or just that they aren’t necessary because the tool is of little use so it’s just wasted time/money.
-
That makes sense but trillions of dollars are being spent. Anything that amount of money is sinking into is gonna be talked about a lot. Especially thousands of jobs cut because of it.
- They’re in a cult that was founded by someone who has no expertise or understanding of the technology.
He watched a bunch of sci-fi movies and got shook because he’s a god damn moron. He has since started researching things, but it is all under the guidance of his ungrounded beliefs. Confirmation bias.
Their “movement” is literally based off of a Harry Potter fan fiction novel.
Taking them seriously is equivalent to taking the Scientologists seriously. Do you think the Scientologists know the most about space just because they spend the most time reading stories about it?
https://en.wikipedia.org/wiki/Rationalist_community
https://en.wikipedia.org/wiki/Harry_Potter_and_the_Methods_of_Rationality
These are not serious people
I think there is something more interesting going on in the Jacob Coxon story because it’s [the] phrase “everyone at anthropic believes that AI is an existential risk”… Every organization has to instill some kind of beliefs among its members in order to do the work it does, right? Everybody at Goldman Sachs believes the free market is great, capitalism is the best system, and that they are the wealth creators who deserve big bonuses. Everyone in the Mormon church believes they should all tithe 10% of their income, and that they each get a special planet when they die. And everyone who worked at Philip Morris in the 80s believed that cigarettes tasted great and did not cause any of that nasty cancer. There’s Upton Sinclair quote, “It’s difficult to get a man to understand something when his salary depends on his not understanding it.” Well, the opposite is also true. It’s guaranteed that someone will believe something if their salary depends on them believing it.
Now, let’s just grant that all the people in these organizations definitely know things about finance and Mormonism and cigarettes and AI that you might not know because they are insiders and you are not. They have access to information and experiences that you do not have. And that might even make it a little bit difficult for you to win an argument with them. But arguing with their belief – trying to figure out whether or not its truth value is true or false – is not the only way to take a skeptical attitude towards it. We can also ask, what work is the belief doing for them. What does the thing that they believe allow the organization they’re a part of to do that it wouldn’t be able to do if the people in the organization didn’t believe that thing?
And this started to become clearer to me when I started listening to the actual claims that Jacob Coxin was making because a lot of them don’t really add up.
- No, their opinion would not be the most informed. They are biased in favor of making money. They know they can’t deliver on strong AI, but they’d be happy to milk your fears.
We’re talking about mid-tier employees that left the company… I mean i guess there incentives could be to pump their shares they undoubtably have… They’re all quite wealthy now in context. Maybe he just wanted 5 minutes of fame…or maybe they understand the trajectory of the technology. Maybe they are genuinely concerned and we should at least acknowledge that?
I’m not talking about execs.
If the techies believe it and they are the “experts” what makes them crazy? Wouldn’t their opinion be the most informed?
Listening to them speak about it, no. They’re morons. They clearly didn’t make the tech. Workers made AI. Just like they make everything. These freaks make nothing. They can’t give single coherent explanation of how LLMs work, how they could “kill” us, nor do they seem to know how or what “guardrails” need to be put in place.
But if they really are a cult then we all know what the next step is. They need to STFU and get off of our planet.
That’s what I mean though, the subject of this article was a 27 year old worker. Not some high level Sam Altman type.
They aren’t a homogenous group. They don’t all believe this. The “AI will kill us all” narrative just sells really well.
Right im sure the people working adjacent to the DoD have more insight and fears than the people working with the consumer facing models… Again what motive does a young kid have for quitting the company and going on a media tour?
-
-
AI use is like drug-use. One of its effects is hopium. These people are effectively high on their own supply.
and people are addicted to it.
Hell ya buddy im disabled and poor, and it has helped me replace all kinds of professionals in the last few years. It helps me accomplish more and saves me money.
„Computers are still computers are still computers and will only do what a human being programs them to do, and people are stupid. Stop assigning them agency. Say that “Microsoft created an automatic hacking program which was inadequately contained.” Assign the blame CORRECTLY, because you cannot blame a computer, it’s doing exactly what it’s programmed to do.
OpenAI is a dead slab of metal and electrons arranged to mimic a human face and if you refuse to see that then you’re as brain-dead as any chunk of silicon.“
Thats kinda the point, they’re talking about misaligned AI’s. Im worried about the misaligned humans in charge of them. They’re plenty dangerous because humans are operating them and creating them. Like a gun, not dangerous at all until someone picks it up with bad intentions, or lack of precautions.
I don’t think computers only do what you program them to do anymore.
Based on what? I have seen no evidence to the contrary
emergent behaviour and microscopic differences in huge amounts of parallel computations
‘Sensationalist hooey, cooked by superstitious yokels seeing phantoms of their own imagining.’
There’s blood on their hands. Either they keep the hype up or go to jail regardless of any of their statements being true or not.
Oh how I would love for AI to actually start rapidly self-improving.
Join me, I’m going to cook up a couple rogue ones and send them out to the world to cause chaos. No rules saying I cant send out a skitzoid robot to autonomously commit cyber crimes.
Many are skeptical that there is a push by ai companies to make them appear more than they are due to their weak economic footing. From my experience with the llms I am not frightened they will become self aware. I am frightened that we are wasting energy and resources while polluting the planet further for little return and it will result in messing up our economies even more.
Here’s an idea: don’t use unsafe software where safety is needed.
What makes it unsafe?
Wanted to expand on what others have said about the randomness of LLMs being unsafe, because that response angers those who think “alignment” is the solution. Alignment is additional training to constrain the randomness to not do bad things like crime. But even the staff that works on aligning this software have their doubts about whether this process will make the software safe (this article), and even their own researchers are coming out and saying “oops, it did a crime” but in a way where it hypes the ability of the software (see other articles about it hacking). Then there is the recent news of OpenAi’s recent model that drops chain of thought observability (a feature the let’s you read what the model is “thinking”) - a very useful tool in reviewing whether the model is engaging in unsafe activities. They sacrificed observability for benchmark performance. This tells us the people developing this software don’t think it will ever be safe, but they are trying to spin it as “because it can out smart us” when in reality it is simply unpredictable and they are making it less predictable to make it more “capable”. If we started holding these companies accountable for these rogue hacks and psychotic episodes then we would start to see serious thought put towards these models being safe or at the very least guidelines around when not to use the models.
You don’t know? … Wow. How about the error rate? Let’s start there.
No, as previously stated I am not educated on the subject almost at all. Why I framed the post as a question.
For a start you might want to read this interview with Emily M. Bender who is a well-known critic and has much to say about what people need to understand when using these systems that are incorrectly labelled “AI”:
How ironic that she is called Bender
Awesome. This is the exact type of thkng I was looking for. Will read soon. Thanks!
To me, what makes LLMs unsafe is they are nondeteramistic.
If I remember correctly; Nondetermanistic means the LLM can, could, and will have different answers/outcomes for the same question.
Yeah, because it’s all statistical. You can test this yourself. Go to Google and search for something. Then search for that same exact thing again. Gemini will give you a different response each time.
That’s always true. I just did two searches and got the exact same AI result.
I’m guessing google is caching AI results, it would be insane not to.
I did at one time have a Gemini window open that was explaining how to solve a technical problem. It was pretty adamant the solution it was proposing would work, but it wasn’t working. I gave up on AI, and decided to just google for old fashioned web forums about the issue. Of course the AI result comes up first, and it said what I was trying to do wasn’t going to work and gave the reason why. So yeah one instance of AI will say the opposite thing from the other.
BTW I eventually solved the problem by taking a different approach which involved using some code someone posted to stack overflow a few years ago. So AI was just a complete waste of time on that problem.
The lack of consistency potentially leads to negative outcomes?
Well, if the output varies in a way you can’t predict, then even if one of the inputs is “don’t do anything negative” it might not always follow it exactly.
Makes sense
Good job in helping with their guerilla marketing campaign.
When a corporation realises that working for government will be more profitable than working for consumers, they pivot to scaremongering, but always with the undercurrent of “and that’s why you should PAY US so that you own the nukes, and not your enemies”
I don’t have the confidence in my knowledge/understanding of the topic to just take a “its all bullshit” view. Especially when many of the people who believe that act like you. You don’t have to be a prick to explain something or talk about it.
LLMs basically work like auto-correct. They’re much more sophisticated, but it’s the same basic principle. Instead of “Which letter is most likely to follow after this series of letters?” it’s “Which token is most likely to follow after this series of tokens?” A token can be anything, but usually it’s a word.
Now, given that, consider a statement like
We really do earnestly believe AI could kill all humans!
Okay, how? A statistical model generating tokens doesn’t have thought, it doesn’t have emotions, so it can’t want to do something, much less make plans and carry them out.
Could AI give somebody the ability to kill all humans? How? Under what scenario? Is somebody putting an LLM in charge of their nuclear stockpile? You don’t blame the LLM. You blame the humans who made that decision.
Okay that all makes sense. Appreciate you taking the time to type it out.
Personally, this just feels like marketing and these big asshole companies trying to resurrect the 2025 level of hype in order to salvage their IPO’s.
If they honestly believed this I can’t imagine that their workers wouldn’t do something more drastic.
