Humans have always overestimated the power of tech. The reach of our imaginations has always exceeded our grasp. I don't think this tech will be replacing people any more than computers themselves replaced people. "Computers" once meant actual people that would sit at desks and compute equations. That job is gone, but its erasure opened up new possibilities that could not have been imagined before. When "modern" computers were first invented, it was assumed that they would replace all humans within a generation. he wrote on his pocket computer from his normal human job that used computers.

The issue is, I'd say the same applies to every model that produces useful outputs.

That you know of. Your lack of awareness is not an indication on the stance of the technology itself. If you have a problem with Grok, as I do, then condemn Grok, Twitter, and that pathetic man child that owns them.

If you tell somebody that the LLM they're using is unethical, they'll nust go use another convenient corporate model.

Yeah, so actively fight against these toxic corporate activities. Fuck OpenAI. Fuck Google. Fuck Twitter. But the instant negative reaction to their shared tool is getting quite a few additional people caught in the blast.

And "public data" is not enough for me, because that typically means scraping copyrighted content from public websites, I'm not aware of a model that uses only data with permission

Again, this is an expression of your ignorance, not reality.

OLMo 2 was trained on Wikipedia and other fully public forums. Its training sources and data is fully accessible and open.

GPT-NeoX is an untrained model that you can train yourself. It's literally just the foundations anybody could use.

Pythia is fully open source, and one of its training sources was GitHub. Which does genuinely bring in to question whether or not you can use GitHub to train AI. Personally, I don't see how branching a repository is any different than using the code to train a model. But the current anti-AI trend has people EXTREMELY sensitive to this concept.

Please explain to me how a wristwatch is sentient or conscious.

My wife is a materials science researcher, and works for a university here in Germany. She uses AI to drastically increase the speed that she programs calculations and simulations. Yes, it regularly spits out nonsense, and she's constantly having to refine and tweak it, which she can effectively do because she has a PhD in the field. She has repeatedly told me that the actual programming was always the least favourite part of her job. But now that she can get an LLM to write the actual code, all she has to do is go back to refine it, and spend more time thinking, and less time reading stack exchange.

Another group used Evo 1 and Evo 2 to create novel genomes for bacteriophages in the fight against antibiotic resistance. There is a nuanced debate on whether those models are technically LLMs, since their language is actual sequencing rather than written descriptions. But they are structurally nearly identical.

My personal project is iterally just for funsies, to see if I can mimick human learning and dreaming by combining an LLM with a more abstract NN. So the weights are never frozen, and it actively hallucinates and dreams scenarios between weight adjustments and re-initialization.

I don't know how to make it not lie, and I'm not sure if that's possible. But I see that akin to asking "But how can we make X-rays safe to be used in shoe stores again?" Maybe we can't, but that doesn't mean we should ban them entirely.

You and I seem to agree whole heartedly with the harms. Over use, abuse, misunderstanding, etc. are terrible and causing real harms.

But I see people pointing their ire at every instance of the word indiscriminately, actively antagonizing those who are actively trying to make the changes they supposedly support.

Yeah, except that's a complete shift from "This technology is bad," to "I don't like the way this technology is being used."

You could write a 5,000,000 word thesis on all the ways ANY technology could be abused, and it STILL would not be a refutation from its beneficial uses.

Explain to me why open source LLMs, trained on public data, so both the training data and the resulting model are both 100% free and open source, are bad. For something to be universally bad, it CANNOT have any positive uses. Otherwise the real assessment is that it shouldn't be abused, not that it shouldn't exist at all, which I 100% agree with.

Because that's the claim that needs to be defended. I am not defending OpenAI, Anthropic, Google, etc. at all, so suggesting I am would be dishonest. I am not saying it should replace teachers, or scientists, or programmers, or anything like that, so suggesting I am would be dishonest.

I am not defending the corporate practice of explosive expansion, which causes these major issues. As that is a product of psychopathic capitalism abusing a technology, not an inherent property of the technology itself.

If you want to justify how an emotionless, mindless, will-less technology is inherently bad, you need to address its BEST, most ethical uses. Not its worst. You're making a universal claim. I reject the universal claim. I don't need to explain the abuses, because I am not justifying or defending them. The responsibility is yours to explain how the ethical uses are somehow bad.

https://lemmy.zip/comment/28450273

Why act like an ass?

Because if you post a public message on a public forum that everybody can see, you don't get to decide how your words are used. That's the fair use license. It doesn't require additional consent.

And how precisely is it unethical to train an LLM on public data? You keep making that assertion, but it's built off of a supposed "Lack of consent." When the fair use license was brought up, you seemed to retreat to the concept that using fair use data to make fair use software freely available to everyone is somehow bad?

You need to explain how the process itself is bad, not just use guilt by association. It seems more like you're building an opinion based on vibes, not actual reason.

Good for you. Keep wielding that shield of ignorance indescriminantly. Do you also think cryptography itself is harmful, or are you focusing on the specific use and harm caused? Guess what, the technology that made the crypto stupidity possible was used for far more than just gambling on imaginary coins. Once again, you're conflating the worst implementation of a technology with its use as a general rule. Would you defend a small set of positive and ethical uses of LLMs? Or do you condemn the entire tech, on the assumption that it is inherently bad? Do you know precisely where and how the tech developed for crypto is helpful to you, and improves your privacy and safety?

Lol no. Actually cells are demonstrably NOT sentient. They have no capacity for choice. No capacity for emotion or experience. They are literally molecular machinery. And we(scientists) understand that machinery to a shocking degree.

If you think cells have any degree of sentience, then you must make a similar argument for advanced clockwork machines. Not robots, pure wind-up clockwork automata.

Being a sarcastic ass does nothing to refute my point. I don't have to be an oracle to look back at history and see 100% of technologies finding an eventual use. I don't know how the problems will be solved, and I don't need to know what their uses will be in order for the trend that has continued for 1,500 years of recorded history to most likely be true.

You, on the other hand, have to somehow show that THIS technology is so wildly unique that none of the examples from the last 1,500 years are informative.

Edit: sorry, got some things mixed up in my brain. Just because you can't identify a use for a technology doesn't mean a technology is bad. Unless you are personally familiar with every single technology, profession, and process on the entire planet, and those that will be invented in the future, and can act as the ultimate arbiter of what is and isn't useful. Alcohol is very harmful. But also good at sterilisation. X-rays are very harmful, but also good at medical imaging. Gunpowder is very harmful, but also makes pretty fireworks.

Cool, then go after the corporations shoving tech in places we don't want it.

Go after the data centers that are displacing families and wasting resources.

Go after the media for misrepresenting reality and distorting truths.

Those are all real harms with real actors. Getting angry at the camera because someone took blackmail photos of you is nonsense.

I got called unethical today for using a small-scale locally hosted open model trained on Wikipedia and Stack Exchange. A LOT of the anti-AI people are universalists that have decided carte blanche that it is an evil technology and everyone who uses it is bad.

Are you aware of "Implied lack of privacy?" It means places where you cannot reasonably assume privacy, like a public beach or supermarket. Public forums and public wikis are about as implied lack of privacy as it gets. If you post a guide on stack exchange on how to solve a programming issue, you revoke your rights in how that information is used. This includes using it to create new software. If you decide you don't like a program I created using your guides, well... That's too bad for you.

And no, this is not victim blaming, because I'm not defending actual theft or telling people to protect themselves more. I'm stating you cannot steal something that has been freely offered to the entire world.

Lol, tell me you don't understand anything about the technology without telling me you don't understand anything about the technology. Nobody here is defending scraping artists libraries or using copy written material. Go find a different strawman to cry at.

But you see, it's okay when they gatekeep because they're defenders of truth and humanity! Gatekeeping is a good thing when used on people you don't like! What's that about equal accessibility and ableism? Shut up! AI bad!

Yyyyyyyyup. Even attempts at honest discussion get polarized. "If you use AI(LLMs) you're an unethical AI-Bro!"

Oh... I didn't realize that my self-hosted open source small model trained on Wikipedia and Stack Exchange was unethical. Thank you for showing me the error of my ways.

It's a testament to both how easy and difficult it is to look out in to space.

Easy, because even the weakest signal travels immense distances, as there is almost nothing to slow it down or stop it.

Difficult, because the only reason we can receive this signal is because we can pinpoint the exact location and frequency of said signal.

There could be aliens sending us messages right now, but buried in the noise of the cosmos. 🤷‍♂️

Cool, your personal experience doesn't dictate the global population. And you just reiterated my point. What about the people in your circles that don't talk about it at all?

I've tried to explain to SO MANY people that the technology itself is not ethereally evil, and it's possible to be both pro-AI and anti-corporate abuse.

LLMs are VERY good at bridging the gap between human language, and computer language. That's it, that's their whole gimmick. They're not oracles, they're not God's, and up until ~5 years ago everyone was fucking thrilled to get a Star Trek computer of their own.

We literally have it now...

We have computers that can generate entire images and programs through voice prompting just like in Star Trek. And EXACTLY like in Star Trek, it gets it wrong, CONSTANTLY. Which is why humans are constantly fixing it and correcting it.

I often think about all the times civilizations are refused technology, and it's explained that they need to earn that tech on their own or else it will be abused and misused.

Well we have that tech. We are abusing and misusing it. But this obsession with claiming that something inherent in the tech itself is unethical or wrong is absolutely bonkers.

Lol seriously? They weren't saying it was direct causation, just that it probably didn't help, and there was likely other underlying issues. Is mental health so incredibly taboo that even talking about it is now unacceptable?