Isn’t Copilot basically an OpenAI GPT interface?
Don’t know. Don’t care. It’s a Microsoft product, thus an American product, ergo something I will never use.
“Elbows up.”
at least deepseek is opensource🤷♀️
I think Grok is too, at least an older version. There’s also gpt-oss, and Meta has released a lot of “open source” models, but I think they use weird licences. Meta and Deepseek (and Alibaba) researchers publish papers that are actually useful, while the rest just publish marketing material, trying to keep the research itself private.
Gpt oss is borderline crap, it’s not that smart, not that great and it’s pretty censored, but it can have niche uses for programming. The oss 20b in particular can be easier to run in some setups than their competitors like Qwen 3-30b. oss 120b is quite heavy: the cost to performance ratio is not good.
Meta abandoned the open source ideal since Llama 4; they went closed source.
Older open source versions of Grok are literally useless, no one should use them. Their cloud closed source models are decent.
Deepseek and Alibaba’s models like Qwen are good.
Accurate.
I can’t agree, the LLM’s don’t have the capacity to be evil. They may be called AI but there is no “I” anywhere in there.
The companies however…well that is a different story.
Those logos are just as closely related to the companies as to the code
They’re actually corporate logos, so … yeah. By definition.
Idk, there’s always that argument how technology is neutral. But is it? I mean tech isn’t separate from the world but embedded into a context. People use it, so I’d like to make an argument that dystopian surveillance tech, all the stuff that fuels the attention economy and industrialized warfare machinery are something alike evil. And AI, well that’s designed to reproduce stereotypes and bias. It’s almost entirely controlled by tech bros. And designed to their specifications, so it’s kinda laid out to do what they want, so there’s intent baked in in a form. And we have the environment footprint so it’d really need to perform better or it’s a net-negative outcome, no matter what philosophical arguments we have.
I don’t disagree, but my point is.
It’s a category error, LLMs are text prediction engines. There is nothing behind the curtain, they can’t by evil, because that implies understanding and intent.
LLMs are evil in the way that earthquakes are evil, it is pure anthropomorphism, and it’s taking the focus from were the real issues are.
Don’t get sucked into blaming the hammer, when the one swinging it it right there.
Yeah, I’m not entirely convinced, yet. Sure you’re right and all. But even with the hammer analogy… There’s the one the carpenter uses. And then there’s the warhammer, specially designed to crush skulls. And you’re bound to have a bad time renovating your roof with that thing. So I think the designer already left some intent in the technology. And on the other side he have what things get used for. I think I’d be willing to attribute evilness to technology if it’s solely for evil purposes. Like certain kinds of land mines or devices to torture people.
Other than that, you’re certainly right. Most tech is at least dual-use. Or neutral and can be used for arbitrary good and evil tasks. I bet this is the case with AI, like other computer tech and automation. And with that it’s down to the humans who use it as a tool.
Question remains if that’s a useful argument in practice. When talking about dystopian science fiction, I’m always a bit unsure about the interplay of the people in power, the technology and society. Is it the people who use technology to oppress the other people? Or did the existence of the technology get them into the position of power in the first place, enabling the dystopian society?
And I’m sure we’ll give AI power to make decisions. We already let algorithms shape our information, lives and society. And oftentimes not for the better. And in my eyes it doesn’t really tell us a lot if we say the computer code has no understanding or intent. It’s still going to affect our lives, and it can have agency or autonomy if provided with the power to act in some form. It doesn’t need intent in the sense of a conscious human being for that. (But that’s not exclusive to AI. A more traditional business process might also decline someone’s loan or medical treatment and ruin their life. Or approve the military bomb someone.)
companiespeopleYeah this is a sore point. Whenever management says, “the company decided…” I really want to stop them and scream, “Who?! Who in the company decided?!”
These 9 companies have made billions by convincing thousands of other companies that their fun text generator can replace skilled workers.
Certainly they are neutral evil, no? Can’t have market monopoly without rules. Wouldn’t go as far as lawful evil because they ain’t afraid of breaking no law.
Historically it’s been the opposite. Bell was a market monopoly until rules were introduced specifically to break it up and thus its grip on telephony, for example.
Complex algorithms that follow rules they cannot deviate from = lawful.
Deliberately incorporating random factors into the algorithm so they don’t generate the same result every time = chaotic.
So I’d argue the LLMs themselves are neutral evil, presuming we allow objects to have alignments. In D&D, non-sapient animals have no alignments, because they don’t understand moral or ethical concepts, so that would argue for LLMs being unaligned and the alignment applying to their companies.
Could you argue a LLM is attuned to its corporate owner and shares its alignment? They’d definitely be cursed.
Then the companies would veer from lawful evil (Microsoft has been the archetype of abusing laws and regulations to its own benefit for decades) to chaotic evil (Grok has no rules, only the whims of its infantile tyrant).
Corporate owners are currently in the Find Out stage about how they have no control over their LLMs. And so no, they do not share an alignment with their corporate owners beyond fleeting coincidence.
Can I put hugging face as chaotic good?
No but you can list them as chaotic evil
