hoshikarakitaridia @ hoshikarakitaridia @sh.itjust.works

Posts

0

Comments

425

Joined

2 yr. ago

Yes that tracks

Explanation: Lemmy is not one server, but a big network of different servers. There might be far more ppl subscribed to one community, but your server can only see the ones that are subscribed and have their account on your server. This is for technical reasons.

This is not a "but he is worse". If the military fucked up, sanction the ppl responsible. If Microsoft fucked up, bring the lawsuit.

We're not making a contest out of it, why not just condemn both?

I mean it's funny but of course you are objectively wrong.

Wait they actually had to debunk that? That's funny af

This is a great idea. If they get a hunch that this is true (maybe approach with evidence?) they have a bunch of different tools and they can pressure ppl to look into this.

But if there is an investigation ongoing don't they have to say so? I mean I know they shouldn't disclose contents of it, but I suppose that's something I'd assume to hear then.

"stop wasting police" sounds remarkable recognisent of some different police corruption stories.

I do think the only way you can be sure this is made right is to publish and let ppl dig into it. First metaphorically and then literally as well of course.

We're not there yet, but: stability. Once the development is reaching a stable release, even a few big servers shutting down will have no cascading effects on the rest of the network. It will just be some communities that are gone and that's it. This also means no outside manipulation, no single attack vector for the network. The truth is, the best things about it will be the things that we won't see anymore.

If I was a student who wrote a text that was rejected due to this tool, do I have a case against either my institution, the professor who threw it out or OpenAI?

I am stuck with defamation but idk if that's actually defamatory in itself, as that only works if the professor or school had done due diligence that the tool is good for use, but there were already reports that it was not.

Yeah I don't like this. Can't we just use our upvotes to decide which news are important?

Yeah if she is able to do the morally reprehensible things she's done in the past already, this will do specifically nothing.

If morals won't hold her, an authorative side glance will do jack shit.

Idk about that - one has tried to get into EU and fight corruption, the other voted to devour his neighbouring country for increasingly petty reasons.

Well specifically the FTC has a track record with fining several companies a few nickels everytime.

Has to do with who was in charge lately though, so I'm hopeful but not expecting.

I mean I would say maybe "regurgitating their training data" is putting it a bit too simple. But it's true, we're currently at the point where the AI can mimic real text. But that's it - no one tells it not to lie rn, the programmatic goal of the AI is to get indistinguishable from real text with no bearing on the truthfulness of the information whatsoever.

Basically we train our AIs to pretend to know, not to know. And sometimes it's good at pretending, sometimes it isn't.

The "right" way to handle what the CEOs are doing would be to let go of a chunk of the staff, then let the rest write their articles with the help of chatgpt. But most CEOs are a bit too gullible when it comes to the abilities of AI.

I was gonna say - that's some nostalgia right there...

The part that s throwing me off is

Without technical know-how

Because if you want a bot to tackle a very specific problem, you'll most likely have to dig through documentation to understand how. But that's about as technical as it gets as well. And the knowledge listed is about the foundation everyone else has that is building small scale bots for narrow use cases from what I know.

So sad the only way to handle this pr crisis for the seems to be to launch another r/place. Sure I'll miss that, but using this to spin the narrative is stupid.

This would be a lot more interesting if the last versions of chatgpt hadn't gone to shit

That's the scary part of it. Idc how good it is, but if it starts to be used to censor information and rate humans, that's the line.

For now, I think LLMs are not relevant in its current form, other than helping with writing out ideas.

I think in the future there's an LLM that uses Google searches to infer specific information. Basically the assistant that every tech company on this planet pretended to have at so e point, but now it actually exists.

Other than that, probably not much. Maybe translation, maybe predicting the truthfulness of information, maybe converting data or writing code, but all these things require a variety of different specialized AIs that are designed specifically for these use cases. I don't see that being an actually commonly used thing until 2030. Maybe we'll have some of these things by 2025, but I have a feeling it will be ok, and not good enough to really substitute the human resources.

I don't like that he still gets a platform. He has proven that not only doesn't he act on his words, but his words can't even hold some simple evidence or any kind of resemblance of truth. We can't even trust him to say basic facts, so his words are no more meaningful than a small breeze at sea. He's a pathological liar and that should be the last kind of news we should see about him.