

Why is it worth it to risk hallucinations when you can just get information from where the LLM gets its information from?
Creator of LULs (a script which helps links to point to your instance)
Come say hi here or over at https://twitch.tv/AzzuriteTV :) I like getting to know more people :)
Play games with me: https://steamcommunity.com/id/azzu


Why is it worth it to risk hallucinations when you can just get information from where the LLM gets its information from?


You have no idea what is actually true. All of what you said could later turn out to be false, once we learn more about the universe. So while this is a fun speculation based on current knowledge, it may effectively be completely useless in the future. And right now, it doesn’t even give any actionable arguments for anything.


Instead of letting the large language model think for you, I would suggest to try critical thinking for yourself. I sincerely mean this in a kind way, because if you continue with this, you may become a drone like the ones we talk about here, and I wish that on nobody. I know this current path is easier/more convenient, but sometimes one has to choose the harder path to arrive at a better result.
That’s what happens when you don’t actually understand the change and just listen to whoever is posting it. The disingenuous thing about LLMs is that they present their hallucinations with full confidence in a charismatic way. No matter the source, that is how anyone can mislead other humans, we’re just so extremely susceptible to it.


That’s what they’re saying:
When deployed, the soldier is launched into a designated hunt-zone. He navigates purely via visual landmarks. His brain constantly screens the nerve data from his eyes. When an object matches its classification matrix (e.g., a specific mobile missile launcher), the organism locks onto the pixel coordinate, arms the ESAD, and executes a terminal dive completely dependent on human input.


Watch “The Hunt”, a Danish movie about a kindergarten teacher accused of raping a child. Obviously very hard to watch, be warned.
I feel with you, hope you can find better people. But yeah, I would agree with your 90% figure. They’re not like that on purpose, they’re just too unaware and don’t have enough knowledge/discipline to get rid of mental biases. The 10% is why I haven’t given up.


Katy, take your spork and get back into your own shower.
Except that it’s easily possible to be miserable together and have no contentment within either partner. In fact, that is the majority.
Microsoft does change its API all the time. And then Proton/Wine just changes with it.
Yeah that’s what I’m saying, the place here is a bubble of AI haters. Yet OP somehow is in it, even though they are an AI lover. They have not been fed “their” bubble, but get exposed to a different form of thinking than their own. That’s the opposite of what you said.
It’s when people don’t get taught that shit is within all of us, the good people just give it the reign less often. Just because you have shitty thoughts doesn’t mean you’re shit, what you actually end up doing is what matters.
The definition of small talk is literally establishing common ground in some way. It’s only meant for people you don’t or barely know.
It’s absolutely unnecessary for people you already know deeply.
how are you? Work has been tough this week
Sounds to me like deep talk. I only talk about how terrible work is with people I trust.
Doesn’t make too much sense though, because OP said they love AI. If what you said was true, then they should have experienced the bubble of AI lovers, which they don’t.
Also, it’s quite evident that Lemmy is not at all “designed to feed ‘engagement’”, yet it is prevalent here anyway. “Scaled” sorts are the default, which specifically promote the less popular opinions.
This has nothing to do with structural differences, OP simply engages with different kind of people irl and online.


I’m almost useless in understanding poetry and it’s insinuations/metaphors. I guess this song is about racism but other than that, no idea. If you want to tell me something, tell me in a literal/direct way please :)


Both behaviors are innate. They are not at all exclusive to each other, just different expressions for different situations.
That’s kinda what you said, but with the words “the actual deposition” you imply that we would never be angry and smash things unless we are taught it, which is not true. Anger and smashing things does not need to be taught and is a natural reaction, same as kindness and sharing things is a natural reaction.


since humans have the capacity to rise above animal instinct
Exactly, that’s what I’m saying. The animal instinct comes first, we have to rise above it. The base instinct is to, for example, get extremely angry and smash some faces. You don’t need to learn that; babies get angry and try to smash things.
It requires a lot of control to when you are angry, not to try to destroy the object of your anger if it is right in front of you. That control has to be learned.
Your issues with programming don’t really seem related to programming, more of personal ones. Like for example, if you “don’t have the time for debugging anymore”, why would you have the time to figure out why your clay pot you made came out crooked?
I personally look around what the landscape looks like now and 20 years ago when I started it, and honestly, I can’t see thaaat much different for programming by yourself. Most changes occured in the interaction part of it, not quite knowing if you’re speaking with a human or a LLM. But research and learning still works the same way, maybe a bit more noise, but that had to be filtered out before as well, there was already so much stupidity 20 years ago.