The thing that bothers me about LLMs is that people will acknowledge the hallucinations and lies LLMs spit out when their discussing information the user is familiar with.
But that same person will somehow trust an LLM as an authority on subjects to which they’re not familiar. Especially on subjects that are on the edges or even outside human knowledge.
Sure I don’t listen when it tells me to make pizza with glue, but it’s ideas about Hawking radiation are going to change the field.
The same used to be said of newspapers (and still ought to be). That is, it’s funny how accurate and informative they appear to be until the topic changes to something about which you have intimate knowledge.
The logical leap to generalise from that is impossible for far too many people and is also an easy trap for those who can make it.
This is literally the Dunning-Kruger effect in action - people can’t evaluate the quality of AI responses in domains where they lack the knowledge to spot the bs.
They don’t realize that the chatbot’s “ideas” about hawking radiation were also just posted by a crank on Reddit.
Anyone consulting LLMs for anything already had a preexisting psychological condition called Stupid as Fuck.
You get totally different answers to “is X healthy” vs “is X unhealthy”
But yeah, if ChatGPT tells you to order restricted substances on the internet, probably don’t do that
if ChatGPT tells you to order restricted substances on the internet, probably don’t do that
so what am I supposed to do with my 3.8 tons of gunpowder now??
Mix it with water and finger paint on a 12’ by 12’ canvas
Then do a big art attack
Social networks are all about making and keeping people angry to make people come back. AI is all about brown-nozing and giving any information with absolute confidence to keep people coming back.
Social networks are all about making and keeping people angry to make people come back
And I hate that it works. Every time I hop over to Reddit out of habit, the posts and comments are so irritating that it makes me want to refute them. But then I come to my senses because I remember that I’ve been banned and can’t comment. Also, I don’t want to contribute to that site in any way.
I’m slowly, but surely, weaning off that site.
After years of bullshit, corruption and nepotism, we as a society (or a critical mass of it) accepted that lies and bullshit is a part of life.
I really think that’s what is going on here, we filled our reality with contradictions and things that drive us crazy, now a large percentage of the population are okay listening to inefficient guessing machines.
Seriously, the fact that hallucinations didn’t kill the hype is, imo, a hallmark of being in a post truth era.
This is not the mindset that made computers and the Internet. Feels more like late stage Rome.
Late stage western Rome was kinda fine, they even had Christianity kinda receding (Goths and Vandals were far more pious Christians, while in Rome there was plenty of support for restoring the old state religion). What really killed them was intrigue and inability to hold the territory militarily.
Late stage eastern Rome can be divided into a few different long periods. If we mean after being crusaded, then it was shattered and trying to recombine into something useful, eventually finished by Ottomans. If we mean before - well, there was the long period of Seljuk conquest of Anatolia, initially a result of the empire being weakened by conquests against its buffer states and allies which didn’t really have to be conquered. But in any case it’s probable that the same thing would happen, the Seljuk armies of the initial invasion time are described as quite numerous, and they had a few military innovations east Romans didn’t have.
I think he meant later stage westen Rome before the Gothic/Vandals/Longobardic tribes conquered it. A paper tiger entirely convinced of is own superiority whilst their military power and world control slowly dwindled.
So basically what you summed up in a sentence. The eastern Roman and Seljuks need not have been brought into this, I think.
Well, lacking troops due to big demographic and agricultural changes outside your control is not the same as being a paper tiger. Their internal turmoil coincided with a military threat.
Also western thalassocracies have been accused of being largely paper tigers a few times in their history, yet from Ottomans to Napoleon to German Empire to USSR no contenders managed to prove it.
OK, what matters is the OP’s point on lies. I agree with that entirely. Lies brought the USSR down, and the same has unfortunately caught up with its Cold War adversaries. It just became visible a bit earlier that this state is not going into space colonizing everything and is not building a classless society with everyone equal and free. But 40 years later, I think, we can confidently say that neither is USA fulfilling its own “cold war promise”. 30 years ago I suspect even many Americans sincerely believed that it is. Everyone thought it was one side’s defeat, while in fact it was a draw with both sides failing.
Oh fun! When Americans lose their healthcare I expect these kind of stories to rise.
“Lose their healthcare?”
What healthcare
I told my wife about this story and she told me that everything on the internet should be taken with a grain of salt. I told her this guy took a few too many grains.
He did his own research! /s
A man uses the internet to poison himself. The story as old as time But if we stick the AI in the title, we can get some sweet clicks out if it.
This is not just “using the internet,” though. AI use ≠ finding some conspiracy-fueled rant on some long-forgotten message board. ChatGPT does not scour the internet or have any sort of meaningful sanity checks on the pattern of words it generates. It doesn’t “know” what it’s saying, nor does it “care.”
If he had done even the most basic of generic internet searches, he would have discovered the DASH diet.
The inclusion of this goober’s use of AI is yet another example why using what is essentially a reinforcement and pattern-generation engine is one of the dumbest things a person can do. It doesn’t seem to matter how many experts remind people of its limitations, so all that remains is pointing out every time somebody does something stupid, so people can at least get a reminder that the other end of the conversation is dumber than they are and only an illusion of intelligence.
It doesn’t seem to matter how many experts remind people…
But this is not new. People have been drinking bleach and giving themselves cyanide poisoning long before the spread of LLM chatbots. Some dare to call it “doing their own research”
Yup reads exactly the same.
Finally evolution back in effect
These days we call it bruhmism.
“For 3 months, he had replaced sodium chloride with sodium bromide obtained from the internet after consultation with ChatGPT.”
I didn’t want to click. But I did so here you go.
Fun fact, mountain dew used to have bromine in it
It sounds more like a crazy person did a crazy thing and happened to use AI.
