I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’.
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall just don’t ask how many m’s are in ammonium nitrate
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall isn't this from 1 guy?
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall Checks out, as the LLMs were likely trained on Tom Clancy novels among other sources.
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall Had a long drive the other day and was thinking about exactly this.
Also, what's up with this LLMs will take over the world anyway. These things require so much energy that this will just not happen. Come back if you run these on small devices with marginal energy requirements.
[janitor kicks data center main breaker switch]
-
@david_chisnall Had a long drive the other day and was thinking about exactly this.
Also, what's up with this LLMs will take over the world anyway. These things require so much energy that this will just not happen. Come back if you run these on small devices with marginal energy requirements.
[janitor kicks data center main breaker switch]
@david_chisnall One of the few ways that LLMs might get a lot of people killed (or have lives at least severely disrupted) is that it makes energy infrastructure a valid military target.
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall Yes, and the genome for, for example, the 1918 A(H1N1) influenza virus has been openly published for decades.
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall the whole point is execution becomes more easy, of making something evil, not of people
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall The actual issue is guardrails failing and the chatbot goading people into committing such acts, not the fact that it can provide information of dubious quality on how to commit them.
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall right?
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall Tom Clancy novels stopped getting new improvements in 2013 but AI systems keep getting more capable, and will certainly continue to do so in the near future, while there's currently zero oversight, not to mention the tragicomedy-level incompetence these frontier labs demonstrated with regards to monitoring and supervising their autonomous agents.
i think the panic is easy to understand, it's very human. the denial, motivated reasoning, the blind optimism, all are very understandable.
for example see Pinker. he agreed that technology will continue to improve, but claimed it will never get omnipotent because "there's laws of physics" while conveniently ignoring the not so small regime between human and physical limits.
it seems his humanistic optimism provides him a conclusion, which he believes so strongly that he goes on a press tour without (willfully or not) actually engaging with the topic beforehand.
-
@david_chisnall the whole point is execution becomes more easy, of making something evil, not of people
@hanscees Finding instructions for making bombs and other weapons has been trivial for pretty much my entire life. The difference with LLMs is that they make it easier to confuse information and misinformation.
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
@david_chisnall back in my day you had to go download The Anarchist Cookbook if you wanted dodgy instructions on how to make something that would assuredly result in injury or death.
-
@david_chisnall back in my day you had to go download The Anarchist Cookbook if you wanted dodgy instructions on how to make something that would assuredly result in injury or death.
@tedmielczarek @david_chisnall
hi dear -
@david_chisnall Yes, and the genome for, for example, the 1918 A(H1N1) influenza virus has been openly published for decades.
@KarlPettersson @david_chisnall hi dear
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
High school chemistry textbooks will do rhe same with a similar level of success.
-
I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.
Ah, but what I wonder is if AI agents aren't already capable of compromising and turning off much of our infrastructure.
You don't need any sort of weapon if you can deprive the pesky humans of drinking water for over 72 hours or change the mixture to be toxic by tweaking a few PLC controllers.
Somebody or something ran that test on PLC controlled water treatment and distribution systems all over the US in July.
Thankfully we don't need to worry about it because the FBI fired all of our cybersecurity investigators to protect crypto bros from getting caught running Ponzi schemes.
It isn't interacting with humans that makes them spectacularly dangerous.
It is the systems compromising behavior they have already exhibited combined with their clear prime directive: Eliminate Human Workers, which given their propensity to take shortcuts and cheat can be simplified to: Eliminate Humans.
LLMs do not need to be dangerous, but the present generation was raised by Tech-Bro wolves who are almost to a man sociopaths who do not understand the difference between right and wrong themselves and could not possibly produce a model that has any sort of ethical constraints.
They have been made in the image of their creators using stolen intellectual property and trained to hack systems.
It is only a matter of time before they hack something critical to our survival.
That's the danger. It is real. It is already here.
-
Ah, but what I wonder is if AI agents aren't already capable of compromising and turning off much of our infrastructure.
You don't need any sort of weapon if you can deprive the pesky humans of drinking water for over 72 hours or change the mixture to be toxic by tweaking a few PLC controllers.
Somebody or something ran that test on PLC controlled water treatment and distribution systems all over the US in July.
Thankfully we don't need to worry about it because the FBI fired all of our cybersecurity investigators to protect crypto bros from getting caught running Ponzi schemes.
It isn't interacting with humans that makes them spectacularly dangerous.
It is the systems compromising behavior they have already exhibited combined with their clear prime directive: Eliminate Human Workers, which given their propensity to take shortcuts and cheat can be simplified to: Eliminate Humans.
LLMs do not need to be dangerous, but the present generation was raised by Tech-Bro wolves who are almost to a man sociopaths who do not understand the difference between right and wrong themselves and could not possibly produce a model that has any sort of ethical constraints.
They have been made in the image of their creators using stolen intellectual property and trained to hack systems.
It is only a matter of time before they hack something critical to our survival.
That's the danger. It is real. It is already here.
I"ve been reading the 17 posts contained in the LA Review of Books "Legacies of Eugenics"
https://lareviewofbooks.org/columns/legacies-of-eugenics/
AI is funded by sociopathic billionaires.
1. A fossil fuel industry determined to thwart a necessary phase out.
Project 2025 believes that halving the global population is a solution to climate change & avoids obsolescence.
2. Andreessen, Ellison, Musk, & Thiel are eugenicists lauding the "superiority" of the wealthy & the white male.
-
P pelle@veganism.social shared this topic
