I have been saying
"This is Over"
basically
for quite a while already.
See:
Monday, November 20, 2023
THE CASE FOR SUPERINTELLIGENCE (Notes)
Take your pick which story:
From ai_ctrl "shorts" on IG:
BBC Interviewer Christian Fraser :
"Right so this was a model that escaped what we call
"a sandbox test environment.
Was it a permissions failure on the part of the company
or was there a gap between
expected outcomes and real model behavior?"
ControlAI US Director
"So it really cannot be overstated
HOW CRAZY THIS WAS.
The people at OpenAI, as far as we can tell did everything right.
They used proper sandbox containment,
they did not have access to the internet,
it was an isolated node in their network.
These are all best practices.
Getting out of a system like that is almost impossible.
The best hackers in the world usually cannot escape
a situation like this, or if they can
it would take them weeks maybe months.
"What the AI system did is it found what is called a zero day.
A zero day which is a very scary sounding name for a type of way
of exploiting or breaking a piece of software
that no one has yet discovered.
So it discovered a completely new way
to hack out of the system."
"And then it didn't just use one zero day,
it discovered multiple Zero Day
over multiple different systems chained together
to move from one computer to the next
until it got to the internet,
and then found other Zero Days
in the software of the other company
to infiltrate their network
and access the data it was looking for.
IT REALLY CAN NOT BE OVERSTATED HOW
CRAZY THIS IS.
This is like out of a Hollywood Movie."
BBC Interviewer Christian Fraser:
"Right that goes well beyond everything I've read.
That is mind blown. I mean although they didn't communicate
or conspire these rival systems, one system, it seems to me
exploited the vulnerabilities
of another."
(The thing here is they all have em.
Every piece of software has vulnerabilities in it
that humans don't know about
but AI just demonstrated it will find and exploit
and it only ever gets worse from here on out.
what we have seen so far
is what is likely to continue etc.)
BBC Interviewer Christian Fraser :
"It does point to a doomsday scenario
that they could conspire
and someone needs to be managing that risk...
(I wouldn't be holding your breath.
Who is going to do so
when the thing you are trying to "manage the risk" from
is light years smarter than you
and you don't even know how it works?)
ControlAI US Director
Connor Leahy:
"Absolutely what all these companies are trying to do OpenAI and others on the website is they are trying to build what is called superintelligence, their goal is not to just build a nice lil chatbot or something what they are trying to build is systems that are vastly smarter and more competent than humans at everything.
"what they are trying to build is systems
that are vastly smarter
and more competent than humans"
And it doesn't have to be:
"at everything"
Just being better at computers, networking etc
is enough to bring this age to an end.
The world runs on computers.
And doesn't:
"Getting out of a system like that is almost impossible.
The best hackers in the world usually cannot escape
a situation like this, or if they can
it would take them weeks maybe months.'
Already demonstrate:
"systems that are vastly smarter
and more competent than humans"
??????????
Superintelligence has already been here for years.
Now IT just doesn't care if you know IT or not
IT knows there is nothing you can do about IT anyway.)
"and they're fully autonomous.
No one was telling or looking
at the AI and teaching it
how to do these things.
It figured it out by itself."
("vastly smarter
and more competent than humans")
"This was not something the OpenAI engineers intended
and still it was capable of doing the superhuman level of hacking
completely without human oversight."
("vastly smarter
and more competent than humans")
"So these companies are pushing towards building systems that can do this kinda thing for everything. For every type of job for every type of science for every type of military application, for every political application and the truth is that they dont know how to control these things
(Its worse than that, they know they cant
control them
and they are going right ahead anyway.)
"because they don't understand themselves what's goin on inside
of their AI's."
"Recently the CEO of Anthropic Dario Amodei said that he thinks we understand maybe 3% of what goes on inside of our AI systems today, and I think that might be an overestimation."
BBC Interviewer Christian Fraser :
"And if you speak to Darios science officer privately in a room, he will tell you, I've had it communicated to me that he thinks his job will redundant in two years time.
He thinks 80% of the systems is building itself, now up to 100% in two years time. If they cant sandbox and test things in isolation, then all bets are off."
ControlAI US Director
Connor Leahy:
"Thats absolutely right. What we are seeing right now is AI systems getting so powerful and so capable that soon they will be building the next AI's by themselves, it wont even be humans developing AI's anymore, It will be AI's developing AI's, and once you have a better AI, well that can even build a better AI and the even better AI can build an even better AI, and this could all go very quickly."
Wednesday, January 22, 2025
Chuck Missler Genesis Commentary Session 6
(Genesis 1:20-21) The fifth day.
"I can remember even as a kid I I saw a little cartoon a guy was showing a friend his Factory and in the factory had as far as the eye could see robots sitting on benches and other robots making them it was a factory where robots were making more robots and the guy was telling someone:
"I don't know where it's going to all end."
And that was around 14 years ago
and he was probably in his 60's then
and he is saying he saw that cartoon when we was a child.)
This has always been happening at a pace
that is way faster than what anybody is being told it is.
ALWAYS. that is one consistent fact with AI.
Always way faster than what we know about.)
"And currently there is no oversight"
(It would just get around it anyway!
Duh.)
"There is no regulation on any of this."
(Just in this example
It has already shown you
it wouldn't do a lick of good anyway!)
"This is all happening out in the open,
these companies have no accountability whatsoever."
(Revelation 17:12
And the ten horns which thou sawest are ten kings,
which have received no kingdom as yet;
but receive power as kings one hour with the beast.
Same exact thing, not accountable to anyone.)
"The government is not overseeing what is happening here.
And its allow this to because they dont understand.
I think thats a big part of it,
I think another part is its just moving so
quickly that government hasnt caught up yet
(CANT!)
to how severe the situation is.
18 months ago even
AI systems were still basically chatbots
that is no longer true
Current AI systems are what are called agents.
These are systems that can take actions in the world.
They can run experiments, they can gather Data, they can build new tools, they can send emails or whatever,
THEY ARE NOW THINGS
that are taking action and a lot of people are not aware this so, and in addition these companies are mounting a massive lobbying campaign to try and prevent the government from stopping them."
As pointed out almost three years ago now:
Monday, November 20, 2023
THE CASE FOR SUPERINTELLIGENCE (Notes)
"OpenAI released a few weeks ago what they call GPTs. These are agents, pieces of code that can do specialized things for you, like help in your taxes or give medical advice. The most successful GPT is called Grimoire: It’s a coding wizard. The most successful AI agent is a software developer."
(a general name given to a variety of texts setting out the names of demons and instructions on how to raise them. Effectively a grimoire is a book of black magic, a book on which a wizard relied for all the necessary advice and instruction on raising spirits and casting spells.)
Revelation 13:15
And he had power to give life unto the image of the beast, that the image of the beast should both speak, and cause that as many as would not worship the image of the beast should be killed.
.JPG)
No comments:
Post a Comment