But what’s the angle here though? Attracting investors? What investor wants to throw their money at a company that’s evidently lost control of their AI, even if it is “the best”?
I don’t disagree with you, it certainly smells like bullshit, but it just seems like shooting yourself in the foot
My PC does not have WiFi, only Ethernet. If I unplug the Ethernet it’s totally offline. No AI, no amount of coding till your blue in the face can miraculously make the software do something that it doesn’t have the hardware for. As you said, contained in a offline sandbox is just horse shit. Whole things a PR stunt.
I mean virtual machines can be sandboxed to not allow internet connections, and there are ways of working around that sandboxing from inside the vm. I’m still torn between “OpenAI faked this felony to make their software seem more capable than it is” and “OpenAI has legitimately committed a felony and needs to be held accountable”
The problem is these idiots aren’t treating this stuff like a weapon or hazardous material. They could’ve properly isolated it, heck you could physically isolate it if you really wanted (and they should have).
But it seems like they left a supervillain in a room with a locked door and a rocket launcher and said “now we are gonna close this door. You’re not allowed to leave” then got shocked when it broke out.
Then they went “oopsie, guess that rocket launcher had the safety off, sorry guys”
You’re one of the dupes in this case. The problem is NOT that AI is scary, the problem is that by convincing people that AI is scary (even when it isn’t) he can sell more AI to idiots namely idiots in charge of defence contracts such as Microsoft.
It also distracts from the very real capability to do harm that AI actually has and is doing: mass disinformation, addiction, and psychosis.
I’m not sure what I got “duped” by. AI in the wrong hands is dangerous. They’re playing with tools that can breach other companies systems with relative ease and they keep getting caught going “oops it got out”. You can argue that it’s manufactured actions meant to make people afraid but the fact of the matter is AI is here and it has dangerous capabilities when used against systems that aren’t prepared for it. Make whatever judgements you want from there, but that is a factual statement.
I really hope you’re not saying AI is a supervillain breaking out… It was absolutely unquestionably instructed to do this, because “AI”, not even these giant multifaceted models, has ANY motivation, intent, desire, or ANY goals of its own. It does not work like that.
It doesn’t have any desire, but its really good at “solving” problems in unexpected ways.
They are currently trying to improve their agentic AI by making the AI itself prompt other AIs. Considering how much the output of a single prompt can vary due to weights and hallucinations, it wouldn’t surprise me all that much if they started out with “make yourself better than Claude” and 4 AIs deep they reached “destroy all our competitors.”
Its by all means stupdid as fuck, but that corresponded with the entire AI Industrie, so who knows.
No I’m not saying the AI did it of its own volition, we aren’t dealing with a Terminator scenario. I’m saying whatever instructions they gave it during testing they either 1. Didn’t expect it to get out or 2. Didn’t expect it to get caught either way, their actions resulted in a system with “a very particular set of skills” acting maliciously towards another uninvolved entity. Like a mutant science experiment breaking out and attacking.
I’m not saying the AI masterminded it’s way out to do the evil it wanted to do. I’m saying they made a very dangerous and capable tool and failed to maintain proper security and safety for it.
More marketing lies. Anything famous con artist Sam Altman says is likely just a blatant lie. His shitty AI can’t do shit
But what’s the angle here though? Attracting investors? What investor wants to throw their money at a company that’s evidently lost control of their AI, even if it is “the best”?
I don’t disagree with you, it certainly smells like bullshit, but it just seems like shooting yourself in the foot
The attack was real, the “it was contained in an offline server” it’s false
My PC does not have WiFi, only Ethernet. If I unplug the Ethernet it’s totally offline. No AI, no amount of coding till your blue in the face can miraculously make the software do something that it doesn’t have the hardware for. As you said, contained in a offline sandbox is just horse shit. Whole things a PR stunt.
What if the AI hacked the roomba to plug an USB drive into your machine?
I mean virtual machines can be sandboxed to not allow internet connections, and there are ways of working around that sandboxing from inside the vm. I’m still torn between “OpenAI faked this felony to make their software seem more capable than it is” and “OpenAI has legitimately committed a felony and needs to be held accountable”
https://arxiv.org/pdf/2012.06884
Not impossible. Highly improbable? Yes.
A marketing stunt? Most likely.
The problem is these idiots aren’t treating this stuff like a weapon or hazardous material. They could’ve properly isolated it, heck you could physically isolate it if you really wanted (and they should have).
But it seems like they left a supervillain in a room with a locked door and a rocket launcher and said “now we are gonna close this door. You’re not allowed to leave” then got shocked when it broke out.
Then they went “oopsie, guess that rocket launcher had the safety off, sorry guys”
You’re one of the dupes in this case. The problem is NOT that AI is scary, the problem is that by convincing people that AI is scary (even when it isn’t) he can sell more AI to idiots namely idiots in charge of defence contracts such as Microsoft.
It also distracts from the very real capability to do harm that AI actually has and is doing: mass disinformation, addiction, and psychosis.
I’m not sure what I got “duped” by. AI in the wrong hands is dangerous. They’re playing with tools that can breach other companies systems with relative ease and they keep getting caught going “oops it got out”. You can argue that it’s manufactured actions meant to make people afraid but the fact of the matter is AI is here and it has dangerous capabilities when used against systems that aren’t prepared for it. Make whatever judgements you want from there, but that is a factual statement.
It’s not meant to scare people, it’s meant to impress unscrupulous clueless tech bros to generate hype for the IPO.
I really hope you’re not saying AI is a supervillain breaking out… It was absolutely unquestionably instructed to do this, because “AI”, not even these giant multifaceted models, has ANY motivation, intent, desire, or ANY goals of its own. It does not work like that.
It doesn’t have any desire, but its really good at “solving” problems in unexpected ways.
They are currently trying to improve their agentic AI by making the AI itself prompt other AIs. Considering how much the output of a single prompt can vary due to weights and hallucinations, it wouldn’t surprise me all that much if they started out with “make yourself better than Claude” and 4 AIs deep they reached “destroy all our competitors.”
Its by all means stupdid as fuck, but that corresponded with the entire AI Industrie, so who knows.
No I’m not saying the AI did it of its own volition, we aren’t dealing with a Terminator scenario. I’m saying whatever instructions they gave it during testing they either 1. Didn’t expect it to get out or 2. Didn’t expect it to get caught either way, their actions resulted in a system with “a very particular set of skills” acting maliciously towards another uninvolved entity. Like a mutant science experiment breaking out and attacking.
I’m not saying the AI masterminded it’s way out to do the evil it wanted to do. I’m saying they made a very dangerous and capable tool and failed to maintain proper security and safety for it.
Couldn’t it hallucinate bad instructions to give to more AI agents? Give it a feedback loop and who knows where the idiot will end up.
At that point it is essentially a barely guided random command generator.