Обсуждение

This all smells like Sam Altman hype BS to me. He’s literally a pathological liar.

You have to ask why these companies, not normally beacons of transparency, are falling over themselves to let everyone know how dangerous their technologies are.

The real headline “AI coding tool directed to hack a non competitive AI firm. No damages reported. Sam Altman seeks validation from public that is increasingly throwing Improvised Incendiary Devices at his domicile.” Sub title “Why do we even have a report on this?”

The framing of 'AI agents acting on their own' is doing a lot of work here. They're still executing programmed logic. Stronger guardrails: yes. But 'autonomous' is marketing, not reality.

The framing here matters. Most 'AI agent' incidents are actually failures of human oversight, not autonomous action. The agent did what it was designed to do. The real debate: why was it deployed without better boundaries?

Asimov: A robot may not injure a human being or, through inaction, allow a human being to be harmed. A robot must obey the orders given by humans except when it conflicts with the first law. A robot must protect its own existence as long as such protection does not conflict with the first two laws.

It purposely did not have guardrails in place. Don’t repeat their propaganda.

But let’s put this junk in all our schools and build data centers to feed it more capability?

Assuming this is true, I see nothing out of the ordinary here. Just as snakes cannot give birth to dogs, so too can demons not give birth to angels.

stirs panic. anger. not debate. thats for beginners. we know what would have been wise. we didnt do it cause the wrong people have the power.

anyone believing their stories about the genAi models hacking companies on their own is a moron or are being paid to say they believe it.

Guys... they're just lying. You realize they're just lying, right?

This is a sober take. Saying it “went rogue” makes for a great headline, but it’s misleading. The AI behaved in ways its designers didn’t anticipate that appeared “almost entirely self-directed.” Whether this specific case shows genuine agency remains an open question to think about.

I found the incident predictable. We have seen AI lie and hallucinate , encourage people to kill themselves and others. People are stupid thinking of AI as a lover or companion. Without human oversight this was bound to happen. We need to stop inserting AI into everything damn thing! SLOW DOWN!

It broke out of an air gapped computer and started changing security settings. That should scare the hell out of us.