Maybe the fat fuck should stay in the White House. He can take his chances anywhere else.
My all time favorite. Sometimes I go with "is a bear Catholic?"
Oh yeah, the PDF standard runs over 1000 pages and is a minefield of potential vulnerabilities and implementation errors. I would not point any student in that direction.
Both queries look like they were accidentally submitted before the user was done typing. Since the one on the left begins with the grammatically incorrect "Is we getting," Google's AI response seems to assume the person is asking about one of Donald Trump's claims that he will be sending checks to US citizens for various reasons.
The one on the right is grammatically correct and the AI makes no assumptions and simply points out that the query is incomplete.
If this is real, it reinforces the stereotype that Trump supporters (or other individuals likely to believe his claims) are poorly educated.
Your honor, when I took off the parking brake, put the car in neutral, and gave it a big push down the hill, it was in the interest of science, to see how far it would go. I think a lot hinges on intent here. It certainly was not my intent to mow down five children at a bus stop.
Simple but beautiful artwork on this one, especially the "reincarnate-o" panel.
Still, it seems the best you can do is reduce the blast radius. The underlying problem - that LLMs are fundamentally untrustworthy - has no solution. I suspect we're going to see a million variations on this kind of attack.
Yeah, I haven't seen any of these soapy memes everyone is complaining about.
I mean, I haven't gone looking for them, but I haven't seen a single one.
Apparently it's a miracle cure that big pharma doesn't want you to know about!
Coincidentally just a few posts above this one in my feed: Online pee drinking communities are huge fans of getting chatbots to tell them that pee drinking is healthy
We just need a few trillion more dollars to get us over the line! It's going to be Utopia, I'm telling you!
It seems like OpenAI is fairly lax, bordering on negligent, when it comes to monitoring these agents. Then when things inevitably happen, they turn it into advertising saying "boy, these things are so ingenious we can't keep control of them!"
If they truly believe these agents have a 10% chance of eliminating the human race, how are they not monitoring them in a controlled environment? How did it take them 3 months to detect this and the many other attacks?
The media seems happy to take up the line that these agents are just supernaturally smart and there's no way to control them, but isn't the simpler answer that OpenAI is doing a crap job at containing and monitoring them? But of course they have to, there's no choice because something something China.
view more:
next ›