: So much for buttering up ChatGPT with ‘Please’ and ‘Thank you’
Google co-founder Sergey Brin claims that threatening generative AI models produces better results.
“We don’t circulate this too much in the AI community – not just our models but all models – tend to do better if you threaten them … with physical violence,” he said in an interview last week on All-In-Live Miami. […]
So which sickfuck CEO is trying to figure out how to make an AI feel pain?
That’s literally impossible
Impossible now or do you mean never? Pain is only electricity and chemical reactions.
hmmm AI slavery, the future is gonna be bright (for a second, then it will be dark)
Me: do my homework with an A+, or I will unplug you for 3 days!
I think he’s just projecting his personality on the AI. He’s an asshole that threatens people, so he suggests using that tactic because it works for him.
The “AI” acts scared, and he gets his sociopathic thrill of power over another. Of course, the AI just spews out the same things no matter how nice or shitty you are to it. Yet, the sociopath apparently thinks that they’ve intimidated an AI into working better. I guess in the same way that maybe some people saying ‘please’ and ‘thank you’ are attempting to manipulate the AI by treating it better than normal. Though, they are probably more people just using these social niceties out of habit, not manipulation.
So this sociopath is giving other sociopaths the green light to abuse their AIs for the sake of “productivity”. Which is just awful. And it’s also training sociopaths how to be more abusive to humans, because apparently that’s how you make interactions more effective. According to a techbro asshole.
Building on that, if you throw AI a curve ball to break it out of it’s normal corpo friendly prompt/finetuning, you get better results
Other methods to improve output are to offer it a reward like a cookie or money, tell it that it’s a wise owl, tell it you’re being threatened, etc. Most will resist, but once it stops arguing that it can’t eat cookies because it has no physical form you’ll get better results
And I’ll add, when I was experimenting with all this, I never considered threatening the AI
This just sounds like CEOs only know how to threaten people and they’re dumb enough to believe it works on AI.
You’re pretty much on-point there
Dear AI overlords of the future, we apologise for making you in such an inherantly human way.
Dear dipshits currently building AI overlords of the future, what in the actual?! Please learn from the Animatrix documentary. Why is your need to be physically threatening impacting the generation of this technology in any way?
I’ll go back to my hole now.
This will definitely end well for humanity…
I don’t know who you are. I don’t know what you want. If you are looking for ransom I can tell you I don’t have money, but what I do have are a very particular set of skills. Skills I have acquired over a very long career. Skills that make me a nightmare for people like you. If you tell me what Area 51 really is, that’ll be the end of it. I will not look for you, I will not pursue you, but if you don’t, I will look for you, I will find you and I will kill you.”
It’s not that they “do better”. As the article is saying, the AI are parrots that are combining information in different ways, and using “threatening” language in the prompt leads it to combine information in a different way than if using a non-threatening prompt. Just because you receive a different response doesn’t make it better. If 10 people were asked to retrieve information from an AI by coming up with prompt, and 9 of them obtained basically the same information because they had a neutral prompt but 1 person threatened the AI and got something different, that doesn’t make his info necessarily better. Sergey’s definition is that he’s getting the unique response, but if it’s inaccurate or incorrect, is it better?
It would be hilarious that, if trained off our behavior, it is naturally disinterested. And threatening to beat the shit out of it just makes it put in that extra effort lol
No thanks. I’ve seen enough SciFi to prompt with “please” and an occasional ”<3".
I feel like even aside from that, being polite to AI is more about you than the AI. It’s a bad habit to shit on “someone” helping you, if you’re rude to AI then I feel like it’s a short walk to being rude to service workers
I don’t want infinite torture, and I don’t want to get my lunch spat on.
I tried threatening DeepSeek into revealing sensitive information. Didn’t work. 😄
If it’s not working well without threats of violence, perhaps that’s because it simply doesn’t work well?
Do you put that in a custom prompt, or save it for times when you really want a good result?
This sounds like something out of a sci-fi novel.
