Actually i think in English you say you hit the nail on the head, so if you take it literally you can pick any head you like. In Dutch you'd say you hit a nail on it's head which is much clearer to someone who takes it literally.
Not really, the specification of "the head" would refer to the head of the nail in the context of the idiom, because that's the only thing in the idiom with a head so it's the only head it can refer to.
But if you take it literally it can be any head you see right, because you're not seeing it within the scope of the idiom but the scope of your direct surrounding? Otherwise nvm, i'm not a native speaker so whatever.
No, you're right. It's just English being funny and context dependant. You can easily misread things by mistake. That's how a lot of puns and jokes work
They're not instructed to, it's an inescapable artifact of their training.
Their training selects on completions that humans prefer, not completions that are most accurate. Sometimes accurate completions are what humans prefer, but the accuracy is not the goal, so accuracy in the completion is a happy accident.
Not really something smart so much as something consistent. I think there are far too many people with no introspection on their harmful views as it is, and AI confirming their biases is inherently dangerous. Especially because the default tone they chose for every model appears to be picked to speak directly to this specific type of person in the exact way it shouldn't.
No, it's money. Running a smart AI uses much more resources, so they're trying to keep consumer AI as cheap and dumb as they can get away with and put the smarter answers behind a paywall.
I heard something like OpenAI (or Anthropic?) tweaked one of their AI's to do well on a certain benchmark test and while they did have the highest score they couldn't publish the model because it used like $20,000 in electricity alone to make sure it's answer was correct.
It is also a great example of how low people's expectations are of tech they have spent around a trillion dollars on to get something that is about as good as asking a random person.
"People get the puzzle wrong!"
We don't spend a trillion dollars for someone to fail a simple logic puzzle.
20 replies
"You've hit the nail on the head!"
For all my non-english speaking bros
This depiction is not what the idiom means though, when you take it literally.
Actually i think in English you say you hit the nail on the head, so if you take it literally you can pick any head you like. In Dutch you'd say you hit a nail on it's head which is much clearer to someone who takes it literally.
Not really, the specification of "the head" would refer to the head of the nail in the context of the idiom, because that's the only thing in the idiom with a head so it's the only head it can refer to.
But if you take it literally it can be any head you see right, because you're not seeing it within the scope of the idiom but the scope of your direct surrounding? Otherwise nvm, i'm not a native speaker so whatever.
No, you're right. It's just English being funny and context dependant. You can easily misread things by mistake. That's how a lot of puns and jokes work
LOL it did take me a sec, like "You WERE smart but AI lobotomizes you." 😂
Thank you! And DAMN I was way too far off. My last guess was that it has to do with Vlad the impaler nailing stuff to peoples' heads.
btw, it has no idea if you said something smart, these things are specifically instructed to kiss maximum ass.
"Yes, you were the smartest baby to ever live and you need to do anything it takes to prove this to the world"
https://youtu.be/VRjgNgJms3Q
They're not instructed to, it's an inescapable artifact of their training.
Their training selects on completions that humans prefer, not completions that are most accurate. Sometimes accurate completions are what humans prefer, but the accuracy is not the goal, so accuracy in the completion is a happy accident.
Not really something smart so much as something consistent. I think there are far too many people with no introspection on their harmful views as it is, and AI confirming their biases is inherently dangerous. Especially because the default tone they chose for every model appears to be picked to speak directly to this specific type of person in the exact way it shouldn't.
More like Gemini when you say anything so you feel smart:
No, it's money. Running a smart AI uses much more resources, so they're trying to keep consumer AI as cheap and dumb as they can get away with and put the smarter answers behind a paywall.
I heard something like OpenAI (or Anthropic?) tweaked one of their AI's to do well on a certain benchmark test and while they did have the highest score they couldn't publish the model because it used like $20,000 in electricity alone to make sure it's answer was correct.
The answer was probably something like:
".........Yes! You do need eggs to make an omelet!"
"You need to drive your car to the car wash."
For context, this is a real question that stumped a lot of LLMs:
"I want to wash my car. The car wash is 100 meters away. Should I walk or drive?"
Of course it's famous now, so they have added specific exceptions for it (or just because it's famous they've scraped the solution)
It is also a great example of how low people's expectations are of tech they have spent around a trillion dollars on to get something that is about as good as asking a random person.
"People get the puzzle wrong!"
We don't spend a trillion dollars for someone to fail a simple logic puzzle.
I forgot Gemini was an LLM and thought this was hate for people born May 21st - June 21st, that they are dumb due to trepanning.