>Nah, even worse the skull isn't gonna be an obstacle for a 9mm
You are actually moronic if you believe this. My old man killed a black bear that had buckshot embedded in its skull and there are countless cases of Elephants and Rhinos shrugging off smaller caliber rounds and going sicko mode on humans afterward. Even something as small as a feral hog can shrug off small-arms fire.
An adult Tyrannosaurus Rex is an 8 to 10-ton carnivore that regularly bit other members of its species on the face with the most powerful bite to have ever evolved and took absolute beatings from hadrosaurs and ceratopsians that outweighed modern bull elephants significantly. It had keratinized scales on much of its head which were especially prominent on its brow ridge and sustained horrific sorts of injuries that would have killed most mammalian predators.
You are not killing a Rex with a 9mm.
While a elephant sized creature might have a skull able to bounce a pistol round, your talk of smaller game being bullet proof is pure mythology. Thousands upon thousands of hogs are dropped every year by 22lr and 17hmr.
Those are all issues of shot placement. Bullets are extremely lethal. A 9mm is far from ideal for a Trex but it can very, very easily penetrate a skull, organs, etc. It could be quite dangerous at close range against a Trex, but ballistically, it's absolutely possible.
It also needs to go through skin (pretty tough and thick) AND hit vital organs aka the brain, wich is small af.
tl;dr : good luck with a 9mm
Can you kill an elephant with a 9mm?
Why do Elephant Guns exist then?
https://i.imgur.com/pzJgWQK.jpg
>Nah, even worse the skull isn't gonna be an obstacle for a 9mm
You are actually moronic if you believe this. My old man killed a black bear that had buckshot embedded in its skull and there are countless cases of Elephants and Rhinos shrugging off smaller caliber rounds and going sicko mode on humans afterward. Even something as small as a feral hog can shrug off small-arms fire.
An adult Tyrannosaurus Rex is an 8 to 10-ton carnivore that regularly bit other members of its species on the face with the most powerful bite to have ever evolved and took absolute beatings from hadrosaurs and ceratopsians that outweighed modern bull elephants significantly. It had keratinized scales on much of its head which were especially prominent on its brow ridge and sustained horrific sorts of injuries that would have killed most mammalian predators.
You are not killing a Rex with a 9mm.
Those are all issues of shot placement. Bullets are extremely lethal. A 9mm is far from ideal for a Trex but it can very, very easily penetrate a skull, organs, etc. It could be quite dangerous at close range against a Trex, but ballistically, it's absolutely possible.
Problem solved.
An elephant was killed with 87 rounds of various pistol calibers and buckshot, completely randomly dispersed throughout it's body. Load this thing with copper solids and you'll absolutely drop a T Rex.
>unironically posting chatgpt
you understand that it just makes shit up out of whole cloth right? especially if it's something it hasn't encountered before
that said yeah obviously. it's not like dinosaurs evolved running away from wild apes with hipoints
The newest variant of chatGPT actually has a semblance of understanding, able to understand and seperate concepts and solve real problems.
It's rather remarkable actually since it is in essence a trained chatbot, but it's allowed to learn workarounds and build learning loops and all sorts of fun stuff.
So no, it doesn't always "just make shit up", though it's intelligence is unfortunately limited to the internet's collective intelligence as it uses posts as training data resources and it is able to make mistakes.
Unfortunately it does not learn in real time, but is a static model they update over time that can adapt to short term functions and commands.
when i say encountered before i mean directly within its training data. the more logical leaps it needs to go through to extract something novel the less likely the response is to have any correlation to reality. and when it does go off the rails its only ability to deal with that is to cover its tracks as well as possible by making it look passable. this is inherite and unrelated to the improvements made in the model, probably because it's a pure forward feed of data through the network without the ability to scrutinize its own decisions in real time.
it is, at its core, a chatbot that functions the same way chatbots from the 1990s functioned, just with orders of magnitude more training data. i can already feel AI fall setting in, and i suspect this time the winter will be harsher on the heavily invested corporations than the one in the 1990s was.
that, and corporate ML algorithms are getting so thoroughly mogged by open source communities they're running to the government desperately begging to regulate their open source competition out of existence
>it's a consequence of it being a Markov process
i believe that is equivalent to what i described, a feed forward only model. it has the ability to recognize flaws in its own reasoning and change its responses (approximating this with prompting makes the base model perform MUCH better), but in a single run it can only work with what it has already done even if the road it has gone down is erroneous.
that's part of the problem, the other problem that seems completely separate is these models have no way of self reflecting on what they DON'T know. no matter how well it performs it will always provide a potentially erroneous response to any question it doesn't know the answer to
>So no, it doesn't always "just make shit up"
Literally all it does is make shit up. Sometimes the shit it makes up is true and sometimes it's not. It can't evaluate mathematical equations, but it knows that, statistically, this collection of characters almost always precedes this other character. That's good enough most of the time, until it isn't.
> The newest variant of chatGPT actually has a semblance of understanding, able to understand and seperate concepts and solve real problems.
It doesn't >It's rather remarkable actually since it is in essence a trained chatbot
It's not actually, a lot of it's training data was stuff like tutorials. Which is why it's remarkably good at doing stuff that is straight paste from the internet, like how to create some basic javascript function or HTML page.
If you stray away from that, it fails completely even if it's just the same thing applied in a slightly different way. >So no, it doesn't always "just make shit up",
It does, literally.
As
>So no, it doesn't always "just make shit up"
Literally all it does is make shit up. Sometimes the shit it makes up is true and sometimes it's not. It can't evaluate mathematical equations, but it knows that, statistically, this collection of characters almost always precedes this other character. That's good enough most of the time, until it isn't.
says, it's literally a very big language model that is essentially trying to predict what words follow another(it deals with words and not characters as that anon says since it uses word embedding and stuff, which means that it can't create new words like it's often done).
Look at around 1:50 in this video, ChatGPT literally created out of thin air a 'UnityEngine API' for Python that doesn't exist. Probably because python is very used, so normally APIs are available to it.
Wrong entirely there was an entire talk in how GPT4 is rather nuts. It can do things that are highly complex that are outside of tutorials and apply concepts to other concepts. You can try it using obscure programs like the unicorn experiment.
It can make mistakes but it's not perfect either. It even corrects itself.
It's arguably better than most humans at tasks.
Don't downplay it to "it's just a chatbot that makes shit up" because it has far more utility than that.
I'm sorry you had to find out this way, but most humans are just making shit up as they go too. Natural language models are really really really good at bullshitting. So good that they are actually correct in most cases, but at the most basic level it's still bullshitting.
>it deals with words and not characters as that anon says since it uses word embedding and stuff
Technically it uses arbitrarily sized sub-word tokens. Not to long ago someone had their language model leaked (I think it was Meta) and that included the full list of tokens, and many of them were fragments of words, some just single characters. Obviously every model works a little differently, but I would think that stuff like mathematical operators would each be their own tokens.
Yeah the penetration constant is wrong, it's assuming a perfectly hard bullet that can't deflect and doesn't deform, and we all know bullet construction is a big deal.
Anything that could kill an elephant would wreck any dinosaurs. They're just big animals unfortunately rather than magical monsters. I wish they were monsters.
In terms of "dumb luck shots" some butthole poacher in Indonesia killed an elephant with four shots of .22LR, although I guess Asian elephants are a bit smaller
But in terms of "holy frick he did this consistently" WDM Bell killed 800 African elephants with 1,200 rounds from a 7mm Mauser, which seems like a shockingly small amount of gun considering everybody by WW1 called it doctrinally inadequate to kill humans
> Bell kept tally on that Rigby-Mauser. Expending 1,200 rounds of military full-metal jacket, the rifle took 800 elephants, with an average of 1.5 shots per kill.
>Anything that could kill an elephant would wreck any dinosaurs.
Actually it's even easier, dinosaurs overall evolved lighter skeletons, which was one of the reasons why they grew so big, so a dinosaur skull would probably be less dense than one of a elephant, probably less dense than a crocodile one as well
An actual T rex is just a giant Vulture so while I don't think 9mm would do the job anything in .30 cal is going to get the job done with a hunting load. I'm also guessing if a rex gets shot at it will probably fricking run.
Apparently in Lost World they were going to have the T-Rex attack a Navy Seal Base in San Fran until they realized exactly how fast a T-Rex would be dropped by a bunch of guys automatic rifles and scrapped the idea.
I heard that one of the new Jurassic movies were supposed to have the Indo-Raptor being auctioned to militaries like what the frick is that supposed to do. It's expensive, the size of a truck, not bullet proof and going to light up on thermals like a Christmas tree... great combo.
I don't know, I don't watch trash, I am just taking Jenny's word on it.
Why? Are you a T rex
maybe
Nah, even worse the skull isn't gonna be an obstacle for a 9mm...however maybe if it had crocodile-like skin, thicker then it could save it.
However, it is rather disappointing to think about it...
>Nah, even worse the skull isn't gonna be an obstacle for a 9mm
You are actually moronic if you believe this. My old man killed a black bear that had buckshot embedded in its skull and there are countless cases of Elephants and Rhinos shrugging off smaller caliber rounds and going sicko mode on humans afterward. Even something as small as a feral hog can shrug off small-arms fire.
An adult Tyrannosaurus Rex is an 8 to 10-ton carnivore that regularly bit other members of its species on the face with the most powerful bite to have ever evolved and took absolute beatings from hadrosaurs and ceratopsians that outweighed modern bull elephants significantly. It had keratinized scales on much of its head which were especially prominent on its brow ridge and sustained horrific sorts of injuries that would have killed most mammalian predators.
You are not killing a Rex with a 9mm.
While a elephant sized creature might have a skull able to bounce a pistol round, your talk of smaller game being bullet proof is pure mythology. Thousands upon thousands of hogs are dropped every year by 22lr and 17hmr.
.45 acp then huh?
Those are all issues of shot placement. Bullets are extremely lethal. A 9mm is far from ideal for a Trex but it can very, very easily penetrate a skull, organs, etc. It could be quite dangerous at close range against a Trex, but ballistically, it's absolutely possible.
Mag dumping it would undoubtedly create wounds that might fester and go septic killing it eventually
For immediately killing it, you'd probably need to land an eye shot which could then easily make it to the brain
Problem solved.
An elephant was killed with 87 rounds of various pistol calibers and buckshot, completely randomly dispersed throughout it's body. Load this thing with copper solids and you'll absolutely drop a T Rex.
It also needs to go through skin (pretty tough and thick) AND hit vital organs aka the brain, wich is small af.
tl;dr : good luck with a 9mm
ok, I see it would a difficult shot, but still...all the firepower you need to kill a t rex is a damn 9mm
A shitload of 22lr could probably do the trick but how much time do you have ?
People don't hunt big fives with 9mm for a reason.
>unironically posting chatgpt
you understand that it just makes shit up out of whole cloth right? especially if it's something it hasn't encountered before
that said yeah obviously. it's not like dinosaurs evolved running away from wild apes with hipoints
The newest variant of chatGPT actually has a semblance of understanding, able to understand and seperate concepts and solve real problems.
It's rather remarkable actually since it is in essence a trained chatbot, but it's allowed to learn workarounds and build learning loops and all sorts of fun stuff.
So no, it doesn't always "just make shit up", though it's intelligence is unfortunately limited to the internet's collective intelligence as it uses posts as training data resources and it is able to make mistakes.
Unfortunately it does not learn in real time, but is a static model they update over time that can adapt to short term functions and commands.
when i say encountered before i mean directly within its training data. the more logical leaps it needs to go through to extract something novel the less likely the response is to have any correlation to reality. and when it does go off the rails its only ability to deal with that is to cover its tracks as well as possible by making it look passable. this is inherite and unrelated to the improvements made in the model, probably because it's a pure forward feed of data through the network without the ability to scrutinize its own decisions in real time.
it's a consequence of it being a Markov process.
it is, at its core, a chatbot that functions the same way chatbots from the 1990s functioned, just with orders of magnitude more training data. i can already feel AI fall setting in, and i suspect this time the winter will be harsher on the heavily invested corporations than the one in the 1990s was.
that, and corporate ML algorithms are getting so thoroughly mogged by open source communities they're running to the government desperately begging to regulate their open source competition out of existence
>it's a consequence of it being a Markov process
i believe that is equivalent to what i described, a feed forward only model. it has the ability to recognize flaws in its own reasoning and change its responses (approximating this with prompting makes the base model perform MUCH better), but in a single run it can only work with what it has already done even if the road it has gone down is erroneous.
that's part of the problem, the other problem that seems completely separate is these models have no way of self reflecting on what they DON'T know. no matter how well it performs it will always provide a potentially erroneous response to any question it doesn't know the answer to
>So no, it doesn't always "just make shit up"
Literally all it does is make shit up. Sometimes the shit it makes up is true and sometimes it's not. It can't evaluate mathematical equations, but it knows that, statistically, this collection of characters almost always precedes this other character. That's good enough most of the time, until it isn't.
> The newest variant of chatGPT actually has a semblance of understanding, able to understand and seperate concepts and solve real problems.
It doesn't
>It's rather remarkable actually since it is in essence a trained chatbot
It's not actually, a lot of it's training data was stuff like tutorials. Which is why it's remarkably good at doing stuff that is straight paste from the internet, like how to create some basic javascript function or HTML page.
If you stray away from that, it fails completely even if it's just the same thing applied in a slightly different way.
>So no, it doesn't always "just make shit up",
It does, literally.
As
says, it's literally a very big language model that is essentially trying to predict what words follow another(it deals with words and not characters as that anon says since it uses word embedding and stuff, which means that it can't create new words like it's often done).
Look at around 1:50 in this video, ChatGPT literally created out of thin air a 'UnityEngine API' for Python that doesn't exist. Probably because python is very used, so normally APIs are available to it.
Wrong entirely there was an entire talk in how GPT4 is rather nuts. It can do things that are highly complex that are outside of tutorials and apply concepts to other concepts. You can try it using obscure programs like the unicorn experiment.
It can make mistakes but it's not perfect either. It even corrects itself.
It's arguably better than most humans at tasks.
Don't downplay it to "it's just a chatbot that makes shit up" because it has far more utility than that.
I'm sorry you had to find out this way, but most humans are just making shit up as they go too. Natural language models are really really really good at bullshitting. So good that they are actually correct in most cases, but at the most basic level it's still bullshitting.
>it deals with words and not characters as that anon says since it uses word embedding and stuff
Technically it uses arbitrarily sized sub-word tokens. Not to long ago someone had their language model leaked (I think it was Meta) and that included the full list of tokens, and many of them were fragments of words, some just single characters. Obviously every model works a little differently, but I would think that stuff like mathematical operators would each be their own tokens.
>no, it doesn't always "just make shit up"
nobody said it "ALWAYS" does, dipshit
>0.09 joules to penetrate 15cm of bone
>doubt.jpg
ChatGPT is like this moron at high school that knows the general idea but messes up the application.
Yeah the penetration constant is wrong, it's assuming a perfectly hard bullet that can't deflect and doesn't deform, and we all know bullet construction is a big deal.
Anything that could kill an elephant would wreck any dinosaurs. They're just big animals unfortunately rather than magical monsters. I wish they were monsters.
In terms of "dumb luck shots" some butthole poacher in Indonesia killed an elephant with four shots of .22LR, although I guess Asian elephants are a bit smaller
But in terms of "holy frick he did this consistently" WDM Bell killed 800 African elephants with 1,200 rounds from a 7mm Mauser, which seems like a shockingly small amount of gun considering everybody by WW1 called it doctrinally inadequate to kill humans
Bell rarely had to shoot twice, I don't know where you're getting your figure from. He even one-shot dropped many elephants with a .275 rifle.
> Bell kept tally on that Rigby-Mauser. Expending 1,200 rounds of military full-metal jacket, the rifle took 800 elephants, with an average of 1.5 shots per kill.
>Anything that could kill an elephant would wreck any dinosaurs.
Actually it's even easier, dinosaurs overall evolved lighter skeletons, which was one of the reasons why they grew so big, so a dinosaur skull would probably be less dense than one of a elephant, probably less dense than a crocodile one as well
shoot out its kneecaps first, cripple it
Been documented that a .22 will kill a grizzly, but fricked if I'd stake my life on it
.22 would bounce inside and make mush out of its brain.
.45ACP (Anti-Cretaceous Problems) would also stop it right on its boots.
An actual T rex is just a giant Vulture so while I don't think 9mm would do the job anything in .30 cal is going to get the job done with a hunting load. I'm also guessing if a rex gets shot at it will probably fricking run.
>shoot each eye once with bb gun
>come back in a week
Apparently in Lost World they were going to have the T-Rex attack a Navy Seal Base in San Fran until they realized exactly how fast a T-Rex would be dropped by a bunch of guys automatic rifles and scrapped the idea.
I heard that one of the new Jurassic movies were supposed to have the Indo-Raptor being auctioned to militaries like what the frick is that supposed to do. It's expensive, the size of a truck, not bullet proof and going to light up on thermals like a Christmas tree... great combo.
I don't know, I don't watch trash, I am just taking Jenny's word on it.
?t=640
I would lambast you in front of the class if you crticially presented that 0.1 joules will bore a 9mm diameter hole through 150mm of skull bone.
AI is literally just making shit up my dude it is not a reliable source of information
you’d need a heavier grain +p round. I would try it with a standard 115.
What did tiborasaurus ever do to you?
i would not trust a hallucinating predictive keyboard on anything in particular
Can you kill an elephant with a 9mm?
Why do Elephant Guns exist then?