Seems like BOT really thinks GPT-4 is noteworthy because it's good at... baby math and programming.
Kek, give it any exercise from any graduate STEM book and it will fail miserably.
"MUH GPT-4 DOE"
Seems like BOT really thinks GPT-4 is noteworthy because it's good at... baby math and programming.
Kek, give it any exercise from any graduate STEM book and it will fail miserably.
It's not enough at any of these things, it's only good at predicting what usually comes next after what you just typed. When will you retards understand this?
BOT is just /v/ retards and 4chan tards spamming bait threads.
/v/ is fine but 4chan and /int/ should be removed from BOT.
That's literally all a human brain does, just more complex and all the time.
Most people can't do graduate STEM exercises which is why its a big deal. Right not GPT-4 is around college sophomore level.
Yeah and it's getting better faster than your average student.
Stop lying, dumbass.
How am I lying? GPT-2 was released 4 years ago and kids were on 1st grade are now on the 5th, meanwhile ChatGPT solves sophomore problems and aces the bar. Another 4 years and the kid will still be in school while ChatGPT will be doing master level stuff at the very least
yeah, another thing is the breadth of its knowledge. even if it might not be super deep in all subjects yet it's more than good enough to have to teach it the basics. sometimes it can quite funny as well. then again I'm sure it will be telling I'm a good Bing jokes as it is murdering us in our beds
moron, it can't even generate simple sets.
>Paper says it's not good with math
>Haha lmao, it's a fraud it's not good with math
Do retards really?
It's the simplest math question. even easier than the majority of SAT math questions, GPT3.5 and GPT4 are good at solving SAT math problems but apparently, they can't answer simpler math questions.
I don't know what high school you went to but i would rate that question harder than most SAT questions simply because we didn't have that kind of questions in school.
its a natural language model not a mathematics program
chatGPT did the multiplications correctly, but multiplied the wrong numbers. It failed at logical reasoning.
It is becoming twice as good in every 6 months. If the progress didn't flat out, llm is going to replace every single white collar job.
they will probably release a "tock" (nothing) MAJOR UPDATE next after the glowies give them a stern talking to.
can your wife's bf?
it wasn't specifically fine tuned for programming and it's already better at leetcode than junior devs
GPT-4 is already AGI.
GPT-4 is the anti-Christ
GPT-4 is literally Hitler
I wonder what it failed at. The easy leet codes are trivial and it only got 31/41.