44 thoughts on “GPT 4 with Vision and DALLE 3 Examples and Use Cases”
That is fucking nuts. Imagine what things will be like in a few years time!
Dude this guys got you all fooled and you believed him lmao
What does it mean that it can "write the code"? Is it just writing the HTML/CSS, or is it also writing the app functionality?
🤯🤯🤯🤯🤯🤯🤯🤯🤯🤯
I gave it a six panel comic I drew, so it definitely hadn't seen it before or have it explained. Still it knew exactly what's going on in that comic and explained it 10/10.
Actually, since the comic is in English it is implied that items would be written on the board from left to right….
This means that the square would have been written first…. revealing…. to the other 2 parties that they were indeed thinking of 3 different items.
Granted the full epiphany that all 3 parties were thinking different shapes would happen upon the 3rd member drawing the 3rd object.
However, given the positioning I think we can all agree that it was very likely that each character was below their corresponding shape.
Otherwise the comic writer would have moved the characters into different positions, as demonstrated by their colors, if they wanted to show the act of moving to draw on the chalk board.
Is it possible that party 2 or 3 went first yes. However given the way comics read, English is written, and people in English speaking countries act given the above, it would be illogical to assume otherwise without the writer of the comic presenting additional evidence.
This means that once the square was revealed it illuminated the concept of their false share understanding with one another.
Granted there was clearly an inclination of misunderstanding, or they wouldn't gone through the act of visualizing what they had in mind, but I would argue the Ai understood the tipping point was at moment the square was drawn and the other two characters observed the 1st's results.
This means that not only did Ai get it right, but it potentially understood the series of events far better than those debating the contrary.
I think it's important to note this means that not only was the Ai likely not wrong, it was also arguably smarter than anyone who didn't immediately intuit that. 😮
Programmers are shaking rn
Insanity. So cool.
Big thanks to all the A*holes ho contributed online with code and similar stuff to train GPT and similar AI. Big knowledge used to fill some pockets! Don't care about the hallucinations! Just keep on going!
chatgpt is now a million times smarter than me
bro i dont wanna chat with gpt anymore its gone too far🥲
All I want is software that goes through all my image files and accurately describes them, and retitles them based on that description (or puts the keywords into a meta field).
just coded my own neuro-sama clone thanks for the prompt, at 4:15
This thumbnail really made me laugh !
Yuuup gonna feed ol buddy ALL my handwritten notes for a business model (was afraid of pc notes to maintain perfect security)😅
I can read twitter without hand holding, thanks for the attempt at making an informative video.
crazy
Awesome, now put it in my glasses with a tiny camera and let me speak to it and we have made science fiction into reality.
6:33 just take a screenshot of the dish and prompt it yourself you dummy ^^
Waldo was nowhere near the center of the image.. This could be used as a universal translator if it can pick up patterns etc..
wish we could get some legal protections against training AI on copyrighted works. this whole thing is SO FUCKED
GPT to write a GPT?
This is absolutely insane
I used it to study for my midterms exams, really good.
At 15:55, Technically if you say "You can't put B next to A" you can still put A next to B and it will not be false. If you were to say that "B can't be next to A" then it would not be correct to put A next to B. I'm not sure if ChatGPT can differentiate these minute differences though.
Now just imagine applying our psychology knowledge to this and build it into everything, a perfect matrix!
More reasons to make San Francisco into a city-state and leave them to their fate.
I have the GPT ap on my phone and it can listen to and respond with speech. I played some whale click recordings and asked if it could translate it to human speech.
Give to it Voynich Manuscript…
I think the scariest part about AI is its teaching us how Dogshit humans are at processing information LOL
you're going to get us all killed
Will we end up with chat gnt with holograms or nigjtvision
let me get some deblumes pimp please
AI is truly our better, and I wholeheartedly welcome what it can do. It is akin to one's own child, certain to outlive its parents, and even more certain to reach even higher potentials.
The UK are starting to use AI to determine if people deserve benefits. Now that is scary
OMG!!
This is pretty cool and all. But if someone weaponizes this system, were screwed.
What a time to be alive, boys. What a time.
honestly, i'm not impressed with Dalle-3
Gary's the kinda guy who won't be impressed by a person that has been trained.
That is fucking nuts. Imagine what things will be like in a few years time!
Dude this guys got you all fooled and you believed him lmao
What does it mean that it can "write the code"? Is it just writing the HTML/CSS, or is it also writing the app functionality?
🤯🤯🤯🤯🤯🤯🤯🤯🤯🤯
I gave it a six panel comic I drew, so it definitely hadn't seen it before or have it explained. Still it knew exactly what's going on in that comic and explained it 10/10.
9:50
Think of a red truck
thanks for cutting programmers sallery…
Light themed twitter…. Truly a psychopath
ppl need to code them to understand NOT JUST DO!
Actually, since the comic is in English it is implied that items would be written on the board from left to right….
This means that the square would have been written first…. revealing…. to the other 2 parties that they were indeed thinking of 3 different items.
Granted the full epiphany that all 3 parties were thinking different shapes would happen upon the 3rd member drawing the 3rd object.
However, given the positioning I think we can all agree that it was very likely that each character was below their corresponding shape.
Otherwise the comic writer would have moved the characters into different positions, as demonstrated by their colors, if they wanted to show the act of moving to draw on the chalk board.
Is it possible that party 2 or 3 went first yes. However given the way comics read, English is written, and people in English speaking countries act given the above, it would be illogical to assume otherwise without the writer of the comic presenting additional evidence.
This means that once the square was revealed it illuminated the concept of their false share understanding with one another.
Granted there was clearly an inclination of misunderstanding, or they wouldn't gone through the act of visualizing what they had in mind, but I would argue the Ai understood the tipping point was at moment the square was drawn and the other two characters observed the 1st's results.
This means that not only did Ai get it right, but it potentially understood the series of events far better than those debating the contrary.
I think it's important to note this means that not only was the Ai likely not wrong, it was also arguably smarter than anyone who didn't immediately intuit that. 😮
Programmers are shaking rn
Insanity. So cool.
Big thanks to all the A*holes ho contributed online with code and similar stuff to train GPT and similar AI.
Big knowledge used to fill some pockets! Don't care about the hallucinations! Just keep on going!
chatgpt is now a million times smarter than me
bro i dont wanna chat with gpt anymore its gone too far🥲
All I want is software that goes through all my image files and accurately describes them, and retitles them based on that description (or puts the keywords into a meta field).
just coded my own neuro-sama clone thanks for the prompt, at 4:15
This thumbnail really made me laugh !
Yuuup gonna feed ol buddy ALL my handwritten notes for a business model (was afraid of pc notes to maintain perfect security)😅
I can read twitter without hand holding, thanks for the attempt at making an informative video.
crazy
Awesome, now put it in my glasses with a tiny camera and let me speak to it and we have made science fiction into reality.
6:33 just take a screenshot of the dish and prompt it yourself you dummy ^^
Waldo was nowhere near the center of the image.. This could be used as a universal translator if it can pick up patterns etc..
wish we could get some legal protections against training AI on copyrighted works. this whole thing is SO FUCKED
GPT to write a GPT?
This is absolutely insane
I used it to study for my midterms exams, really good.
At 15:55, Technically if you say "You can't put B next to A" you can still put A next to B and it will not be false. If you were to say that "B can't be next to A" then it would not be correct to put A next to B. I'm not sure if ChatGPT can differentiate these minute differences though.
Now just imagine applying our psychology knowledge to this and build it into everything, a perfect matrix!
More reasons to make San Francisco into a city-state and leave them to their fate.
I have the GPT ap on my phone and it can listen to and respond with speech. I played some whale click recordings and asked if it could translate it to human speech.
Give to it Voynich Manuscript…
I think the scariest part about AI is its teaching us how Dogshit humans are at processing information LOL
you're going to get us all killed
Will we end up with chat gnt with holograms or nigjtvision
let me get some deblumes pimp please
AI is truly our better, and I wholeheartedly welcome what it can do. It is akin to one's own child, certain to outlive its parents, and even more certain to reach even higher potentials.
The UK are starting to use AI to determine if people deserve benefits. Now that is scary
OMG!!
This is pretty cool and all.
But if someone weaponizes this system, were screwed.
What a time to be alive, boys. What a time.
honestly, i'm not impressed with Dalle-3
Gary's the kinda guy who won't be impressed by a person that has been trained.