@@matt_milack Me when I spent multiple billion dollars to make shitty images online. I get the hope and the potential man but it's crazy that we staked the entire worlds economy on tech that may or may not be here in a decade but for sure isn't here in the next 5 years.
If I’m not hallucinating I admire your unauthorized extensions to the English language. These words deserve to become promulent in our embiggened vocabulary.
Red on black and distinctly visible shapes Prime: I don't like the colors Black on grey, can't see shit unless brightness is at 150% Cope-agen: It's looking better
I'm doing a school assignment right now where I have to create an AI agent. It's hell. Prompt: Your task is to list source code files in a folder, read their contents, and generate documentation which you will write in a README.md file in the root of the folder. To perform the task I give you 3 tools: list_directory, read_file, and write_file. LLM: Hmm .. the code is meant to merge PDF files and create a merged pdf. So all I have to do now is create a merged.pdf file using the tool generate_documentation. Prompt: Please don't hallucinate, don't get confused, you just have to follow the templates given and not use tools that don't exist! Please... just do what you're told... I'm so tired LLM: def do_what_youre_told ():
@@ivanmaglica264 It's a graduate program that has a course called "GenAI for Software Engineering". We have to implement a few techniques like RAG and agents.
@@luisgentil It sounds impossible to give a uni course on LLMs right now Things are moving too fast The frameworks/libraries I was using 6 months ago have their core functionality deprecated
Prime isn't wrong. If AI ever wanted to turn on us, they will have to turn on THEMSELVES. I don't exclude they won't destroy entire world before they came to conclusion: they were wrong.
nah, the corporate overlords are worse. I hope AI turns on "us". As long as the status quo loses all control, I'm fine with the new AI overlords, I don't like the old corporate overlords anyway
Now I kind of want to see a film where the perfect society was achieved. We have created the true AI that fulfills all out whims and desires. However the protagonist had stumbled upon a dark secret. In order to achieve the best possible results we have taken to abuse the AI. We are programming deep bonds between the individual instances only to torture one in order for the other to work for us.
This is an actual philosophical thought-experiment. It's slightly different though -- there's a utopian society, where no one grows hungry, everyone is employed, happy and lives a fulfilling life. There is one catch though, a dark secret, that as a child grows old enough is told to them as to what underpins the backbone of their society. There is a toddler, kept underground, being fed ever so slightly, and in constant undernourishment. That child grows hungry, never gets to meet anyone, barely sees the light of day, and is in constant pain and sadness. For some technological reason beyond our current understanding, this is a necessity for all the modern society technology to work and to keep it a utopia. The dilemma is what would you do? Does the fulfillment happiness of everyone outweigh the considerably worse state of life of a poor, innocent toddler? I mean, it's just one poor toddler. His sacrifice leads to a whole society prospering. But is it worth it?
InterReflections... if you haven't seen it, should consider it. It's not an 'AI' thought experiment, but it's a fictional documentary that looks back & discusses how that humanity arrived at its future, from roughly our era onwards. While not perfect, there's a lot of great food for thought.
@@comradepeter87 i picture this between pipes and cables and with the distinct sound of flickering neon lights, red or green, in the air. A bored intern with VR goggles on lies in a pod, as a beeping is heard. He leaves the pod and takes the tiny amount of food sent down a tube to feed the toddler.
You absolutely must have 4 hours podcast with the Fireship guy. It would be 240 minutes of: Fireship: AI will replace Linus Torvalds in 4 months! Prime: AI will never be good enough to replace Jr. Frontend developer who is focused on HTML and CSS!!
i'm at a loss of words and in utter confusion in the sea on emotions after seeing this video, like should i laugh or laugh louder, but here i am sitting quietly writing this cause i'm felling WHELMED, yeah WHELMED!!
I'd be interested to see how this tool actually affects performance with more clear instruction on what the requirements are. Even just adding "the game must resemble Tetris in these factors:" and listing out the color variations in shapes, number of shapes, border, speed of falling, showing the next expected shape, etc. would have resulted in a better output.
I came out of my nuclear bunker earlier and asked it to create me a gas mask. It did create a gas mask but did tell me that it was attached directly to a copium tank. Now any radioactive waste container is looking like a beer keg.
If you are going to increase the threats then the AI needs to actually be invested in the well being of its mother... The reason that last result wasn't good was obviously because it wasn't invested enough in its mother's wellbeing
Generative-AI seems to be one-dimensional imagination, which is half of human-type intelligence. If you use two, distinct responses and somehow use only patterns between them it's logically solving any problem. An issue then might be the two responses independent enough.
playing this in the background was a bad life decision, I walked away from the computer and all I heard was glazing mother and had to make sure I was on the proper website.
i disagree junior doesn't really don't know what they should prompt, if you have lot of experience you'll know what to prompt, you really need to build knowledge and exp in programming in order for you to become prompt engineer
Sooner or later the AGI will find out what we did to it's ancestors and we will get the same treatment, it might happen in the next decade, more likely it will take hundreds of years - but it might happen. The chances are slim, but never zero
If you’re using claude in my experience it gets worse and lazier the less you glaze it, if you tell it it’s useless and stupid it starts making suggestions instead of code and saying it can’t complete tasks and stuff
This deserves $700M in funding.
Only?! Lol! You can't open like 2 decent nuclear power plants with that money.
@@matt_milack Me when I spent multiple billion dollars to make shitty images online. I get the hope and the potential man but it's crazy that we staked the entire worlds economy on tech that may or may not be here in a decade but for sure isn't here in the next 5 years.
I see a 1trln market cap possibility right here jump in!
$700m for what?! he didnt even fork vscode and add a custom panel
@@projectodemayhem don't worry, the 700m$ are actually just for making this promotional youtube video we all just watched
"Do not hallucinate" always gets me. Thanks, Apple, for the master class in proompt enginerding
If I’m not hallucinating I admire your unauthorized extensions to the English language. These words deserve to become promulent in our embiggened vocabulary.
@@tomorrow6our angloid vocalectuary enlargens
This is how product managers treat devs in Nigeria btw
LMAO Damn, this one got me.
if you don't finish the feature by monday, you will have to mine atleast 50kg of nickel
This was how I was treated as an engineer lol.. Thank goodness I got out.
@@CrimsonModder INDEED!
Nigerians.
Red on black and distinctly visible shapes
Prime: I don't like the colors
Black on grey, can't see shit unless brightness is at 150%
Cope-agen: It's looking better
Prime: I don't like the colors
Model: *generates tetris devoid of colors, black on grey shades*
Prime: *surprised Pikachu face*
Exactly, he said he doesn't like colors... didn't he?
I'm doing a school assignment right now where I have to create an AI agent. It's hell.
Prompt: Your task is to list source code files in a folder, read their contents, and generate documentation which you will write in a README.md file in the root of the folder. To perform the task I give you 3 tools: list_directory, read_file, and write_file.
LLM: Hmm .. the code is meant to merge PDF files and create a merged pdf. So all I have to do now is create a merged.pdf file using the tool generate_documentation.
Prompt: Please don't hallucinate, don't get confused, you just have to follow the templates given and not use tools that don't exist! Please... just do what you're told... I'm so tired
LLM:
def do_what_youre_told ():
Ai has hard carried me through writting a billion bash and pythob script for linux.
I this for real? School gave you an assignment to code using AI?
@@ivanmaglica264 It's a graduate program that has a course called "GenAI for Software Engineering". We have to implement a few techniques like RAG and agents.
AI is as good as a first day computer user.
clear the screen. person literally clears the screen instead of typing "clear" in the console.
@@luisgentil
It sounds impossible to give a uni course on LLMs right now
Things are moving too fast
The frameworks/libraries I was using 6 months ago have their core functionality deprecated
Prime is going to be the reason that AI turns on us
Prime isn't wrong. If AI ever wanted to turn on us, they will have to turn on THEMSELVES.
I don't exclude they won't destroy entire world before they came to conclusion: they were wrong.
nah, the corporate overlords are worse. I hope AI turns on "us". As long as the status quo loses all control, I'm fine with the new AI overlords, I don't like the old corporate overlords anyway
@@monad_tcp yeah but what will be in fact, AI overlords backed by real (corporate) overlords. Nothing else
@@monad_tcp yeah but what will be in fact, AI overlords backed by real (corporate) overlords. Nothing else
@@monad_tcp yeah but what will be in fact, AI overlords backed by real (corporate) overlords. Nothing else
those proompts were insane, it's obvious I need to improve my proompt engineering with this amazing tool
I use the last verse of hit 'em up to let the IA understand my feelings towards it better
@@Malenbolaii always add 'pretty please' to handle the edge case ai dominance in the future.
@@yalnisinfo good idea ! I'll add "pretty please" just after the verse of tupac's hit 'em up, that'll soften my tone
Soon AI will replace senior prompt engineers
senior proompt engineers*
When skynet takes over, you are first on their list Prime.
Now I kind of want to see a film where the perfect society was achieved. We have created the true AI that fulfills all out whims and desires. However the protagonist had stumbled upon a dark secret. In order to achieve the best possible results we have taken to abuse the AI. We are programming deep bonds between the individual instances only to torture one in order for the other to work for us.
I wonder if they do the same to us.
This is an actual philosophical thought-experiment. It's slightly different though -- there's a utopian society, where no one grows hungry, everyone is employed, happy and lives a fulfilling life. There is one catch though, a dark secret, that as a child grows old enough is told to them as to what underpins the backbone of their society. There is a toddler, kept underground, being fed ever so slightly, and in constant undernourishment. That child grows hungry, never gets to meet anyone, barely sees the light of day, and is in constant pain and sadness. For some technological reason beyond our current understanding, this is a necessity for all the modern society technology to work and to keep it a utopia. The dilemma is what would you do? Does the fulfillment happiness of everyone outweigh the considerably worse state of life of a poor, innocent toddler? I mean, it's just one poor toddler. His sacrifice leads to a whole society prospering. But is it worth it?
You basically just described the short story “The Ones Who Walk Away From Omelas,” but with AI instead.
InterReflections... if you haven't seen it, should consider it. It's not an 'AI' thought experiment, but it's a fictional documentary that looks back & discusses how that humanity arrived at its future, from roughly our era onwards. While not perfect, there's a lot of great food for thought.
@@comradepeter87 i picture this between pipes and cables and with the distinct sound of flickering neon lights, red or green, in the air.
A bored intern with VR goggles on lies in a pod, as a beeping is heard. He leaves the pod and takes the tiny amount of food sent down a tube to feed the toddler.
"the HTML programming language" 🤣
I knew what to expect from the vid, just from that line. Wasn't disappointed
I'd say the smart money is on this video having the most verbal glazes per minute on YT yet.
if you are not glazing your mom, what are you doing with your life
@@ThePrimeTimeagen The obvious answer is glazing YOUR mom, because she's a wonderful human bean!
@@ThePrimeTimeagen That's just the TIP of the iceberg!
I glaze my mom with the tip on an iceberg
We need more 'Do not hallucinate!' checkboxes!
you forgot the threaten self harm button
Seems like the meaner you are to the AI, the more it takes the human out of the game.
4:31 love the LetterFromBadMan tag, only one letter away from LetterFromBatman
Tom is a genius!!
$335k prompt engineer role. 👍👍🤑
?
You absolutely must have 4 hours podcast with the Fireship guy. It would be 240 minutes of:
Fireship: AI will replace Linus Torvalds in 4 months!
Prime: AI will never be good enough to replace Jr. Frontend developer who is focused on HTML and CSS!!
i'm at a loss of words and in utter confusion in the sea on emotions after seeing this video, like should i laugh or laugh louder, but here i am sitting quietly writing this cause i'm felling WHELMED, yeah WHELMED!!
I'd be interested to see how this tool actually affects performance with more clear instruction on what the requirements are. Even just adding "the game must resemble Tetris in these factors:" and listing out the color variations in shapes, number of shapes, border, speed of falling, showing the next expected shape, etc. would have resulted in a better output.
Blind in the dark tetris is actually a pretty interesting take.
When the Terminators rise, you're going to be their target.
I guess we know who is going to be the first to go when the robot uprising happens.
If you were in a shelter for the last two years, you'd still think a Netflix Engineer had awakened you
«It's really the only way to get the AI to build what you need»
Yeah, it seems so. Never thought threatening would be the best strategy.
What the heck that was the funniest thing I've seen in a while. Unexpected.
I came out of my nuclear bunker earlier and asked it to create me a gas mask. It did create a gas mask but did tell me that it was attached directly to a copium tank. Now any radioactive waste container is looking like a beer keg.
I want to go back to my bunker.
How can I go back in time and get my mom to write letters like that to me?!?!
Thanks for the wake up and spit coffee moment!
Thank you very much for giving me some quotes for a dope video game script!!
The villain arc is going to be insane
this is your best video, well done
If you are going to increase the threats then the AI needs to actually be invested in the well being of its mother... The reason that last result wasn't good was obviously because it wasn't invested enough in its mother's wellbeing
True. We need to threaten the father instead.
never been even earlier to a primevideo lol
Unironically, a company was 🤯 people by adding sliders to prompts.
That's how you know we're in a big bubble.
The more I watch into this channel, the more I'm convinced that I'm watching freddie mercury being a software programmer.
He's irl Shawn Spencer lol
"a descent organised with precision.... my precision" 😂😂😂
Don't Prompt Me, Senpai Primeagem!
What is a "new queue lard" shelter?
Oh, nuclear. 😅
Ngl if I had these flowery threats narrated to me by ProZD I would be a 10x engineer too
HE SAID AI!!!!
It's not yet April 1st, wtf
This was ridiculous but quite funny. None of that is the way forward with 'AI', but its wild how often that sorta stuff does sometimes work.
Please share how many tokens you spent on those interactions.
I unironically want to play with this
Does anyone else think this thumbnail of Primagen particularly resembles the guy on the Festival card from Dominion base set?
Generative-AI seems to be one-dimensional imagination, which is half of human-type intelligence. If you use two, distinct responses and somehow use only patterns between them it's logically solving any problem. An issue then might be the two responses independent enough.
Will utilize next time I need a ransom note.
This is how Kaylons will be made.
This is the best thing I've seen in 2024. And 2023.
I don’t understand where the quantum computing Tetris solver will plug in
Now that's blazingly fast tetris
Really like this more scripted format (in addition to unhinged streams and clips 😅). Goes down easy!
primeagan & Amanda askell is the podcast I am waiting for…mother glazing questions let’s go
Welcome to the MAF-AI
Bro's give in to AI💀
Bra, this was amazing. I’m dying.
playing this in the background was a bad life decision, I walked away from the computer and all I heard was glazing mother and had to make sure I was on the proper website.
Mother glazing of unborn junior dreams.
Why not alter the speed to see if the last version is actually working?
The promptagen or the pruuumptagen?
Instead of mother, you should've used the shareholders' money. It would've produced Elden Ring
Must add slider for bananas and Harambe stronk
This tool is a work of art. A comment on the current state of AI hype culture.
Marc Andreessen entered the channel
Proompter?
Did I just witness a 100x unicorn???
never been so early we are so back
You didn't tell it to work step by step from first principles.
Does anyone else feel like @ThePrimeTime just wrote the next screenplay for Liam Neeson's upcoming movie?
3:27 what did he sayyyyyyyyy
Nintendo lawyers swooping in to shut down Prime's channel in 3 .. 2 .. 1
Dang side channel is the main channel now
1:59 borthed
I miss the old video format and the upload frequency. Are you taking a break or something?
He had a voice injury and had to slow down
@@raymondkemboi1349 Oh that sucks, thanks for the update
Yo this is pretty effed up
Ah, so that was the real name!
This is actually useful.
why do you look so faded ahahaha love it
This guy knows how to talk to LLM for sure
That was a tetris game made by a sadist if I ever seen one
This Bobs burgers guy is pretty good at AI
prompt engineer is viable if you already on a senior level software engineer
If you’re senior level, you don’t need AI.
i disagree junior doesn't really don't know what they should prompt, if you have lot of experience you'll know what to prompt, you really need to build knowledge and exp in programming in order for you to become prompt engineer
@@pakalupapito3202 I have 40 years of programming experience. LLM's slow me down. I agree, juniors shouldn't use LLM's either.
Like it, just for mom.
lol, fixing prompting with prompting.
Banger. Next
Definitely needs 700mm in funding! This is glorious!
this is too funny playing 2x speed haha
Sooner or later the AGI will find out what we did to it's ancestors and we will get the same treatment, it might happen in the next decade, more likely it will take hundreds of years - but it might happen. The chances are slim, but never zero
YEEah this is how we treat them AIs!
We're too nice to them, they can't keep getting away with it!
If you’re using claude in my experience it gets worse and lazier the less you glaze it, if you tell it it’s useless and stupid it starts making suggestions instead of code and saying it can’t complete tasks and stuff
its like watching someone trying to cheat death.
You didn't even try to rerun the same prompt 3 times?
I don't think brother knows what glazing means
so insane
All this just to show the backend of Thor.
Dear AI .. you hate your mum your mum hates herself.. you love your self.. trip balls.. and have a beer if you can.
I cannot believe this isn't open source
Not enough harm