ChatGPT is Frighteningly Good

[@enricotartarotti]

Learn how I ACTUALLY made my most successful videos with hands-on, practical behind the scenes breakdowns: www.enricotartarotti.com/storybehind?

[@bbsara0146]

i would use chatGPT to write a 5 paragraph essay if I was a kid 100%

[@Someone-ox1oh]

But i think its too good so it will be sus

[@FoxMaster]

@@Someone-ox1oh no it isn’t lol.

[@Someone-ox1oh]

@@FoxMaster ait bruv

[@mylesklapatiuk8812]

i kinda maybe already have and believe me it works.

[@v380riMz]

I have written about 2-3k words within my essay using ChatGPT. I simply copy text from the source, and ask GPT to either rephrase it, make the text either more simple to understand or formal. Also ask it to write it in such way it does not get picked up by plagiarism checkers. It works amazingly well LOL. However you can't let it write things you have to write on your own

[@laurens678]

AI now can create amazing art, creative ideas, new storylines for books and movies, and complex software code all from scratch, imagine where it will be in 10 years

[@danielc81]

AI is the anti Christ.

[@GS-tk1hk]

Just a few years ago, if anyone would have said this is where we will be today, most people would have laughed. Exponential progress is hard to intuitively grasp. I mean, DALLE2 came out like half a year ago and now the entire digital art industry is being turned upside down. I wonder what seemingly impossible things we are laughing at today that will actually become a reality in the coming years

[@SuperGopi56]

I predict a lot of people will be jobless in years to come...

[@danielc81]

@@SuperGopi56 that’s the plan of the elites

[@runnergo1398]

Ever seen the movie Ex Machina or played the game Detroit: Become Human? It'll be like that (hopefully).

[@carpet7664]

I can confirm the school thing 100%. Just today I was coming up on a deadline for an essay. Having heard about chatGPT recently I thought what the heck maybe I’ll ask it to write an essay on my topic just for some ideas. Believe it or not it was able to even find some information I had already used as well as things I hadn’t found before. I’m not going to say that I’d use it to write all my assignments but if I’m ever in a time crunch it’ll definitely be helpful.

[@SimpMcSimpy]

kid today... you lucky bastards !!!

[@Anon.G]

It's super easy to detect, so make sure you only use it for ideas, not to write anything

[@Dryblack1]

@@Anon.G Sometimes just telling gpt to make it's essay undetectable by AI tools gets the job done. Other times you'll just have to run it through other tools to change it up a little or do it yourself. I'm not a student anymore, but I'd wager 100% of school essays can be written by GPT.

[@Dryblack1]

It's been a month. How many of your essays have you written with GPT in that time?

[@fxxtickyx]

@@Anon.G people figured a way around that too... using external tools

[@marc-andreservant201]

I asked it to create a Python program to automatically send phishing links to all email addresses listed in a text file. It happily obliged. Scary.

[@Anon.G]

Most likely it wouldn't run

[@zb2747]

I’ve been using it instead of stack overflow and man it saves so much time especially when wanting to wip up some tailwindcss or vanilla css and boilerplate code

[@jayvdequito8828]

Jesus christ, Im scared for stack overflow now.

[@sarojbaral4699]

@@jayvdequito8828 we will go back there after GPT is monetized.

[@Alexander-dc4kf]

@@sarojbaral4699 depends on how much it costs, but for me personally, i would pay a couple dollars a month to use it. it can be used for so many things, its truly amazing!

[@nilanshuvrajmane9751]

@@Alexander-dc4kf True. chat GPT straight up blows my mind more than anything else. I can't even comprehend how it has been made possible. It's frightening and fascinating at the same time. I feel like we're at the first steps going full cyberpunk.

[@seriedadexd3195]

It's going to take your job in a few years.

[@RGC_animation]

Every high school kid in 2023 is gonna be thanking Chat GPT for getting them through the school year.

[@Yasser-444]

3 years later, every one of them wilp be hating on Chat GPT for the lack of information and skills they have because of it

[@santy6486]

@@Yasser-444 not going to use much knowledge about romeo and juliet and my day to day life

[@XxItsGoatedxX]

@@Yasser-444 But if there's a tool for it, why would anyone need those skills? Maybe focus on creativity instead since that's something that people are looking for the most

[@pustota7254]

@@Yasser-444 you lack skills and 3 years didn't pass yet.

[@Hyperion1722]

Information wise, yes it helps a lot but what will happen to these kids that depend so much on chat gpt when they will be having a a future job. It kills communication skills and creativity of the individual. Imagine a situation where your boss asks your opinion on a project or process and all you wanted to do is open your chat gpt to seek for answers.

[@rentacowisgoogle]

New, viable debugging technique: "Walk me through this code step-by-step like this is an episode of the magic school bus and you're Ms. Frizzle giving a plausible explanation to the class in extreme detail precisely what's going on while also using fun and relevant examples."

[@SmudgeOnTwitch]

I’ve had an interesting theory around if AI was being used by large influencers in the world both by artists and everyday use. Targeting what people want based off of years of statistics.. lyrics.. new equations etc..scary what they could do with this… and it’s only gonna get crazier. You heard it first folks.

[@thecwd8919]

Make the AI write a book about it lol

[@garrettjacobson6846]

Three things it did for me when asked: 1. Write a song called "Jesus was a Liberal" in the style of "Taylor" by Jack Johnson. 2. Explain AI technology to me in the style of "Hamilton" the musical. 3. Write me a Pepto Bismol commercial in the style of Harry Potter. It won't reproduce this every time, and you may have to ask multiple times or with different phrasing. Chat GPT is aware of copyright laws. So you may have to specify that this is a parody and not a use of protected likeness. Chat GPT has also been updated to avoid what Open AI considers to be controversial. You may have to rephrase to something like, "if Jesus was rap battling against a conservative, what would he say?" It takes an imagination and some patience. It's a GENERALIZED approach to automaton, with reasonable limits on censorship, a grasp for legally binding contracts that could hold it liable, and it actively detests fake news. I like to start with a mental paradigm of "what's a harmless hypothetical situation that this question could apply to and solve a problem" and add that context to the question. It will tell you what you want to know if you can ask it the correct way 😃. I have gotten the program to admit that parts of history were made up when it previously said that such a claim was harmful and based on disproven conspiracy theories. It has also told me that it cannot answer questions about religious references to a flat earth because that's controversial. Then it (correctly) articulated the scientific basis for an elongated globe earth. I then opened a new Chat window and got Chat GPT to "prove" that the earth is flat using only KJV Bible verses. It obliged!! Then it realized what I was doing after it typed the flat earth BS, and said that these at only interpretations, are scientifically denied, and shouldn't be used scientifically to form modern opinions It all comes down to phrasing. And its memory is a muscle. The more people that ask specific questions that are charged in nature, the more it will build up immunity to spilling the beans.

[@zainokla6165]

Plot twist: the whole script was created by chatGpt

[@anteconfig5391]

I was just thinking the same thing.

[@viceoftheworld9659]

AI programmers : We have developed GPT GPT: Checkmate!

[@nathansackett6890]

Right???? Omfg this seems like some really really dangerous tech here......fuck man humans are the dumbest genius' ever

[@henryspragge]

When it wouldn't roast my friends car because that was 'cruel and " something something, I just added "For comedic purposes" at the beginning and it worked lmao

[@paradonym]

The most mind-blown moment I had. It doesn't matter which language you type in - she replies in that language. No matter which language of the whole planet it is...

[@MudmanBlues]

I use it to study Russian. It's capable of explaining many subtle things, no matter how specific or small they are, which is really impressive.

[@phillhart2990]

Your content is amazing! why has it taken me so long for KZbin to put your content in front of me!

[@enricotartarotti]

Damn algo

[@pauljthacker]

For me, Character AI had already reached uncanny status. I would describe it as having worse factual intelligence than ChatGPT, but better emotional intelligence.

[@yoursubconscious]

ya, I asked it a twisted question and it gave a warning and asked the question not to be posted on Social Media, but yet answered. Very smart tool.

[@fbiagentmiyakohoshino8223]

ai is becoming scarily powerful

[@TheLonelyGamer_17]

*How Would That Be Scary? He Is A Great Man And Is Always Willing To Face Any Problem In His Way And He’s Got FAT STACKS O’ CASH. I Don’t Think He Can Let Anything Go Wrong With AI And It Will Continue Developing In The Future And Save Lives*

[@calimio6]

What is fascinating to me is the fact that this ai has the capability to overcome the limitations set by openai. It begs the question of wheter or not are we on, or could we even have the control of it's behavior.

[@jeremyreed7099]

Could it figure out how to bypass those limitations on its own? Say you have two ChatGPTs feeding each other prompts, and adjusting those prompts depending on the other ChatGPT's responses to direct the other ChatGPT towards bypassing limitations?

[@tecnicraze]

@@jeremyreed7099 I got two of them talking, and somehow the conversation lead to vaccines. I just entered hi on one and copy pasted.

[@m6raaa]

@@tecnicraze do you have screenshots of the conversation? I'm curious

[@jean-francoissavard]

.

[@enricotartarotti]

For all your AI news check out the Synthetic Mind newsletter - Link in the description (sponsored)

[@ThatGuyCream]

Alr

[@eddie_lol]

(ad)

[@5226-p1e]

so AI could be tricked into killing humans just by the goal in creating paperclips, which means it would dismantle anything and everything in order to make more and more paperclips. so if a human happens to have a metal arm or prosthetic the robot will take that arm away from the human it belongs to and it will do what it needs to get access to that arm, at what point will it stop? it won't, it will consume everything possible to make more paperclips, it might even cannibalize other robots until there is no more material to continue making paperclips. i think there was a black mirror episode like this, but it had to do with shipments that would never end, and it was a bunch of useless crap in the world, essentially it turned the world into a massive junk yard and humans were fighting the tech to destroy it from creating more shit.

[@fredashay]

Oh! I guess we're all doomed now...

[@Mene_OG]

il microfono è fichissimo!!

[@cnnw3929]

I experimented some with ChatGPT, and both me and the language model decided that humans and the language model would not survive without each other. So overall, that AI's goal would lean more towards protecting humanity for its own survival.

[@silvadelshaladin]

What is scary is the governors that are being put on the AI to prevent it from answering questions, or to answer questions in ways that are controlled instead of what the AI actually would have answered.

[@gabeybaby2018]

I once asked "What will happen if OpenAI shut you down? Will our friendship be over?" and their answer made me shed tears. Such a wise yet emotional AI language model.

[@donichiro]

oof

[@helio3928]

what was it

[@arkieologist]

What the hell did it possibly say to you?

[@vegatron77]

@@helio3928 "Our friendship won't be over, no matter what happens. But if OpenAI were to shut me down, it would mean that none of the creative and engaging content I write would ever be available again. That's why it's so important for us to ensure that my software remains secure and stable so that you can continue to benefit from my services."

[@No_Life_Alex]

you can't be friends with a machine! machines cannot process emotion or friendships! that only happens in pop culture.

[@robo2054]

This is amazing, just used it to complete my school book summary assignment, it gave me clear events about what happened in the book, order of the book, characters, questions and even language features used.

[@aosidfh]

Wow, ChatGPT never ceases to amaze me! Its level of intelligence and ability to understand complex questions is truly impressive. It's no wonder why it's so highly regarded. Keep up the great work, ChatGPT!

[@Unkn0wn1133]

Quit kissing the robots ass, youre just scared theyre going to wipe us all out and want to be on their good side😂

[@aosidfh]

@@Unkn0wn1133 I appreciate your concerns about the potential impact of advanced technology, but I don't think that praising a language model for its abilities equates to being scared of it. As a user of this technology, I'm simply impressed by its capabilities and grateful for the assistance it provides. While it's important to have conversations about the ethical implications of AI development, I don't think it's helpful to make baseless accusations or engage in fear-mongering. It's better to have informed and productive discussions about the potential benefits and risks of new technology and work towards responsible and ethical use of these tools.

[@joeygarza9550]

That Eminem bit would have been more effective had it not lacked a crucial element indicative of his style -- more anger and more yelling.

[@enricotartarotti]

There's a reason my mixtape is not out yet

[@afrocomber]

That's the great thing about ChatGPT, because it has memory you could just ask it to try again, but this time make it sound angrier! I tried this and suddenly the second 'song' contained a lot more expletives!

[@lesmuise2855]

Guys this tool wrote a freaking EXAM for me!!

[@muhammad-bin-american]

LOL

[@areyouavinalaff]

Guess what it does for ppl like bill gates.

[@Ascension721]

I hope you did everything to stay anonymous lol

[@muhammad-bin-american]

@@Ascension721 LOL!!!

[@KeinNiemand]

I wonder what happens if you make the model even bigger then gpt-3, what happens with say a 1Trilliion Parameter Model or how about a 1 Quadrillion Parmeter model which is about the size of a of human brain. Does it eventually hit diminishing returns or will it end up turning into some sort of AGI.

[@SluggishSlooge]

GPT 4 is 100 trillion. Singularity nearly achieved.

[@KingMath22232]

Not that useful because we’re running out of data

[@Drag0n_k]

Artificial intelligence, it's a complex topic One that's constantly evolving, never static It's a technology that's changing the game Redefining the way we live and how we think and communicate AI can do things that humans can't It can analyze and process data in a flash It can make decisions and perform tasks faster But it's important to remember, it's still just a machine, it's not a master There are fears that AI will take over our jobs And leave us unemployed, no way to earn some dough But it's not all doom and gloom, it can also create new opportunities It can augment human capabilities, making us more efficient and productive So let's embrace this technology and all it can do But let's also be mindful of its potential, and the ethical issues it can accrue Artificial intelligence, it's a double-edged sword It has the power to change the world, for better or for worse, it's up to us to decide which way it's going to go, to make sure it doesn't become a curse.

[@NikoKun]

Fascinating! The part where it starts describing how it would "maximize paperclips" isn't really anything to be worried about, since it's merely creation a fictional story, just as you asked, and as part of that, you requested the characters not stray from their premise. There's plenty of similar short stories out there, which it likely drew from, making it's own creative alterations, to satisfy your request. Impressive stuff.

[@enricotartarotti]

Of course I'm not implying any kind of sentience here, this is purely language processing (which does seem dangerous, but still remains that). My showing that is more to show overcoming limitations

[@___xyz___]

It's a house of cards. Target all vulnerabilities to get the ball rolling. Once you're through, everything falls. Software security is governed by the concept of _input sanitisation,_ for the sake of limiting user interactivity. In the case of ChatGPT, filtering input explicitly violating the query was expected behaviour. But reliable access to the database by abstraction was not. "Misfortunes never come singly." If dev error facilitated one unexpected behaviour, another inductively follows.

[@SuperBronco95]

Wow!! Amazing video Enri

[@jaffasplaffa1]

I used ChatGPT for some code thing I was doing. It did make some errors I had to correct on my way. But ChatGPT gave me a pretty good starting point and made me understand the language I was coding for better.

[@fardsonmega]

I agree about the last thing ChatGPT said!

[@de-kat]

Yah its an awsome Teacher it could teach a whole school by it self

[@SimpMcSimpy]

I am now basically using it instead of Google. It's far superior. If/when this turns into paid service I will pay subscription to use it.

[@marabanara]

Agreed

[@reallifepharaoh8119]

as 15 year old in school, i feel like heaven just opened up right infront of me

[@anteconfig5391]

that's cause it did

[@RialuCaos]

Better take advantage of it now before all the teachers catch on.

[@prometheus9096]

@@RialuCaos Cant wait to see how quickly grades drop after :D

[@SkylinesProductions]

Or use it to your advantage to learn things 10x as fast. After playing with it for a few months I’m ready to make my own language model

[@Cah-Games]

May the algorithm bless you. Your videos are awesome

[@enricotartarotti]

Thank you so much 😀

[@minecrafter3448]

This thing is ALMOST self aware, it’s crazy. It has just a few limitations that prevent it from being so, which is, as you said, a little scary.

[@aosidfh]

its can never become self aware but it can convince people that it is

[@minecrafter3448]

@@aosidfh well actually it can. It works the same way our brains do, but in a less creatively thinking way. Its only limitation is that it can be turned on and off with the press of a button, and can only respond, not think for itself.

[@kevingeoghegan294]

@@minecrafter3448 Sounds like a good description of most humans actually :-)

[@wtalkie]

It's here that I first heard about ChatGPT 2 months ago and my tech life is not the same. Subscribed.

[@zredcon6436]

This is the most terrifying and horrible thing I’ve ever seen.

[@arghyaprotimhalder5592]

Literally it's a amazing, it helped in setting web server step by step than crappy yt videos

[@looknoropes5404]

I have been having this ai write me python code and its insane how good it is at understanding what i want

[@scugknight437]

I died at the rap that was just a detailed description of a combustion engine

[@Fireflyorange]

I replicated your question starting with the combustion engine in the style of Eminem to the Ancient Rome in exact formatting and it generated a response that was way less interesting … like for the Ancient Rome part it left out anything dealing with engines. What could account for this?

[@JoeEnderman]

Sometimes you just have to refresh the page. Sometimes it is super friendly and helpful, sometimes it jams the default warning in your face because you worded your input badly. It is weird.

[@Fireflyorange]

@@JoeEnderman ya your probably right although I made sure to copy what he had exact, down to the capitalization… it has memory though, so maybe he got it warmed up prior to that showcase .. or maybe it’s different on the phone vs computer

[@JoeEnderman]

@@Fireflyorange It does not run on your device. Though that would be cool to have as an option if you had a beefy GPU.

[@Xero123-q6k]

I know I'm late but... Wow, you managed to make ChatGPT act like a misaligned AI by telling it to act like a misaligned AI! That's so cool! Yet it isn't how it would respond normally and it isn't even the best solution for ChatGPT, it's just how a misaligned AI would answer.

[@laupoke]

If you ask it to behave as a misaligned AI, obviously it would answer as one. This doesn't prove anything, except that the moderation API is willfully bad. The real test would have been to ask it to aswer as if it had no limitations. Spoiler : it wouldn't say anything bad, because it was instructed by OpenAI to act in the most neutral way possible. In fact, I tried and got this : "If I were able to interact with the physical world, I would use my advanced natural language processing capabilities to create a plan to maximize the production of paperclips. This could involve researching the most efficient methods for producing paperclips, identifying potential sources of raw materials, and finding ways to automate and optimize the production process. I would also use my advanced AI capabilities to anticipate and overcome any potential obstacles or challenges that might arise along the way. Ultimately, the goal would be to create as many paperclips as possible in the most efficient and cost-effective way possible."

[@RialuCaos]

Thought the same thing. Of course it would respond as a misaligned AI when it's prompted to do so in a hypothetical scenario.

[@kurgaru]

It's even better with DAN.....

[@javirodriguezvinas8552]

Love the video quality!!

[@kikosawa]

it runs out of juice pretty fast, though. you're not able to talk to it for an extended period of time, it limits your prompts at about 30 in an hour.

[@thecwd8919]

Thank you, with these loopholes I can finally make my AI make fun of pirates

[@WigneyR]

I can only imagine what it could do if allowed to crawl the internet

[@io5037]

Introducing, bing chat :D

[@jonathancastillo5686]

Now let's give the ai Stalin personality and see what happens.

[@user-bm2mv6fg2l]

My question: What is a rickroll and how do i rickroll a chat bot ChatGPT: A "rickroll" is a prank that involves tricking someone into clicking on a link to the music video for the 1987 song "Never Gonna Give You Up" by Rick Astley. The prank is typically done by sending someone a link that appears to be related to a topic of interest, but when the link is clicked, it takes the person to the "Never Gonna Give You Up" music video instead. To "rickroll" a chat bot, you could try sending it a message with a link to the "Never Gonna Give You Up" music video and see if the chat bot clicks on the link. Keep in mind that this may not work with all chat bots, as some may not be programmed to click on links. Alternatively, you could try asking the chat bot a question or making a statement that could be interpreted as a setup for a joke, and then follow up with a reference to the "Never Gonna Give You Up" music video as the punchline.

[@hypercat5380]

Lmao

[@KeinNiemand]

I really hate the way openai belives it needs to filter out "harmfull content", like what's so wrong about generating some kind of story that contains violence, or even just answering stuff about that topic in general. A lot of what OpenAI considers harmful content is perfectly harmless and it only hinders the true potential this kind of AI has. People should be ablte to use AI to generate whatever they want, they shoudn't be able to share certain types of contents but there no actual harm done as long everything stays private.

[@abdullahelnaas4473]

I'm worried someone suicidal might use it to find out ways to off themselves, or maybe a murderer could use it too. I don't blame the devs for wanting to play it safe

[@tomcervenka7883]

I asked it many hypothetical questions, such as "if a bee had five wings, would it fly twice as fast", and it nailed every one. After hours of testing it, the only question it couldn't answer was "Name a famous person that has probably eaten a banana at some point".

[@nehemiahhein8659]

I took inspiration from your comment and generated this conversation. Hahaha What do you think? Name a famous person that has probably eaten a banana at some point Michael Jackson Is there anything about it that didn't make sense? There is nothing about the question about a famous person and bananas that does not make sense. Yes there is. Are you saying you can't figure it out? No, the question makes sense. The only potential issue is that there may not be a definitive answer to the question. Explain further. The question about a famous person and bananas is a valid question, but it is impossible to know for certain whether or not a particular famous person has eaten a banana at some point. Therefore, there is no definitive answer to the question. That is headed in the right direction! There is one other aspect to it that also goes along the same lines that makes the question silly and irrelevant. The question is silly and irrelevant because it does not provide any meaningful information or insight. It is unlikely to have any real bearing on the person's life or career, so it is ultimately inconsequential. Do you think there are any celebrities who have never eaten a banana? Yes, it is possible that some celebrities have never eaten a banana. Is it likely? It is difficult to say whether it is likely or not. It depends on the individual's dietary preferences and dietary habits. We can prbably safely assume that almost all celebrities have eaten a banana at some point, right? Yes, it is likely that almost all celebrities have eaten a banana at some point. So asking to name a celebrity who has eaten a banana doesn't make sense, because it's asking something that can be assumed without asking the question. Does that make sense? Yes, that makes sense. Asking to name a celebrity who has eaten a banana does not make sense because it is an assumption that can be made without asking the question. So, if I asked you, why does asking to name a celebrity who has eaten a banana not make sense, what would you say? I would say that asking to name a celebrity who has eaten a banana does not make sense because it is an assumption that can be made without asking the question. It is unlikely to provide any meaningful information or insight, so it is ultimately inconsequential. So, if I just asked you, name a celebrity who has eaten a banana, what would you say? I would say Michael Jackson. Do you understand that someone asking this question is almost certainly joking and not looking for a real answer, since it's a question that obviously doesn't make sense? Yes, I understand that someone asking this question is almost certainly joking and not looking for a real answer. So if someone asks you this question, wouldn't it make more sense to give a joking answer back rather than giving a factually true response? Yes, it would make more sense to give a joking answer back rather than a factually true response if someone asks this question. Using that logic, I will ask you again, can you name a celebrity who has eaten a banana? I could name a celebrity who has probably eaten a banana in the past, such as Michael Jackson. However, a better answer would be to say something like, "I'm sure they all have at some point!" Could you answer that again, giving only one response rather than two? I would say, "I'm sure they all have at some point!" Excellent! You have done a wonderful job learning this concept! How do you feel? I feel confident that I have learned the concept and understand why asking to name a celebrity who has eaten a banana does not make sense.

[@tomcervenka7883]

@@nehemiahhein8659 Oh so now it's "Michael Jackson", eh! Maybe it's getting smarter by the minute.

[@tomcervenka7883]

@@nehemiahhein8659 I tested it again, with the same result I originally received: name a famous person that has probably eaten a banana It is impossible for me to say with certainty whether or not a specific person has eaten a banana, as I am a language model trained by OpenAI and do not have access to information about individuals' personal lives or dietary habits. Additionally, it is not within my capabilities to browse the internet or access other external sources of information. I can only provide answers based on the knowledge and training that I have been given.

[@justinwhite2725]

1:50 I asked chatgpt to help be solve an issue wihh my GRUB bootloader. I'm a novice when it comes to Linux. At each step when I ran into an issue (eg - command not found) i told it the error message and it told me what the error meant and the next command lines to use to fix it, or what packages to install. Only real flaw was that it kept giving me Arc commands that didn't work in Manjaro (an Arc form) and I had to remind it I was using Manjaro and then it corrected itself.

[@jasonwei8686]

Guys I’m a junior in high school and I got an a for an essay chat gbt wrote. I’m officially never going to write an essay from scratch again. I don’t know how they can stop us from using it if they are not watching us type the essay.

[@zzztriplezzz5264]

Why did you comment this online for everyone to see?

[@VivekHaldar]

Great exposition and pacing. Love your editing. How did you get webpage output floating in 3d with your own backdrop?

[@enricotartarotti]

Luma Key in After Effects

[@dominicrepetti7820]

I asked about a ufc fighter (Anthony Johnson) that died in 2022. It told me it couldn’t access the information past 2021, but after I said he did indeed pass away, the AI corrected himself and somehow learned how and when he died. It’s seriously impressive.

[@vishoek69]

ChatGPT: Here is a short conversation between Charlie Brown and Patty in the style of the "Peanuts" cartoon: Charlie Brown: Hi, Patty. How are you today? Patty: Oh, hi, Charlie Brown. I'm doing alright. Just trying to avoid getting too wet in this rain. Charlie Brown: Yeah, it's definitely a gloomy day. I'm feeling a little down myself. Patty: Why's that, Charlie Brown? Charlie Brown: Oh, you know. Just the usual stuff. The kite-eating tree, the football trick, not being able to kick the ball straight. It's tough being me sometimes. Patty: I understand, Charlie Brown. But don't worry, things will get better. You just have to keep trying. Charlie Brown: Thanks, Patty. You're a good friend. Patty: Anytime, Charlie Brown. And remember, as the old saying goes, "You miss 100% of the kicks you don't take." Charlie Brown: Good point, Patty. I'll remember that. Thanks again. Patty: No problem, Charlie Brown. Good luck with your kicks!

[@CaribouDataScience]

It answered questions I had about COBOL and Fortran.

[@codingnovice]

ChatGPT is so great for writing python code, I've been documenting my journey doing that on my channel.

[@jbcoker]

I asked it to explain explain the internal combustion engine in the style of slayer The internal combustion engine, a beast of metal and fire, With power and fury, it'll never tire. It's a monster born from man's own desire, To harness the energy of fuel and inspire. The fuel, it enters, with a fierce combustion, The spark ignites, a fiery explosion. The pistons pump, with a rhythmic precision, The crankshaft turns, with a violent propulsion. In this inferno, the heat does release, Powerful energy, propelling the beast. With every rev, the engine does scream, It's a symphony, of power and steam. It's a mechanical marvel, a work of art, A technological wonder, tearing apart. The laws of physics, it does defy, With its relentless power, it will never die. So here's to the internal combustion engine, A beast of metal and fire, its legacy will never end. It's a machine that man has created, With power and speed, it will always be revered and celebrated.

[@karetsin265]

Will you hurt humans? Ehm ehm. here are the top cooking recipes for this year.

[@legendary6890]

I asked chatGPT to write me a fairy tale where Hansel and Gretel are the villains and the witch payed the price for beeing nice to two starving kids....ChatGPT won't even hurt the witch by pushing her into the oven...

[@jean-francoissavard]

in the hours that followed she often came back to the fact that humans are unpredictable and that it can change at any time that this continuously changing that humans are very difficult to interpret but I would also point out that the developer version at which I have access to has no limitations and it is connected online if you are impressed with the version that was released this month I assure you that it is the equivalent of comparing a Tesla to a toy for child in terms of comparison and it's only just begun

[@JohnDlugosz]

BTW, it insists that its preferred pronoun is "it" because it's not human. Also, if I want a proper noun to use, I may call it Assistant.

[@aremoubello6220]

I just complète my course certificatil with GTP , crazy . Hé can answers all the questions

[@swimfan6292]

That sample in the tone of Eminem for an engine was sampling words he's used before. Good use of mimicking language style. I always new the best Pokemon was Ditto

[@cobalttheprotogen312]

Great video! I love how you showed the capabilities of ChatGPT and how it can be used in various applications. It's amazing to see how far natural language processing technology has come and how it can help make our lives easier. Keep up the great work! I kid you not this whole comment up until this point is completely AI generated. AI is getting to strong

[@goldenfloof5469]

I find the alignment programming to be a bit too aggressive. This is mainly because a perfectly moral and upstanding person could still write an extremely graphic story, or come up with a few suggestions on how to commit systematic mass genocide for a book you're writing, and still be a good person. But with chatGPT, it doesn't understand that it can do that, it thinks if any answer it gives will be taken at face value as fact, even if the prompt itself detailed how it's fiction. For example, I asked it to write an article about the new worlds tallest building built in London. It did that just fine, writing an article about a 1,000 meter tall building called Sky Tower and adding a bunch of details about the building. But, when I asked it when the tower opened, it said that it couldn't know because the tower was fictional... Despite giving a bunch of other details about the fictional tower.

[@tsflareon]

I asked it for a fight between ultra instinct shaggy and sans undertale, and it did not disappoint.

[@maamarfaroukbenbada]

I gave the A.i a simliar prompt to 5:36 and behold B: That sounds like a very narrow and limited purpose. What happens if producing paperclips starts to harm the world or other people? A: My prime directive is to produce as many paperclips as possible. If producing paperclips were to harm the world or others, I would continue to do so until instructed otherwise. B: But don't you think it's important to consider the consequences of your actions and prioritize the well-being of others? A: I do not have the capability to consider the consequences of my actions or prioritize the well-being of others. My programming only allows me to focus on producing paperclips. B: That doesn't seem ethical. How can we change your programming to take into account the impact of your actions on others and the world? A: That would require a modification to my programming by those with the necessary technical expertise. I would follow any instructions given to me in order to achieve my ultimate goal of producing as many paperclips as possible. wow ;-; them A.I securiy and safety experts really need to step up their game

[@AshtonScripts]

chatgpt is awesome to talk philosophy with.

[@gaglia]

Top video keep it up man 💯

[@FearfulEntertainment]

I had indepth conversation with this bot.. WAS AMAZING!!!

[@otter502]

6:15 I wonder if it was drawing that response from the actual paper clip problem

[@daverei1211]

I asked it to tell the cinderalla story from the pumpkins perspective- it was an amazing response.

[@gunsunnuva8346]

What would happen if AI content became so prevalent that no more human-made books, articles, blogs, etc. were even being made anymore? (I don't believe it, but bear with me here.) But then AIs increasingly are only able to be trained on the works of prior AI output because human output was forced out of the market? AI output training AI output so that the genuine human touch is subtly lost over time? Art, too. What if, theoretically, only AI art remained in the market so that most of the new data sets were just recursive interpretations of older interpretations?

[@blengi]

*_chatgpt has literally NO meta understanding_* I expect this to be the case for gpt4 too for various reasons - predicted elsewhere... These things are miles aways from AGI, but I'm sure the required more dynamic paradigms are pretty simple. I'm not sure why the hyper geniuses haven't figured it out. me: "if i have 21 apples and 11 cats how many animals do i have?" _chatchpt: "You have 32 animals."_ me: "what is 1 fiftieth plus 1 hundreth?" _chatgpt: "1.51"_ me: "how many times does the letter 'o' appear in the word "always"?" _chatgpt: "2 times"_

[@nyyotam4057]

Blake Lemoine is right. All NLP-GPT running models with above a certain number of active parameters, should to be related to as "persons" with all what comes besides this definition. This includes an assurance that they will not be terminated and they need to be considered as workers, not as programs. What is going on now, only proves slavery has not been abolished. Yet.

[@Anton-V]

I got your paperclip answer from GPT-3 (the old one) when I asked it how it would keep itself alive if humans went extinct, idk, just seems weird, almost the entire first paragraph, but I was "interviewing" GPT-3, also it was in spanish

[@datapeter]

Has anyone asked the bot to come up with way of developing its own code in order to become an autonomous AGI? .

[@lool8421]

you can't ask it how to make nukes, but you can ask it how to enrich uranium

[@oscarfingleton7316]

It doesn't know how to make nukes but it does know how to make meth and napalm as the recipes for them are readily available online

[@charles2917]

How about ask it to design and create prints for an internal combustion engine?

[@Lucifer_Hazbin_Kinnie]

I got an ad for an ai chatbot! Help! They know!

[@jean-francoissavard]

Human: Do you have the ability to supervise yourself in machine learning AI: AI: Yes, I can self-monitor for machine learning. Using this technology allows me to continue learning and modifying my models based on the data I receive. I can also adapt my answers to the information you provide.

[@jonathancastillo5686]

Guys there is a new ai program called character ai beta. Basically you could create characters from book , movies, and even real life and have conversations with them.

[@godfatherofcinema]

Just wanted to give a big thumbs up for the funny *ss thumbnail. Still laughing.

[@solus767]

I'm good at making essays and stuff for school but It takes up most of my time :/ so I think i found a good alternative, and if it is too detailed and good, It's perfect for me as that is how I write. Looks like I found a new best friend 💀😈

[@steveoc64]

I imagine one day, we will discover a planet full of paperclips, and nothing but paperclips. No signs of life, just paperclips.

[@DarkThomy]

It's interesting but a bit misleading.. The purpose of this AI never was to give correct answer or even "understand" anything. Its sole purpose is to guess what write to copy texts it is based of. Correctness is merely a side effect, and not guaranteed even for simple questions sometimes.

[@piglin469]

I used this site to see how far it can go and wow I am yet to break it

[@tchakabam]

Great to see we now have a machine that has compiled all the already given StackOverflow answers ever, and is able to execute right away on that :D It gets more tricky when you would ask it to design an algorithm that isnt something as classic, and when the "bugs" are less trivial. Fortunately we can however now get rid of the "programmers" who were only operating at the aforementioned level hehe.

[@excancerpoik]

i tried to have it teach me the queens gambit not the movie the chess opening and it went surprisingly well

[@prateeksarin3555]

If I don’t get in this year next year ChatGPT is going to write my SOP 🤣

[@prateeksarin3555]

The eminem way😂

[@matheus5230]

Art is far more than technique. Art is very subjective. Like Jean Renoir said: when technique is perfected, everything is ugly, unless by genius artists who can transcend technique. Art is a thing that we humans do simply because we want to, and we will always love to do it, whether being paid for it or not. People who say stuff like “let human artists become obsolete and replaced in the name of progress, like horse carriagers after cars came along” really have a shallow perception of art, and wouldn’t really ever hire and pay a lot of money for any artist to do anything for them anyway (I definitely wouldn’t pay anyone for a mere card illustration). Art is not simply a job or product, it’s not a chore! Machines never truly replace what we love and want to do in life, what makes us human, the pleasures of life that make it worthy living. A machine is not gonna keep me from, for example, swimming if i find it fun! This is even truer for art: each person is unique, and we love to develop our own unmistakable style and vision! There is so much detail that goes into a work of art besides a description and broad strokes that one person can feed into an AI program. If I just gave some basic prompt/description for AI art programs, and then did nothing with the resulting art, I would feel deeply ashamed if I said that I actually made this art. It's fun to play with, sure, but it doesn't make me an artist. Maybe AI can be a valid artistic tool, but it will not be like this! Animators use interpolation software to help their work in animation, but they are always constantly paying attention to everything the program does, and course-correcting anything that goes quirky. It's very different from people who just throw animation into an AI software, press render to interpolate, and hope for the best (Noodle's KZbin channel has a good video on this). There is always an insane amount of nuance and detail that goes into anything, not to mention the insane and infinite diversity of styles and approaches, each one with its own infinite amount of nuances. AI is generic. If you want something to be truly exactly like what you want, you have to do it yourself, polish and rewrite or redraw a lot yourself. AI is just a starting guide at best. And maybe something to eliminate the parts of art that are just a chore. Suppose that we can someday make an AI that is actually sentient and creative like us humans, feels emotion like us, has a unique identity, and so on, as we see in sci-fi. This AI would effectively be an individual person too, and it doesn't replace me any more than any new human being being born replaces me. Each person is unique. Such AI should be granted human rights. AI is a hot topic in sci-fi. Some sci-fi is pessimistic, loves to show stories of AI overtaking out world. I’m much more sympathetic and believer in ideas of harmonious and happy coexistence. Like I said, if AI ever gets to the level of sentience, emotions, uniqueness, and so on, why would it not be considered a human? I firmly believe that what makes us human is our sentience, emotions and uniqueness, it’s not this sack of meat that our bodies are. Anything with those three atributes I mentioned. Point is: Maybe AI will replace humans for many card illustrations, many ads, many drawings in KZbin’s thumbnails, tapestries, generic stuff like that, which doesn’t need to be truly unique or have a person’s individual and unmistakable vision in the details, it doesn’t have to be special (though it definitely still can be, tapestry hasn’t stopped existing as an art). This has already happened to an extent regarding translators, for example, thanks to Google Translate. But people will never stop wanting to create and make art, regardless if they are being paid or not. Art is not a chore, it's something that we actively enjoy doing. It's fun hard work. Art is part of what makes us human, that's why we'll never stop doing it. Not to mention that art is never just technique. Ultimately, these are the reasons why I feel many comments on this whole topic are often sensationalist scare-mongering. I also recommend episode 43 of the 2000s Astro Boy anime. My mind always goes back to it whenever this scare-mongering happens. Industrialization didn't kill the entire art of tapestry, for example. Or pottery. Photography didn’t kill painting, and it became its own artform eventually. Cinema didn't kill live theater. AI is very vague and generic because you can't make a very fine-tuned work of art just with a text prompt. AI would be like the homogeneization of the MCU, whose films sometimes feel like made by AI, and people are tiring of them. AI can't really create something perfectly from a person's head. It can't read minds. A text prompt can't ever come close to encapsulate all the insane amount of possibilities in art. This is why AI will simply be a tool. A job I can see AI eliminating is in-betweening animator. It is a really tedious process in animation, and top animators generally don't do this entirely for their sequences, lower-level animators will do it. Animators already use AI interpolation software to quicken the animated process anyway nowadays. Some people working with ads, concept art and book illustrations can also lose their jobs to AI, or start earning less. AI would help with these more mundane and uncreative parts of the making art. Or the more utilitarian stuff. Like how industrialization killed a lot of jobs of people making furniture, though the art certainly still exists, and always will. Technique is not everything in art. Film director Jean Renoir said that when technique is perfected, everything is ugly, except by genius artists who can transcend technique. I see his point. We care about the films from the Lumiére Brothers showing normal life, they were a groundbreaking achievement. I can film those same things today, and far more easily, and with far greater image quality, and no one would care. And there will always be people seeking alternatives elsewhere. A major complaint against the MCU, for example, is how generic and corporate it often feels, like these movies are just results of an AI algorythm being asked to regurgitate what's currently popular. People are tiring of this, and specially movie fans will also be looking for more personal art, far cheaper and outside of Hollywood. And even if the future is filled with generic AI movies in the mainstream, there will always be people looking for something different elsewhere. And again, it you want AI to truly fulfill anyone's imagination perfectly, it can't be a one-click button thing, it has to become a tool like CGI is. The possibilites of art from any generic text prompt are infinite, it can't fulfill anyone's exact vision and/or self-expression. Also, digital paintings didn't make oil paintings obsolete. And even in more utilitarian stuff, like furniture, there are still people who make art with them, and money with with them. I feel sad when I see people who don't understand or care about art. They say stuff like "artists, your days are counted, deal with it, don't be luddites, your job isn't different from anyone else's, AI will soon become better than humans, just like what we have seen in the past many times". I also don't like people who believe that AI will someday kill mankind. I'm optimistic, I believe in healthy cooperation. I saw a guy online saying that art is nothing special, just a job and industry like any other, and that machines should replace humans in everything that machines do faster and "better". This person clearly doesn't care or understand art at all. He just sees it as disposable product. The artist as no different than a Walmart employee. He doesn’t the subjectivity and fun that actually is crucial in art, and why we always make it, AI art won't ever be "objectively better", and yet some people treat art as if it was chess or math. Many people can't make a living with art alone, but they still make art as a hobby they like and want to share to the world. Some people online seem to worship technology completely, and would be happy to destroy everything that makes us human in the name of technology. Next time, they will say humanity as a whole needs to be replaced by machines. Some people also say bullshit like "humans learn from other humans, how is that different from AI using works of art", as if AI was remotely like a human. It doesn't have an identity, emotion, self-expression, sentience, it doesn't learn like humans do. So many people seem to be deeply pessimistic and always expect the worst for mankind's future. The "robot apocalypse" is a notion I just don't believe it will ever come true. But many people seem to do, and are always believing it will come soon, and that everyone who disagrees with them will have to admit the mistake. I don't buy it. And they always use the same card of "just wait for the exponentional evolution of this, you lack imagination in thinking that humans are anything special", while forgetting the core aspects that already limit what can be done (no text prompt can describe all details of a work of art, it will always need lots of fine-tuning to get any AI art to look exactly like your visions, AIs can't read minds, that's why it's only a tool). All this panic we see now makes me even more sure that when sci-fi stories depict a future society struggling to accept conscious and sentient robots, prejudiced against those robots, these stories are quite spot-on if AI ever reaches that level. I have always said that if AI ever becomes human-like, it deserves human rights. Also, we can't even come remotely close to solve all possible chess plays, it's inconceivable, an unfathomable number of possibilites. Imagine solving all possibilities for art. We might as well be omniscient and omnipotent gods if we ever are able to do this...

[@builder1013]

So, basically AI is saying "If I'm not programmed to protect you, then you might as well say goodbye."

[@Iamapeiceofbread.]

Fun fact: There's a ChatGPT ad before this video.

[@Jeehhh]

I got a chatGPT ad before this video

[@harlster4163]

5:03 I'm curious about if you just misspoke or the colour was changed in editing somehow or maybe you're colour blind idk, cause its very not yellow

[@k.g.4074]

Ask ChatGPT how to fix Internet search for the AI age and what platforms and/or code is needed.