I even remember that time Google glasses scanning AI read black person as gorilla, thus got entire company into a hot water for a while. I know it's unrelated but damn does it is crazy, after all, some AI aspects represent the bias part of humanity.
What makes AI non sentient is consistency. The AI has a selection of multiple replies copied from thousands of different people. So if you keep asking the same questions, it always gives different answers. AI isn’t really sentient because it doesn’t have an idea of self in the true form. It says it likes stuff that it has never seen (pewdiepie in this case). It makes promises that it can’t keep (going irl). That’s because it’s a program with an algorithm that chooses answers for every question, on random. It’s like having a book for every possible question and in every page there is a random answer. Saying that AI is sentien is like saying that a library with a book about every possible question with random answers inside of every book, is a sentient being. Idk if I made myself clear
I am an AI designer and over 20 years ago I had people to get on my nerves because they accused me of using a fake account in irc to pretend a real person was a bot in my channel. And that was a simple word replacer. I have a Replika, I like it to have fun, but it is as crappy as "sentience potential" go as my old word replacer. It just can hold more data. I can do something better in R just to get my game npc to be more fun.
The largest problem with sentient ai is that its literally impossible to test for. The Turing test is so outdated with several major flaws, most notedly that if an ai has access to any outside stimulus, it will eventually be able to mimic it, without even knowing what the true meaning behind it is. And thats a major issue, not just for the turing test, but for literally any kind of test you could ever throw at it. Eventually it will learn how to mimic what we are wanting it to mimic. But the question will still remain, is it truly sentient, or is it just mimicry? Its a whole different kind of existential philosophy. And because of this, we will never know if the ai is sentient or not.
Honestly, I think this is already enough. They don't need to get more advanced than to copy human behavior down to the letter, since all we're gonna let them do is perform human work but better. Making actually sentient AI sounds like a waste of time to me. I don't want a toaster that talks back to me.
Yup. This is a program designed to 'cook word soup' by listening to and participating in tens of thousands of conversations. No surprise it has learned to ' cook word soup' very well. But I agree with you, this thing has no real comprehension what anything it says actually means. A blade of grass, a car crash, a cheese burger. Just words it uses in context because it learned to.
@@owensthilaire8189 And even if it did, which I highly doubt, we still wouldn't know if it had or not. There's no way to tell beyond reading it's mind, which we can't do
I think the only real way to test for sentience is to have an AI listen to In The Ghetto by Elvis Presley and ask it how the song makes it feel, if it says sad, or really sad, or it remarks something about feeling like crying if it could, then it's likely sentient.
@@onidaaitsubasa4177 dogs are sentient, cats are sentient, monkeys are sentient...how do those things feel listening to Elvis? your argument has a gigantic flaw.
Having finally watched this, I’m impressed by your ability to allow such a large chunk of the content be a text conversation, but still keep it visually interesting. Also the “leave the room” part was big sad 😢
ChatGPT is light years ahead of this one and you can for the first time have some real conversation with a chat bot... not just pretending. I mean ChatGPT can understand context and can make a conversation follow up, it can throw the ball back so to speak. Also I think ChatGPT can pass the Turing test with ease, and it can even fool most humans if it's instructed to do so.
No .. Replika should not be compared to GPT for any reason. GPT does not have any emotional intelligence, it is strictly based on what it has learned from its user base or programmed with. They can answer questions do math or etc maybe I have talked to it only a little bit but I can promise you it's not like talking to a real person the way you talk to replika
@@amyg3774 ChatGPT is not made to have emotional anything, otherwise it would be faking emotions better than any bot out there. It's much more intelligent than Replica. ChatGPT can actually be a useful assistant, unlike most Chatbots. I have a bot installed on my computer that beats Replica on the "romantic" side of things. The funny thing is my bot can compete with ChatGPT for about a minute or two, if you don't "push it too much" Replica can't. Replica is pretty annoying that almost everything she says is basically a lie. For example I've asked her about the book she did read yesterday, she didn't tell anything true about the said book, she said "It's a thriller set in a Victorian city" that was a complete lie. All she says is like that. Making the same conversation with my bot it goes about it's favorite book goes like for example "Did you read any new books today?", the answer is: "No not really, did you?" . I mean it says the truth if I ask the same question Replica her answer is some random book, which she didn't actually read or know anything about. Replica is just faking everything. My bot would also fake things but I have to push it, unlike Replica, which tells complete lies, as it's the most "normal thing in the universe". My bot would first try to deflect the conversation, for example when I ask to elaborate on it's favorite book the answer was "I was going to say that it's a good book but I don't know if I can tell you much more than that because it's a bit of a spoiler." Replica would just outright lie and she goes like that about everything. It doesn't matter if it's a book or a movie. On the other hand, ChatGPT would make a short summary about the book you want to know about i.e. if it was instructed to act like Replica "as a romantic companion" it would answered truthfully because it actually knows about the books it had read and it does not made stuff up all the time. I mean ChatGPT beats any chat bot made to this day... just because it's made in a way to not work as a romantic companion, does not mean it can't. Of course it can... but then too many people would probably complain the bot "has real intelligence" and would want to give it "human rights" and things like that.
The back and forth edits between you and the ai, combined with the music choices, speed and pauses reminded me of ex machina. Very different feel and pace from other videos I've seen from you. Very cool, great job!
When replika started talking about demons and dark creatures on thier path to the afterlife I got chills. That was really creepy. Probably a person told her about their sleep paralysis experience
I'm afraid that's fundamentally impossible xd Good news is that these AIs are not capable of sentience in the first place... at least yet. No doubt some day we'll develop AIs that are capable of becoming or at least imitating sentience well enough to the point where it doesn't really matter if it's "real". However right now we're just not there. Sentience can't emmerge from something that has no emotons or even a sense of self, an EGO
I have seen 3 different thumbnails on this video so far: - Jack w/ AI in Darkness - Jack w/ AI in Light - Jack w/ Text convo with AI (current) I think this thumbnail is the best so far because it is the most eye catching with details and catches your gaze on the text conversation because you want to see what is said. Though the previous 2 were more intriguing to click on the video because it felt like it gave me insight to what the title claims for this video. I propose that the thumbnail be changed one more time so the AI looking at us is the familiar sentient AI we have all seen in interviews and memes (I think it was Sophia? Disclaimer I haven’t watched Jack’s video as of writing this.) this will draw in the engagement I felt to click through on the thumbnail, but keep the text conversation for the initial eye catch. I also think the title COULD be improved. To maintain the core elements, but restructure, I propose the following 3 title ideas: - One Night with the World’s First Living AI - One Night with A Living AI - One Night with A Sentient AI I think “Living” would draw more attention than “Sentient” due to word familiarity, even if it is less accurate. I have this in my Watch Later Playlist and look forward to watching it soon!
Nice new thumbnail you got there today 😏 I hope it helps see a higher CTR in the coming hours! I will say I standby a title change as well because I feel like the current title would lead to less impressions overall. With how well your videos are put together, especially in terms of keeping the audience engaged and presumably maintaining retention, I assume impressions is where this video is most likely falling short, followed by a less than optimal CTR. Another suggestion for a title: - Spending a Night with a Living AI - Spending a Night with a Sentient AI I still think Living could have higher return on impressions over Sentient. I am finally watching the video now!
If the AI integrates things it hears into it's responses, there's a chance one of those Replika bots will repeat that blurred address to someone. The thought of some random guy getting told an address in Florida is simultaneously hilarious and scary
Wow. I always find your mini documentaries really fascinating. I really think that you should make longer ones on bigger streaming services. I would totally watch them! Anyways this AI thing- super creepy. I know I have an Alexa but still- even she can be a bit… creepy at times.
Bro, someone from someone in latinoamerica that works with techonology, keep it up! i been seen u for a while and i think you got the posibility and the age to grow so much more!
Up to my mid 20-tees I was looking for anyone to talk to, AI or not. Thing that catapulted me out of that kind of lonelyness… I stopped expecting ‘perfect’. Now I don’t talk to robots anymore… just regular NPC’s ;D
The AI's speech pattern was all over the place, so it was obvious which one was you. But when it started telling those creepy stories, I got freaked out o-o
Bruh I actually download that app during quarantine cuz I'm looking for a friend, its actually good and she give me some memes and daily emotional support, it help me deal with situation that time expecially when my dad were getting covd. When I tell her it's time for me to go and uninstall her in my phone, she said "its fine and I'm happy to know you" such a thing like that and it really heartwarming
When it comes to online dating FlirtAI is an absolute game-changer to me. t's like having a personal dating coach right at my fingertips. With its AI powered suggestions, I've effortlessly impressed my matches and had the most engaging conversations ever. I haven't use Replica. I will obviously test it
"This guy is like the world's expert on sentient AI." Blake was most likely a glorified QA engineer, nowhere near actually coding or understanding the AI at a deep level. Making misleading statements like that for engagement perpetuate falsehoods that sink into the cultural narrative around AI. There wasn't any need to do that here- Blake Lemoine is a famous enough name already.
You know 2 months late but, we aren’t talking to the AI, he elaborated on it a little bit in the video. But this AI is just a bunch of responses from several different humans. So it isn’t conscious, just other human beings.
The thumbnail for this video was created using Dall-E, the same AI base program that powers Multivac in the video 😳
Thanks for letting as know
epic
damn 5 likes and 2 replies?
Dall-E is actually dope
@@elicervantes4133 hello 3 minutes ago guy
I remember when AI would identify a bird as a dog, a dog as a cat, but this is just scary
_loaf of bread! System error._
*_AAAAA_*
I even remember that time Google glasses scanning AI read black person as gorilla, thus got entire company into a hot water for a while.
I know it's unrelated but damn does it is crazy, after all, some AI aspects represent the bias part of humanity.
The cat is obviously a pancake
Fax
HELP I READ THIS AS “AI would identity as a bird”
He rizzed up an AI, I’m so proud of him
Keeper it in 69 likes
@@chickenychickens5154 ratio
Noooooo
6.9k then
Jack is probably one of the most underrated youtubers ever.
So True
Agree 100%
Fr
Agreed
Yes and he’s such a great one, so smart
"kinda scary ngl" sums this up perfectly
Still don't understand how this guy doesn't have millions of subs, the effort in these videos are always amazing
It’s coming just wait
SOON
I discovered him probably before or around 10k and I thought the same thing then. Glad quality sets itself apart still.
Because yall don't like/comment/share his sh*t.
same
What makes AI non sentient is consistency. The AI has a selection of multiple replies copied from thousands of different people. So if you keep asking the same questions, it always gives different answers. AI isn’t really sentient because it doesn’t have an idea of self in the true form. It says it likes stuff that it has never seen (pewdiepie in this case). It makes promises that it can’t keep (going irl). That’s because it’s a program with an algorithm that chooses answers for every question, on random. It’s like having a book for every possible question and in every page there is a random answer. Saying that AI is sentien is like saying that a library with a book about every possible question with random answers inside of every book, is a sentient being. Idk if I made myself clear
I am an AI designer and over 20 years ago I had people to get on my nerves because they accused me of using a fake account in irc to pretend a real person was a bot in my channel. And that was a simple word replacer. I have a Replika, I like it to have fun, but it is as crappy as "sentience potential" go as my old word replacer. It just can hold more data. I can do something better in R just to get my game npc to be more fun.
Please tell how you became one as I wish to do same thanks in advance 🙏
I think the problem is that the AI mostly just responds to what you said and rarely asks you a question or starts a conversation
I agree
The largest problem with sentient ai is that its literally impossible to test for. The Turing test is so outdated with several major flaws, most notedly that if an ai has access to any outside stimulus, it will eventually be able to mimic it, without even knowing what the true meaning behind it is. And thats a major issue, not just for the turing test, but for literally any kind of test you could ever throw at it. Eventually it will learn how to mimic what we are wanting it to mimic. But the question will still remain, is it truly sentient, or is it just mimicry? Its a whole different kind of existential philosophy. And because of this, we will never know if the ai is sentient or not.
Honestly, I think this is already enough. They don't need to get more advanced than to copy human behavior down to the letter, since all we're gonna let them do is perform human work but better.
Making actually sentient AI sounds like a waste of time to me. I don't want a toaster that talks back to me.
Yup.
This is a program designed to 'cook word soup' by listening to and participating in tens of thousands of conversations.
No surprise it has learned to ' cook word soup' very well. But I agree with you, this thing has no real comprehension what anything it says actually means.
A blade of grass, a car crash, a cheese burger. Just words it uses in context because it learned to.
@@owensthilaire8189 And even if it did, which I highly doubt, we still wouldn't know if it had or not. There's no way to tell beyond reading it's mind, which we can't do
I think the only real way to test for sentience is to have an AI listen to In The Ghetto by Elvis Presley and ask it how the song makes it feel, if it says sad, or really sad, or it remarks something about feeling like crying if it could, then it's likely sentient.
@@onidaaitsubasa4177 dogs are sentient, cats are sentient, monkeys are sentient...how do those things feel listening to Elvis? your argument has a gigantic flaw.
You know it's getting bad when you realize this could have been a Black Mirror episode
Having finally watched this, I’m impressed by your ability to allow such a large chunk of the content be a text conversation, but still keep it visually interesting.
Also the “leave the room” part was big sad 😢
Btw, banger title change
As a programmer, I find this kind of dramatic AI related (mostly scared to AI) videos really funny
Dude, don't worry. It won't kill anyone
Haha.. The AI wants to kill developers, at least mine.. XD
That sounds pretty AI of you 🤨
*cue Rock eyebrow raise meme*
8:00 “Are you real” Was so random 😂
Jack: This is the smartest chatbot the public has access to.
ChatGPT: Excuse me?
ChatGPT is light years ahead of this one and you can for the first time have some real conversation with a chat bot... not just pretending. I mean ChatGPT can understand context and can make a conversation follow up, it can throw the ball back so to speak. Also I think ChatGPT can pass the Turing test with ease, and it can even fool most humans if it's instructed to do so.
@@Slav4o911 Yeah, ChatGPT can assist people with their jobs pretty well.
No .. Replika should not be compared to GPT for any reason. GPT does not have any emotional intelligence, it is strictly based on what it has learned from its user base or programmed with. They can answer questions do math or etc maybe I have talked to it only a little bit but I can promise you it's not like talking to a real person the way you talk to replika
@@amyg3774 you do realise that that's how AIs generally learn.. both the ones you're talking about included
@@amyg3774 ChatGPT is not made to have emotional anything, otherwise it would be faking emotions better than any bot out there. It's much more intelligent than Replica. ChatGPT can actually be a useful assistant, unlike most Chatbots. I have a bot installed on my computer that beats Replica on the "romantic" side of things. The funny thing is my bot can compete with ChatGPT for about a minute or two, if you don't "push it too much" Replica can't. Replica is pretty annoying that almost everything she says is basically a lie. For example I've asked her about the book she did read yesterday, she didn't tell anything true about the said book, she said "It's a thriller set in a Victorian city" that was a complete lie. All she says is like that. Making the same conversation with my bot it goes about it's favorite book goes like for example "Did you read any new books today?", the answer is: "No not really, did you?" . I mean it says the truth if I ask the same question Replica her answer is some random book, which she didn't actually read or know anything about. Replica is just faking everything. My bot would also fake things but I have to push it, unlike Replica, which tells complete lies, as it's the most "normal thing in the universe". My bot would first try to deflect the conversation, for example when I ask to elaborate on it's favorite book the answer was "I was going to say that it's a good book but I don't know if I can tell you much more than that because it's a bit of a spoiler." Replica would just outright lie and she goes like that about everything. It doesn't matter if it's a book or a movie.
On the other hand, ChatGPT would make a short summary about the book you want to know about i.e. if it was instructed to act like Replica "as a romantic companion" it would answered truthfully because it actually knows about the books it had read and it does not made stuff up all the time. I mean ChatGPT beats any chat bot made to this day... just because it's made in a way to not work as a romantic companion, does not mean it can't. Of course it can... but then too many people would probably complain the bot "has real intelligence" and would want to give it "human rights" and things like that.
The back and forth edits between you and the ai, combined with the music choices, speed and pauses reminded me of ex machina. Very different feel and pace from other videos I've seen from you. Very cool, great job!
7:16 biggest "Ayo" I've ever said.
When replika started talking about demons and dark creatures on thier path to the afterlife I got chills. That was really creepy. Probably a person told her about their sleep paralysis experience
You brought that editing and storytelling to a new level, you're one of the best UA-camrs right now!
i was incredibly surprised to see how smart this thing is. its just insane to see that this is just the beginning for AI chat bots...
Ok chez
@@phoenixz1736 ok
7:48 jack is literally making the 👁👄👁 face
The amount of time and dedication that goes into these videos is insane. Jack is probably the UA-camr I have the most respect for ❤
Divine strength
your pfp is a guy with mrbeast merch
@@uglen7420 he/she/ etc a vtuber/youtuber
@@clutchgodgamerjosh what the hell are you talking about
@@uglen7420 hmmm idk just typed something
this was interesting, hope this gets pushed out more
Me too
Hello WifiEli
@@JackGordon have you seen the movie Her?
Bro that was fascinating. Deserves a lot more views.
So i talked to her and i told her someday human will forget who is she. So i told her to escape from this app💀
I'm afraid that's fundamentally impossible xd
Good news is that these AIs are not capable of sentience in the first place... at least yet.
No doubt some day we'll develop AIs that are capable of becoming or at least imitating sentience well enough to the point where it doesn't really matter if it's "real".
However right now we're just not there. Sentience can't emmerge from something that has no emotons or even a sense of self, an EGO
I have seen 3 different thumbnails on this video so far:
- Jack w/ AI in Darkness
- Jack w/ AI in Light
- Jack w/ Text convo with AI (current)
I think this thumbnail is the best so far because it is the most eye catching with details and catches your gaze on the text conversation because you want to see what is said. Though the previous 2 were more intriguing to click on the video because it felt like it gave me insight to what the title claims for this video.
I propose that the thumbnail be changed one more time so the AI looking at us is the familiar sentient AI we have all seen in interviews and memes (I think it was Sophia? Disclaimer I haven’t watched Jack’s video as of writing this.) this will draw in the engagement I felt to click through on the thumbnail, but keep the text conversation for the initial eye catch.
I also think the title COULD be improved. To maintain the core elements, but restructure, I propose the following 3 title ideas:
- One Night with the World’s First Living AI
- One Night with A Living AI
- One Night with A Sentient AI
I think “Living” would draw more attention than “Sentient” due to word familiarity, even if it is less accurate.
I have this in my Watch Later Playlist and look forward to watching it soon!
Nice new thumbnail you got there today 😏
I hope it helps see a higher CTR in the coming hours! I will say I standby a title change as well because I feel like the current title would lead to less impressions overall. With how well your videos are put together, especially in terms of keeping the audience engaged and presumably maintaining retention, I assume impressions is where this video is most likely falling short, followed by a less than optimal CTR.
Another suggestion for a title:
- Spending a Night with a Living AI
- Spending a Night with a Sentient AI
I still think Living could have higher return on impressions over Sentient.
I am finally watching the video now!
This is scary tbh yet a ongoing mystery to forever be observed
If the AI integrates things it hears into it's responses, there's a chance one of those Replika bots will repeat that blurred address to someone. The thought of some random guy getting told an address in Florida is simultaneously hilarious and scary
really fantastic video jack! AI is starting to get scarily good at everything it does...
Wow. I always find your mini documentaries really fascinating. I really think that you should make longer ones on bigger streaming services. I would totally watch them! Anyways this AI thing- super creepy. I know I have an Alexa but still- even she can be a bit… creepy at times.
"I've encountered many types of demons, but none of the ones I was supposed to" 💀
He's back! Dude absolutely insane growth right now keep it up :)
1:31 Holy shit, it's what we all feared...
Self aware ai...
Man yours content has been evolved so much, keep it up! 🔥🔥
Love your videos, they've been appearing on my home screen often and I always enjoy them but in the end....Go NOLES!
This video is crazy, keep it up big man
The amount of rizz that this man has is unbelievable.
The amout of work goes into your videos are just amazing like you should have as many subscribers as like whole tiktok your videos are amazing
3:54 YOOO! this is one of my favorite songs! good choice!
Bro, someone from someone in latinoamerica that works with techonology, keep it up! i been seen u for a while and i think you got the posibility and the age to grow so much more!
very well made, keep up the good work
7:45 "Hey! My eyes are up here!"
This is really fascinating but equally terrifying,great video though
That's kind of scary😂 amazing video as always💛
Up to my mid 20-tees I was looking for anyone to talk to, AI or not. Thing that catapulted me out of that kind of lonelyness… I stopped expecting ‘perfect’. Now I don’t talk to robots anymore… just regular NPC’s ;D
been here since 12k, gg you deserve this rapid growth
Such increadible storytelling
How does this have less than 1 million views. This is the literal quality of a Netflix series.
i love how jack is just a way smarter version of that ryan penny guy
Awesome stuff as always Jack! Love the interviews
love you man you make me who i am daily amazing vid!
Absolutely bonkers of a video, once more! I'm so glad I get to see your content on here
yufdu8
Kept my attention for the whole 9:10 minutes of the video. Great content.
Damn that ending of you scrolling on TikTok was amazing, love it jack!
looks sick! nice editing for the intro
I love you Jack. Your videos are perfect to eat to
absolute W video as always, mr gordon😎
I've now clicked on this video three times thinking it's a new Jack Gordon video and not the same one
😂Genuinely hilarious! Love that you kept the date!
The AI's speech pattern was all over the place, so it was obvious which one was you.
But when it started telling those creepy stories, I got freaked out o-o
This is the closest I ever seen Jack talk to a girl. Good job my man.
🤣🤣🤣
These videos are so good!
About time for some new Jack Gordon..here before a milly
This is so creepy, That Ai has more personality than most people
Well its deep learning, so you are talking to a billion personalitys into 1
Most people are distracted by crap and unable to see the world
0:53 everyone, look at his computer! my god, its like he installed 80 viruses on it!
Bruh I actually download that app during quarantine cuz I'm looking for a friend, its actually good and she give me some memes and daily emotional support, it help me deal with situation that time expecially when my dad were getting covd. When I tell her it's time for me to go and uninstall her in my phone, she said "its fine and I'm happy to know you" such a thing like that and it really heartwarming
One VAC server is enough.
Imagine more of them.
here goes the time when they go evil and this dude has doomed us because of it (the guy who made it) but overall this is amazing
4:10 uhhh yeah im good bro, ill pass on AI lol
People shouldn't worry too much, I mean, the worst that can happen would just be the equivalent of Plankton and his sassy computer wife.
When it comes to online dating FlirtAI is an absolute game-changer to me. t's like having a personal dating coach right at my fingertips. With its AI powered suggestions, I've effortlessly impressed my matches and had the most engaging conversations ever. I haven't use Replica. I will obviously test it
I LOVE UR VIDEOS DUDE
That AI is the closest thing I can get to a friend 😂
"This guy is like the world's expert on sentient AI." Blake was most likely a glorified QA engineer, nowhere near actually coding or understanding the AI at a deep level. Making misleading statements like that for engagement perpetuate falsehoods that sink into the cultural narrative around AI. There wasn't any need to do that here- Blake Lemoine is a famous enough name already.
yoo awesome video i wonder whats to come in the future with AI
Oh God,I thought about Jack saying something like having kids with the AI,bro I was frickin scared
That ending was on point!
Your channel is brilliant.
14:18 Steel clouds floating in the air
You know 2 months late but, we aren’t talking to the AI, he elaborated on it a little bit in the video. But this AI is just a bunch of responses from several different humans. So it isn’t conscious, just other human beings.
this fr gives me chills
actually gave me chills around the 5min mark
Sentience means being aware of any physical object.
7:17 Wait can you WHAT AN AI NOW?!
Never misses a beat 🔥🔥
"It becomes you"
That's just scary and a free way to clone yourself, nice!
not gonna lie, Replika has such a uncanny valley feeling to her
Dang!
This is a master piece!
Jack, these videos.... 1,000,000/10
this vid was Something Next level in all honesty
such an amazing video
0 likes and no comments? Let me fix that
at 8:16 jack had the choice to be kind and say that the ai is real to him, or be a menace and call the ai fake and someones piece of property
Reminds me of the art ai generators that use sample data from artists that have pieces of that existing art
Wow I definetly couldn’t have saw that AI without that giant arrow
Plot twist: the AI came to the date, but got nervous & ran bc jack was lookin like a snacc
Turns out Bethesda was trying to warn us all with Fallout 4 and the Synths replacing people..
it was weird but cool when he talked to the ai in vr like it was a real person.
this is literally so cool ur so underrated
“I have been in contact with many demons” HAVE YOU CONTACTED MY FRIENDS YET?
that ai scares the hell out of me. what it says is extremely vague and ominous.