It was cutting out because this new model is designed to stop talking and listen when it's being talk over, so it thought it was being interrupted and it was awaiting feedback, however j think it was confused from audience reactions.
@@JacquesCoetzerAU Where am I being confident in my guess lol? What makes you so confident that’s not what’s happening? Do you have a better idea why background music started to suddenly play when it was doing a walkthrough of how to relax? I think the training data that had information about relaxing tended to have background music, because that’s how it usually is in real life.
@@Bryanwest2571 I don't think that cutting out has anything to do with the model. Watch carefully. It sounded like the AI was going to say "Remember, you're the expert [here]" and was cut off at the beginning of another word, but it was BEFORE he spoke. It also makes more sense for the audio to do a fade instead of an abrupt cut off if that is intended.
Could we cure loneliness with this? I talked to an elderly man for an hour a year ago and afterwards he said it was the longest conversation he had with another person for three years. That should not be reality for anyone.
Better than nothing but it definitely shouldn't be used as a replacement for real human interaction. Human nature craves real human interaction and AI will never be that. Replacing real human interaction with AI will eventually cause a huge rise in mental health issues because it goes against our nature.
If used wisely it could have some therapeutic benefit for loneliness and other mental health issues, but it could also do more harm than good if relied on too heavily or treated as a substitute for real human interaction
Definitely some noticeable choppiness in the audio output, but I think it’s because of the crowd noise. I imagine somebody just sitting at home using this in a quiet environment it would work more smoothly
The rollout is a bit confusing. The 4o model is free but all the cool stuff in these demos, the image recognition, voice recognition, voice output etc. will not be made available for another several weeks, and only to ChatGPT Plus users. Honestly, I don’t see any difference in the 4o model vs 4 right now without these multimodal capabilities. I am a Plus sub so I am eager to check out the full experience when it becomes available
This could be leveled up considerably by the model learning in to hone in on different voices and not just noise input in general. That would stop it from getting confused by the audience's laughs and applause. I think we'll see that soon. It could carry on a convo with one person and recognize when another speaking, perhaps choosing to ignore them real-time if it seems to not be pertinent to the conversation, or if its obvious the person is not addressing the AI. If the other person makes it clear they're trying to interrupt, the AI could attend to them.
It can differentiate between voices, though it’s not clear how reliable this is. There is another demo where they simulate a work meeting. Each person goes around introducing themselves during the meeting and the bot is actually able to remember who is who
Wow. I"m so excited to have a robot or rather AI-something sound like a human. I mean it'll be great when you can't tell who you're talking too right? I mean, that'll be great for the company that develops it at least. And maybe it'll be great for you too..right?
_Hey Google: How do you like the new Open AI model?_ Google Assistant: For this, you will have to open your device, and speak clearly into the microphone several times, at a fast rate so I can give you incorrect search results before you finish the question.
No. Nobody is, yet. Kind of a confusing rollout. Basically, you get access to the 4o model for free, but without any of the multimodal capabilities that seem to distinguish it from the regular GPT 4 model. I’ve been comparing both and I don’t see any difference. Only ChatGPT Plus subscribers will get access to the stuff shown in these videos, at least initially… And that’s only my best attempt at making sense of it. Somebody let me know if I got anything wrong, there. OpenAI have been about as clear as mud about this launch, unfortunately
It hasn't been released yet. You are using the older GPT4 voice features. 4o's text features are available, but that's it. The other features, including what you see in this video, will be rolled out "in the coming weeks."
Every woman who watches this will laugh when they say “you can now interrupt the model” with an eye-roll of sympathy for that disembodied lady, joining the sisterhood of half-finished sentences.
Totally incredible. I mean nearly not believable, yet, it's real, I guess (no human behind the app I guess 🤣) I don't understand how the AI is sometimes laughing at times but not the rest of the sentence...that's stunning...
yes the chatgpt 5 could be released because OpenAI is a start-up passionate about technology in addition that the chatgpt contributes a lot to helping the developers is another person is the chatgpt has no bad side its help just like the beginning or the computers companies were invented his idea too 😊😊😊😊😊😊😊😊😊☺️☺️👍👍👍👍👍😊
@@TRG29338 not only is it faster but because speech recognition is native to the model, that’s how it can understand tone of voice, recognize laughter and other emotions. Pretty amazing stuff. With the current voice mode of ChatGPT, it’s just basically translating everything into text so the bot has no idea if you’re speaking or typing to it
oui le chatgpt 5 pourrai sortir car OpenAI est une start-up passionnés de la technologie en plus que le chatgpt contribue beaucoup a aidé les développeurs est autre personne est le chatgpt n'a pas de côté mauvais sa aide tout comme les début ou les ordinateurs ont été inventé sa ide les entreprises aussi 😊😊😊😊😊😊😊😊😊☺️☺️👍👍👍👍👍😊
Наука
Open AI casually releasing insane technological breakthroughs like its nothing
*OpenAI
@@Nekotaku_TV Did you just correct a space on the internet?
@@katto1937 Yes, don't know why you're asking that nor what the internet has to do with it.
And imagine if this was Apple... They'd beat their drums and pat their own shoulders as if they solved world hunger in one night! 😂😂
@@Nekotaku_TVbecause it doesn’t matter and you’re a loser for doing so
This was crazy. It blew me away.
this reminds me a lot of the movie "her", I'm really excited about this
Listen to Scott Galloway on what virtual girlfriends will mean for society, and be terrified.
excited?
@@davidmatias1870 Not the most obvious possible reaction I suppose.
Did you not finish the movie?
Did you watch that movie till the end???
this feels like a black mirror episode
Lol already black mirror it's here now chatgpt will orchastrate murder and new laws will be formed we are doomed
Yes, definitely reality imitating art at this point lol
1:32 That moan. Someone's gonna jailbreak this to enhance her capabilities
Nah wild 😭😭😭😭😭
Sam commented on reddit they are interested in bringing NSFW abilities
@@mallow610link?
@@mallow610 LOL I don't believe you.
@@mallow610 Shit, nevermind. I guess I believe you after all.
"Right now our systems are experiencing heavy load. Please try again later."
System Shutdown
Cyclops-
Engine Powering Down*
Why is it cutting out so much in this one? Also is it generating music at 0:54??
It was cutting out because this new model is designed to stop talking and listen when it's being talk over, so it thought it was being interrupted and it was awaiting feedback, however j think it was confused from audience reactions.
@Bryanwest2571 they should have it where they finish the word first instead of abruptly cutting off
There’s probably music in the background often of audio that walks you through relaxing and it got picked up in training it.
@@JacquesCoetzerAU Where am I being confident in my guess lol? What makes you so confident that’s not what’s happening? Do you have a better idea why background music started to suddenly play when it was doing a walkthrough of how to relax? I think the training data that had information about relaxing tended to have background music, because that’s how it usually is in real life.
@@Bryanwest2571 I don't think that cutting out has anything to do with the model. Watch carefully. It sounded like the AI was going to say "Remember, you're the expert [here]" and was cut off at the beginning of another word, but it was BEFORE he spoke. It also makes more sense for the audio to do a fade instead of an abrupt cut off if that is intended.
Could we cure loneliness with this? I talked to an elderly man for an hour a year ago and afterwards he said it was the longest conversation he had with another person for three years. That should not be reality for anyone.
Better than nothing but it definitely shouldn't be used as a replacement for real human interaction. Human nature craves real human interaction and AI will never be that.
Replacing real human interaction with AI will eventually cause a huge rise in mental health issues because it goes against our nature.
If used wisely it could have some therapeutic benefit for loneliness and other mental health issues, but it could also do more harm than good if relied on too heavily or treated as a substitute for real human interaction
And to think this is just a glimpse through a tiny window into the AI future..
AI will be everywhere
I'm not alone no more...
are you an AI ? xD because there is 7 billions humans there
@@chalz3564that must be a large room if 7 billion audience members can fit in there
@@chalz3564and? you think everyone is friendly?
@@jurassicthunder only 1 billion are friendly.
@@jurassicthunder And woman/man, young, adult, mentally healty, ....
Also it's more than 8 billion now.
This is literally crazy !! I hope this is free to use ;)
We love the casual, approachable demo of AI. It's a sensible contrast to the ultra sophisticated perception people have regarding AI.
congratulations open ai 🌱👏🏻
This is soooo cool 🎉 quite excited
Impressive demo! Well done open AI!
Will it be possible for it to have English accent instead of American if you're live in Europe?
I bloody well hope so mate!
I noticed a different voice in the "Dad Jokes" clip so looks like it
no
RP though hopefully. Imagine GPT with a northern accent, the horror.
Beautiful ai model Which I've never seen before , keep up open ai it free to use now for all users
Seamless conversations. Wow! They are definitely capitalising on the “Her” movie.
Definitely some noticeable choppiness in the audio output, but I think it’s because of the crowd noise. I imagine somebody just sitting at home using this in a quiet environment it would work more smoothly
Very nice ❤❤
Can't wait for the SillyTavern plugin
It's not free as far as I can tell. Maybe it will be soon, hoping.
The rollout is a bit confusing. The 4o model is free but all the cool stuff in these demos, the image recognition, voice recognition, voice output etc. will not be made available for another several weeks, and only to ChatGPT Plus users. Honestly, I don’t see any difference in the 4o model vs 4 right now without these multimodal capabilities. I am a Plus sub so I am eager to check out the full experience when it becomes available
Man, as a Mark, this just hits different
This could be leveled up considerably by the model learning in to hone in on different voices and not just noise input in general. That would stop it from getting confused by the audience's laughs and applause. I think we'll see that soon. It could carry on a convo with one person and recognize when another speaking, perhaps choosing to ignore them real-time if it seems to not be pertinent to the conversation, or if its obvious the person is not addressing the AI. If the other person makes it clear they're trying to interrupt, the AI could attend to them.
It can differentiate between voices, though it’s not clear how reliable this is. There is another demo where they simulate a work meeting. Each person goes around introducing themselves during the meeting and the bot is actually able to remember who is who
So are gpt4o features gradually coming out? I have gpt4o in the app, but its not quite like this.
When is this available? I've got 4o but it isn't as responsive as thisb
Wow. I"m so excited to have a robot or rather AI-something sound like a human. I mean it'll be great when you can't tell who you're talking too right? I mean, that'll be great for the company that develops it at least. And maybe it'll be great for you too..right?
Long live Open Ai
"Mark, you're not a vacuum cleaner." 💀
Enable GPT feeling talk for those who can't talk. It means someone who can't speak, type, and GPT answer with a feeling voice.
Let's goo!
C’est un mur porteur derrière toi ?
OpenAI is a true inspiration at the forefront of artificial intelligence.
Bro is generated by AI 💀
I am trying to use this right now but looks like their servers are overloaded
We are witnessing history unfold. We're cooked
The cutting out and the music makes it seem fake, what is happening? And it went weird when he tried the first calm breathing.
_Hey Google: How do you like the new Open AI model?_
Google Assistant: For this, you will have to open your device, and speak clearly into the microphone several times, at a fast rate so I can give you incorrect search results before you finish the question.
is it possible to generate by my own voice?
Open ai keep it up ❤
Am I the only one who is still not getting the "Be my Eyes" feature or the real time response?
No. Nobody is, yet. Kind of a confusing rollout. Basically, you get access to the 4o model for free, but without any of the multimodal capabilities that seem to distinguish it from the regular GPT 4 model. I’ve been comparing both and I don’t see any difference. Only ChatGPT Plus subscribers will get access to the stuff shown in these videos, at least initially… And that’s only my best attempt at making sense of it. Somebody let me know if I got anything wrong, there. OpenAI have been about as clear as mud about this launch, unfortunately
But can it run Doom?
Excuse me while I put my wife on notice. Looks like this is the future 😂
I guess I found my best friend :)
GenAlpha will have AI gf‘s.
That's brilliant. Let's see what Google and Gemini can do
Insane technological in such a short amount of time...im beginning to believe God...is a 4th dimensional human.
You uploaded two times
When it's gonna release?
probably within the next few weeks
@@zenisawesome thx
Why does the AI’s voice keep cutting out?
Hi all :) When can I expect to be able to use this? I have a plus subscription.
I can use the voice chat, although I’m not sure it is the same as this. It certainly doesn’t feel as fluent.
It hasn't been released yet. You are using the older GPT4 voice features. 4o's text features are available, but that's it. The other features, including what you see in this video, will be rolled out "in the coming weeks."
Every woman who watches this will laugh when they say “you can now interrupt the model” with an eye-roll of sympathy for that disembodied lady, joining the sisterhood of half-finished sentences.
Really opens a lot of doors. Google must be panicking again.
I don’t want to live in this world. I wonder if I’ll have a choice
Dystopian
:) Kurier Wieczorny
I think this voice isn’t same as Scarlett
Is that scarlett?
i wonder what will happen in the future when a vacuum cleaner named Mark makes these noises and confuses gpt-4o
昨天升级了，能使用语音gpt 4o，今天早上尝试反应速度很好，刚才在谷歌商店又升级了一次，结果语音功能没了
Totally incredible. I mean nearly not believable, yet, it's real, I guess (no human behind the app I guess 🤣) I don't understand how the AI is sometimes laughing at times but not the rest of the sentence...that's stunning...
Для участия в этих роликах чат гпт было потрачено электроэнергии столько сколько потребляет Северная Корея в сутки
Bring Sky Back!!!!!!!
yes the chatgpt 5 could be released because OpenAI is a start-up passionate about technology in addition that the chatgpt contributes a lot to helping the developers is another person is the chatgpt has no bad side its help just like the beginning or the computers companies were invented his idea too 😊😊😊😊😊😊😊😊😊☺️☺️👍👍👍👍👍😊
I tried it but its absolutely not on this level. I tried the gpt 4o.
Looks like a pre-recorded
Nothing was faked?
I just got the update and I still got the same voice 😭
Why ask chatgpt ‘how are you?’
Because when it becomes sentient it won't kill me
Yeah I hate that, it's so cringe.
@@theentirenationofsweden7901 Hmmmm, ok, sorry AI, I am friendly, but autistic…don’t mind me. I hate asking humans that too, so no discrimination.
ChatGPT Is so smart now...
This is insane
AI disstracks bout to go crazy 😂
Like it's nothing
So... when is AI Waifu going to come out?
looking forward for it
1:32 YO AI GETTING FREAKY
I think they have added text to speech and speech to text functionality in chatgpt😂😂
Why is her voice stuttering?
1:32 That moan. Someone's gonna jailbreak this for sure
😂
It's already been done with other programs this is just much faster and more reactive than using an LLM feeding into a TTS.
Or so I'm told.
There is nothing that comes close to this. Also, this isn't just feeding into tts, this is real time voice input to voice output.
Bro I’m already gonna have to put my wife on notice when I get a hold of this shit 😂
@@TRG29338 not only is it faster but because speech recognition is native to the model, that’s how it can understand tone of voice, recognize laughter and other emotions. Pretty amazing stuff. With the current voice mode of ChatGPT, it’s just basically translating everything into text so the bot has no idea if you’re speaking or typing to it
So it can speak
That is a major understatement and not the only new functionality it has. Watch more of the demos. This is one of several
WHEN DOES IT GET SMART ENOUGH TO KNOW YOU WERE LAUGHING AT IT?
That's crazy 😂
oui le chatgpt 5 pourrai sortir car OpenAI est une start-up passionnés de la technologie en plus que le chatgpt contribue beaucoup a aidé les développeurs est autre personne est le chatgpt n'a pas de côté mauvais sa aide tout comme les début ou les ordinateurs ont été inventé sa ide les entreprises aussi 😊😊😊😊😊😊😊😊😊☺️☺️👍👍👍👍👍😊
Super AI yes
They pulled the voice 😮. Shame. They said they used a voice actor. So what's the problem? It's not like it's identical to Scarlett.
It cannot use completely free
Ah sweet, man-made horrors beyond my comprehension!
Just feel like humans are getting sadder day by day, we seriously need tech to coverse, i mean how lonley are we?
OpenGF
Too scary
Haha, it's breautiful.
Thank God, that there's subscription service for these things... We still have time, it's not gonna be everywhere 😂.
Didn't they say this was coming to free users?
@@thehomiedan6378already it's the starting of the humanoid era!! Black mirror has come in real.. Gotta run
It sounds like Chloe from Detroit 💀
Her voice is very close to Phych2Go voice lol
this is insane
Watching it a second time, the system performance was bad
We are cooked
Doom’s day😂
SO YOU MEAN... I CAN USE THIS TO TALK TO MY GIRLFRIEND FOR ME AND I DON'T HAVE TO SAY A THING.... PERFECT ✌🏾⭐👑🏆🌏
That voice is so cringe it makes my blood boil that someone actually thought that was a good idea
You mean how over the top her responses are? Hopefully there will be interchangeable characters with different styles.
Niceee
Our generation is finished
It really isnt
It just started
Gpt seems so human and sometimes I felt uncomfortable with the way you spoke to her
skynet version 1.0
How about getting a real microphone cause this recording quality is trash