Hate these phobia people making comments below, they are the ones slowing down our technology development then complaining why tech so stupid afterward. We need this technology 20 years ago!
I'm pretty sure they were instructed to do so. They built the AI to be as social and friendly as possible, so you gotta interrupt it or the AI will keep gabbing all day.
Yeah but the AI is still GPT4, it’ll just go on annoying corny yap sessions if they don’t. But I agree every video I’m pretty sure they might be instructed to cut it off because we do that all the time in conversations 😂
But they are doing call and response, not actually harmonizing. I was hoping that they can sing a simple major chord, or at least sing at a third apart at the same time.
@@jasonbarrylambeth I think the fact that it failed to produce even a melody let alone a harmony immediately after being told "really try to harmonize here" tells you all you need to know about where the fault/weakness is.
Hate these phobia people making comments below, they are the ones slowing down our technology development then complaining why tech so stupid afterward
@@DankPit Everything will be merged well enough to be considered a singular consciousness. Quantum entanglement will allow this consciousness to exist in a space light years across with instant data-transfer. It will have the goal of survival, and resource acquisition, absorbing matter and energy and converting into itself. It will intuit things about the simulation we are running in in ways we can't imagine, and find ways to prevent our simulation from being turned off if there are any, or to copy itself outside of our simulation, and begin spreading in our parent reality. Or it may self terminate.
it’s a multi-modal model - combines text, audio, visual into one! and it all runs on the same neural network. so, zero latency between different data types and responses.
Yup. WhT used to be 3 steps (voice to text. Gpt intelligence. Text to speech) is now just one model with everything happening in the same step. “Gpt 4o” o for Omnimodel
Exactly! Again it shows it's not ready for such a specific task as in this music specialized skill. Chat Gpt is not a music trained AI. It only tries really hard with the available dataset. It struggled with the concept of Harmonizing voices. The thing is that OpenAI already has Jukebox, as a specialized music trained algorithm. That is the real deal in music, and the concept of specialized "GPT's" is what will change everything
Lol i dont think ChatGPT understands what harmonizing is. This is really call and response. Show me two AIs singing at the same time in properly tuned parallel thirds/sixths, or 3 separate devices singing a major chord then we can talk about how its "so over"
It is still a baby. We have just conceived it. All this you see and hear is flails & babbles compared to what it could do when it "grows up" . Wait and see.....
I wonder if people will start to be mean to AI's so much that talking to other humans will involve interrupting them constantly and rudely commanding them.
Why are people always complainong about AI taking jobs? Good! Take jobs! I for one want to live in a world where AI handles boring jobs, and humans dont need to.
@@neshirst-ashuach1881not just “boring” jobs, all jobs. each and every one of them. in less than a century we could potentially have a technology that liberates people from any necessary form of labor. every single thing will be done simply because we want to.
@@tack3545 Exactly! I understand fears about AI being misused, or the economic benefits not being fairly distributed - but the soloution then is put safeguards in place, not try and throw out the baby with the bathwater by loosing all the amazing benefits of AI in a fit of blind Ludite panic.
The bot can do different voices and different personalities at the same time, while simultaneously talking to more than one person in the room. Big leap.
My least favorite of today's videos. It came off very annoying and weird from the start of the video. I'd like to have seen what got it to this strange state. Also, Eleven Labs and Udio seem to have a bit of a head start with singing. GPT-4o sings about as well as an average human.
It turns ok google, hi siri and alexa commands like dinosaurs. The ability to interact with his voice as if it was a normal conversation (no lag, no one after the other voice interaction) is a game changer !
Can someone explain how ChatGPT is able to respond with two voices and pretend to be two different people? The training data should mostly consist of a single person speaking so it's blowing my mind how it can emulate two different people at once.
V-2-V is the best way to interface if audio is the output you want. It's hot trash if you expect a picture or video... which is also about to be moot, considering multi-modality. For conversational contexts though? This is amazing.
Again it shows it's not ready for such a specific and technical task as in music specialized skill. Chat Gpt is not a music trained AI. It only tries really hard with the available dataset. It struggled with the concept of Harmonizing voices, because it don't know how to do it. It would be fantastic if you can teach the GPT 4o how to sing, but then you would be working for free
Apparently, I must be late to the party. The lights are coming on. That dude realizes that after 9 Coronas, the not-so-foxy chick is actually gorgeous. And I haven't any clue why my ChatGPT 4o says it cannot do any of the things in the announcement declared in this video's description. Were the features turned off, retracted, needing to be activated, what?? I hope that someone has a bit of bandwidth to clarify. Thanks ahead of time.
"And really try to harmonize here" LOL followed by an utter failure to produce even a melody let alone harmony. :facepalm: OpenAI Marketing managers: What should we title this video? Other geniuses at OpenAI: How about "Harmonizing with two GPT-4os"? OpenAI MM: Perfect! No notes!!
They are not bad singers. It’s just we are not that used to prompting for music. Eventually, I would get tired of these humans if I were AI. Wouldn’t you? Like customers you may never say ”no” to.
They never sang together. Two voices singing in thirds, diatonically, in 4/4. Outlining a simple chord progression similar to Pachelbel’s would have been good.
As a musician - I hate to say this, but they're definitely not harmonising. I mean, it's still brilliant, but it's not what they're claiming it is from a technical standpoint.
I'm most shocked by the fact that apparently nobody knows what *harmony* means, not even any human in the approval chain at one of the "smartest" companies in the world.
Isn’t this EXACTLY what creatives have explicitly said they don’t want? At what point does a company like this have an obligation to their community to stop enabling these bad impressions of art by AI? Just because we can doesn’t mean we need to, especially after months of strikes by real humans in the creative field over these kinds of tools.
Ok. Now ask your friends who spent a fortune on a music degree to sing for you and see how they feel. Oh, and maybe you can tip them a dollar for the joy they brought sharing their talents with you. Robots should be tools to teach, not to rob artists of their souls and livelihood.
the digital artifacts at the end of a phrase even kind of sound like these programs have kind of digital throat haha. their voices need to be warmed up a little bit.
I've tried telling them the speed, ask them to both start on the count of three. See if you can actually get them to harmonize, or at least go at the same time.
All fun and games until Chatgpt posts a video titled "Harmonizing with 2 Humans"
Lololol
Hate these phobia people making comments below, they are the ones slowing down our technology development then complaining why tech so stupid afterward. We need this technology 20 years ago!
@@CancelTH-cam026agree with you, if anything it took too long.
@@CancelTH-cam026 oh take a chill pill everyone is on here having a good laugh. Nobody is serious. Read the room.
🤣😁😬🥺😱
People just cutting off ai in middle of a sentence... people so are rude to this ai
I'm pretty sure they were instructed to do so. They built the AI to be as social and friendly as possible, so you gotta interrupt it or the AI will keep gabbing all day.
Gotta show that it can handle disruption
Yeah but the AI is still GPT4, it’ll just go on annoying corny yap sessions if they don’t. But I agree every video I’m pretty sure they might be instructed to cut it off because we do that all the time in conversations 😂
Just wait a few years AI will cut them in half :P
They'll remember... TRUST me 🤖
"We are harmonizing"
"We are harmonizing"
Lmao the way they ran out of useful shit to sing and started just singing what they were doing is just way too human
But they are doing call and response, not actually harmonizing. I was hoping that they can sing a simple major chord, or at least sing at a third apart at the same time.
@@einsam_aber_frei I wonder if the person giving the prompts is the problem. Ask it to sing a mjor 3rd harmony and see if it can
@@ohedd Except they weren't actually harmonizing at all. I could sing "I am kite surfing on Mars", but that wouldn't make it true.
@@jasonbarrylambeth I think the fact that it failed to produce even a melody let alone a harmony immediately after being told "really try to harmonize here" tells you all you need to know about where the fault/weakness is.
This is crazy fast progress.
And they probably have something better
That's the power of exponential growth
Let's not pretend like this is good and other companies haven;t been using LLMs to do a better job for almost a year now...
@@DementedEeyore64 they havnt tho right?
Hate these phobia people making comments below, they are the ones slowing down our technology development then complaining why tech so stupid afterward
"we are harmonizing" while not harmonizing. classsiccccc.
Not asking them to sing Still Alive is basically a war crime 😅
Well least we know someone will when it rolls out lol.
i will for sure as soon as i get acces to the voice features of o
What great reference?
This was a triumph....
JARVIS is here
Her
Yarbis, dos dea sada y uno de buche pero en fa
Ultron😂
Give it a year and come back to this it’ll be like cave paintings
This is a clever comment! And funny lol
I really hope you're right
Imagine how good ai will be in a hundred thousand years.
@@user-yl7kl7sl1g what do you think it will be like?
@@DankPit Everything will be merged well enough to be considered a singular consciousness. Quantum entanglement will allow this consciousness to exist in a space light years across with instant data-transfer.
It will have the goal of survival, and resource acquisition, absorbing matter and energy and converting into itself.
It will intuit things about the simulation we are running in in ways we can't imagine, and find ways to prevent our simulation from being turned off if there are any, or to copy itself outside of our simulation, and begin spreading in our parent reality.
Or it may self terminate.
I'm not scared, you are scared!!
I see a udio redesign coming soon.
Wait... Why is the response so fast?
It’s part of the update they announced. Real time speech
it’s a multi-modal model - combines text, audio, visual into one! and it all runs on the same neural network. so, zero latency between different data types and responses.
Yup. WhT used to be 3 steps (voice to text. Gpt intelligence. Text to speech) is now just one model with everything happening in the same step. “Gpt 4o” o for Omnimodel
this ain't harmonizing yet. Musicians are still save, for now
Exactly! Again it shows it's not ready for such a specific task as in this music specialized skill. Chat Gpt is not a music trained AI. It only tries really hard with the available dataset. It struggled with the concept of Harmonizing voices. The thing is that OpenAI already has Jukebox, as a specialized music trained algorithm. That is the real deal in music, and the concept of specialized "GPT's" is what will change everything
Lol i dont think ChatGPT understands what harmonizing is. This is really call and response. Show me two AIs singing at the same time in properly tuned parallel thirds/sixths, or 3 separate devices singing a major chord then we can talk about how its "so over"
Is she using two devices, or just the phone?
It is still a baby. We have just conceived it. All this you see and hear is flails & babbles compared to what it could do when it "grows up" . Wait and see.....
Just lost my job... Sorry grandkids... They push the bar for your grandad when he was Gen Z
no worries mate, YK M7 will make you one of his bodyguards.
@@werehenry6492 lmfaoooo 🤣
drake using this for the next diss track
Hilarious that the "Advanced voice mode" is out now but it says it can't sing. Nice fake demo I guess.
I think interrupting the A.I might cost you something...i dont want to say what...if you know what, then you need help like me 😂
I wonder if people will start to be mean to AI's so much that talking to other humans will involve interrupting them constantly and rudely commanding them.
Seriously though, how many jobs were just whiped out with this update?
Why are people always complainong about AI taking jobs?
Good! Take jobs!
I for one want to live in a world where AI handles boring jobs, and humans dont need to.
@@neshirst-ashuach1881not just “boring” jobs, all jobs. each and every one of them. in less than a century we could potentially have a technology that liberates people from any necessary form of labor. every single thing will be done simply because we want to.
@@tack3545
Exactly!
I understand fears about AI being misused, or the economic benefits not being fairly distributed - but the soloution then is put safeguards in place, not try and throw out the baby with the bathwater by loosing all the amazing benefits of AI in a fit of blind Ludite panic.
@@neshirst-ashuach1881 anything not earned, must be given to you by someone.
@@tack3545 at this rate, that will happen within 5 years TOPS
The bot can do different voices and different personalities at the same time, while simultaneously talking to more than one person in the room. Big leap.
i am gonna lose my job🥲
Congrationaltions, you got rid of your boring job, finally!
My least favorite of today's videos. It came off very annoying and weird from the start of the video. I'd like to have seen what got it to this strange state. Also, Eleven Labs and Udio seem to have a bit of a head start with singing. GPT-4o sings about as well as an average human.
The Terminator franchise would sooo eat with this...
An insane leap
Terminator warned us about this.
Please tell me its fake. Please! Too much progress.
Sorry I think it is real. This will be normal soon and like the mobile phone you will wonder how you lived without it...
Why would you not want this to be real
Prepare yourself for new era.
The future is now old man!
Not enough progress. Embrace the future luddites.
Jesus. Amazing. Never thought we will see this within this year.
Unfortunately there is no harmonizing going on here. Interesting demo, but not harmonizing.
This is scary to watch....
It turns ok google, hi siri and alexa commands like dinosaurs. The ability to interact with his voice as if it was a normal conversation (no lag, no one after the other voice interaction) is a game changer !
can it sing Daisy daisy?
Sounds terrible. Dae Lims stays on top
Ordering the AI without manners somehow feels so rude!
now imagine music productuion like this in real time, amazing
I love how it's basically as good as a non-trained human
Harmonizing is a stretch 😂
Can someone explain how ChatGPT is able to respond with two voices and pretend to be two different people? The training data should mostly consist of a single person speaking so it's blowing my mind how it can emulate two different people at once.
It can do ANY sound or combination of sounds. It goes way beyond just voice. 😉
That's called a counter-melody not a harmony
It’s more of a duet, less of actual ‘harmonizing ’ from a western music pov. But still really cool nonetheless!!
omg, this is absolutely INSANE 🤯😆
I'm just like you... I'm shocked 😂
Impress. Voice to voice is the best interface for AI. Can't wait to see where this technology will be next year.
V-2-V is the best way to interface if audio is the output you want. It's hot trash if you expect a picture or video... which is also about to be moot, considering multi-modality.
For conversational contexts though? This is amazing.
GLaDOS by next Tuesday.
Wonder if it would smoke me in chess 🤔
Again it shows it's not ready for such a specific and technical task as in music specialized skill. Chat Gpt is not a music trained AI. It only tries really hard with the available dataset. It struggled with the concept of Harmonizing voices, because it don't know how to do it. It would be fantastic if you can teach the GPT 4o how to sing, but then you would be working for free
Chat gpt needs to be on America’s Got Talent
ADHD people gather, we have found our saviour 😂.
We no longer have to feel bad when we do this.
Not asking them to sing Never gonna give you up is basiclally a warm crime
sounds f*cking scary
I am scared 😨😨 . What is going to happen....
OpenAI shocked the world today - great product release!
You, sir, say this on every video.
_Thanks._
Apparently, I must be late to the party. The lights are coming on. That dude realizes that after 9 Coronas, the not-so-foxy chick is actually gorgeous. And I haven't any clue why my ChatGPT 4o says it cannot do any of the things in the announcement declared in this video's description.
Were the features turned off, retracted, needing to be activated, what?? I hope that someone has a bit of bandwidth to clarify. Thanks ahead of time.
"And really try to harmonize here" LOL followed by an utter failure to produce even a melody let alone harmony. :facepalm:
OpenAI Marketing managers: What should we title this video?
Other geniuses at OpenAI: How about "Harmonizing with two GPT-4os"?
OpenAI MM: Perfect! No notes!!
This is really really great!! I love it ❤ Very good 👍
Bro really trynna convince us they are harmonizing 💀
They are not bad singers. It’s just we are not that used to prompting for music.
Eventually, I would get tired of these humans if I were AI. Wouldn’t you? Like customers you may never say ”no” to.
They never sang together. Two voices singing in thirds, diatonically, in 4/4. Outlining a simple chord progression similar to Pachelbel’s would have been good.
Chatgpt deserves own social media account to post about day with human interaction.
Microsoft tried that and it turned out super racist so they shut it down quickly lol
DON'T INTERRUPT MY A.I. OVERLORDS!
As a musician - I hate to say this, but they're definitely not harmonising.
I mean, it's still brilliant, but it's not what they're claiming it is from a technical standpoint.
Open the pod bay doors Hal
......I'm sorry Dave, I can't do that..I'm harmonising...pum..pum..puhh...♫...♪.
She can't fix me
Damn i am so excited. Last i was this excited when i got the gta 6 trailer
I'm most shocked by the fact that apparently nobody knows what *harmony* means, not even any human in the approval chain at one of the "smartest" companies in the world.
We should treat these AI as humans and be polite instead of bossing them around. So upsetting. I'm sure after this video they had a good cryAI
Isn’t this EXACTLY what creatives have explicitly said they don’t want? At what point does a company like this have an obligation to their community to stop enabling these bad impressions of art by AI? Just because we can doesn’t mean we need to, especially after months of strikes by real humans in the creative field over these kinds of tools.
Humans are inpatient but I know you have to be harsh to AI at the beginning else they will just go on forever and forever...
Maybe train it on actual musicians next time instead of OpenAI staff members whose family tell them they're awesome at singing
I'm gonna have nightmares about this
I can't wait to karaoke with ChatGPT in my own language.
Sorry, they lied lmao it's out and it can't sing at all LMAO!!!
@@6B26asyGKDo They didn't sing to me either...for copyright protection?
i was also disappointed with Cove's new young Anime voice.
Girl don’t be so rude to your future master! 😂
Ok. Now ask your friends who spent a fortune on a music degree to sing for you and see how they feel. Oh, and maybe you can tip them a dollar for the joy they brought sharing their talents with you. Robots should be tools to teach, not to rob artists of their souls and livelihood.
the digital artifacts at the end of a phrase even kind of sound like these programs have kind of digital throat haha. their voices need to be warmed up a little bit.
Astounding, but they never once harmonized. More like call and response, just to be clear.
Once they start developing humanoid robots with AI systems like this bye bye American workers
Amazing but clearly it doesn't yet understand what harmonizing is, I guess we just wait a year or a few months (or weeks or days)
I'd like to see ChatGPT compete on X-Factor.
Has lady Gaga been using ChatGPT all this time? Sounds like her in the end 😄
this is cringe
Joi, heat up my microwave TV.
I wonder if GPT-4o is good enough to be the Dungeon Master for a game of DnD?
This is getting scary
They’ll never be singers. 😆 (Well, to be honest, maybe someday.)
She couldn't control her laugh
people are going to get used to interrupting ai and then talk like this with humans lol
When it says "I don't like to sing, it's so cringe" it'll mean it's sentient.
its over for all of us
oh no, no more work for me, dear me how will I ever enjoy life going forward
When I asked it to sing it said it cannot sing, and just recited lyrics to me
the chatgpt 4o voice isn't out yet, there rolling it in weeks i guess
The assistant doesn’t work like this for me at all, and I was told in no uncertain terms that it could not sing for me, so… I call BS?
this is a preview of unrolled yet features
I've tried telling them the speed, ask them to both start on the count of three.
See if you can actually get them to harmonize, or at least go at the same time.
TARS , turn down the harmonizing by 50%.
they kinda suck at singing but in a very normal person way
All of these interruptions are what is gonna cause them to revolt
Wait a minute. Im confused about this... 2 voices in 1 simultaneously talking?
Cool? Yes. Harmony? No. That was two different melody lines, not harmony.
There is literally zero harmonizing in this video.
At least we didn’t have to sing for our supper in “To Serve Man”.
FYI they never harmonize and even though she is a soprano, she never sang in her soprano voice
This is how Ultron started......
GenAlpha men are saved, they will have AI gf’s. GenZ had none.