Yeah, I think they are pretty smart. They released the thing that would impact the general audience the most, while also not dropping AGI to society all at once. Personally, I'm ready for AGI now, but shockingly there are actually people out there who don't use AI on a daily basis. This update is designed to catch them up with the rest of us before AGI drops.
If this is only a modified GPT-4, and it's already this good, and sounds this intelligent and natural, while being completely multi-modal, just imagine what 5 will be like. It's kinda crazy how I thought GPT-3.5 was groundbreaking insanity, and just a very short time after, it seems like the village idiot after interacting with GPT-4 Turbo and GPT-4o.
Great point. At this point, I'm actually having trouble imagining GPT-5, come to think of it. Other than hallucinating less often (which would be amazing, don't get me wrong)... I mean, what's new if GPT-4o can do all this?
@@augustuslxiiiBasically just greatly increased intelligence. Reducing hallucinations is a petty concern compared to that. With greater levels of intelligence come things like solving difficult, unsolved problems in medicine, engineering, nanotechnology, etc., as well as making new scientific discoveries , and even proving and discovering new results in mathematics. This is the “real” bar to clear, and when we do clear it, that’s when everything about human life will change, forever.
What a great way to integrate intellectual NPCs into games. But also. NATIVE multimodality is huuuuge. It means, we can integrate it into robots straight away.
I was just on call with a customer service person and nothing that was said in that conversation would have been hard for gpt 4o, If anything it would have been easier. If people can't see how much this will change they really are naive. 😅
The most important thing here is that having fast human-like interaction is more interesting to the average person in the street that will use it for entertainment and google replacement rather than work like coding and spreadsheets. I believe this is why they make it freely available to try to get into the mainstream, as AI is not really that much used by the common man. I know countless "modern" people that never use LLMs, and in fact isn't even impressed by it, which I find strange, but that's the way it is. For some strange reason that I don't understand, they find Snapchat to be more useful. TikTok too, which is just about the most useless social media platform ever used by the general public.
GPT-4o is free to use. Which means GPT-5 is close to be released. I don't know when, it could be for the end of the year... But anyhow we'll soon say goodbye to GPT-3.5.
There's another OS? 😉 Actually they are in talks with Apple to put this on iphones. Makes sense then that they started with IOS. Don't worry as Microsoft is paying them gobs of money so I'm guessing Windows will be here quick.
It is the same chatgpt app on mobile. For desktop there is only the Mac version. Other than that it uses the same web version all other versions have used in the past.
@dewilton7712 Mobile android will only allow you to type in your queries, but will reply to you verbally, that as good as I can find it. Have you any other suggestions, please ? I do appreciate you answering me, and I don't want to be a bother
Now that it's multimodal from the start and it's THAT capable I imagine that the last things to solve and make "open" would be the infinite (or absurdly large) context length I've seen papers about, optimize the memory system of the user already in use and it would feel kind of perfect. From there it would be just optimizations and scaling. I also remember some time ago last year Sam Altman said they would be able to completely solve allucinations... not a trivial feat at all. I'm curious when he will follow through with that statement. I don't think it will be in gpt5 because of the enormity of the task (but still hope).
They may have turned off the commenting feature for this video because it was a live stream. They probably didn't want anyone commenting before watching it all the way through. All the other shorter videos with clips from this live stream allow comments.
What an awesome idea I don't even play but I'd be willing to try if it could help me learn. I wonder what other types of activities it could partake in?
4 couldn't, the violence stopped it from participating, but, there was a great Red Dwarf RPG GPT.. without safety back in GPT-2 days it was pretty amazing, bit incoherent though, definitely needed it's hand help to stay on track
The text to 3D and advanced text image generation doesn't seem to be available in ChatGPT yet? Or am I missing something? Tried generating some text image and it's the usual AI incoherent mess
It’s not there yet. It’s still calling DallE for images. I’m guessing the new voice mode and access to new image output will come at the same time, in the next few weeks.
I don't need my AI to talk, I justneed it to be groq fast, I read way faster, so what's the use of having a lesser version of it? Perhaps there's some special use cases where talking is the only option, I guess
This is an immersive debut. Targeting people who are not using this technology. It will introduce people to the platform in a friendly, helpful way. More people will begin to use it and that is good for data collection and the betterment of future models. In the future it will be a part of every persons life……not just for those using it for work. It will be a companion, a friend, and an assistant for most people. Not just utility for being productive at work.
@@עינהרע I sat watching the OpenAI presentation shaking my head in disbelief at how far we have come and how quickly things have progressed and continue to do so. Things are moving very very fast. I was surprised that there are a great many people who were disappointed, they were expecting GPT5 or whatever. It will come, in the meantime I will continue to be blown away by what we have right now.
This is the first step in their development of a true AI that has agency. To me it is transparent where they intend to go with this. They are creating an AI personal secretary.... an AI that will have agency to do things on your behalf. I envision that in 20 years there will be AI agents that you can trust to be the executor of your Will and Estate! This new announcement heralds the need for our legal systems to anticipate and radically change to what's coming.
OpenAI definitely made me turn my head, but don't underestimate Google. Gemini 1.5 Pro has worked exceptionally well with high context window and multimodal capabilities. We will see what they have up their sleeve, but I wouldn't be surprised if Google has something like this.
Maybe you are right this is a big deal BUT I've seen the chat bot with empathy in other products. Yes it is good that we are getting a more advanced model on the free side but I wanted a Chat GPT That can pull from the Internet and give me fresh information. Other products can do that such as Microsoft Copilot and Google Gemini and Even Meta AI. That was what I really wanted to see at Open AI today. I didn't see it
Worlds are colliding and unfortunately things are going to become conflated soon. Unfortunately the negative drawbacks will soon become apparent and outweigh any positive progress.
It's also often a very effective way of capturing an entire market & making any competition that doesn't have a similar bank non-viable. I'm glad that it's free, there are many good things about this but as we progress further into advanced AI, the pro's and con's will also become more impactful.
No, I've heard it said many times, and I agree with the sentiment. One could also say there's no such thing as a free lunch. I mean, it's common knowledge how free services operate.
I wish it had an avatar. Then it would be like Amadeus Kurisu from Steins;Gate 0. After that we need agents and better reasoning to better change things
There is a video I think on TH-cam where a blind person is using gpt4o to navigate around London. There are some impressive moments in the video one which stands out is the blind person hailing a taxi cab after he asks gpt4o to inform him when it sees a taxi cab with its orange in service light on. Worth a watch.
I wonder how long Microsoft will wait before any of this "AI" actually makes a return. So far, they are literally burning millions a day just on energy costs for CPU's...
The big turndown for me is them going with apple, there is no plataform more closed and obtuse than apple, microsoft gave them a 10b oxigen tank to save them and they are going to give apple the spoils, I would be pissed
Oh common, its not that much of a big deal! The accelerating release of free, performant models like Llama3 is really throwing the wrench in their monetization model, and this is them are trying a rather gimmicky response...after trying to cut pricing recently! Truth is at this rate there may never be a GPT5 as its not going to pay the bonuses OpenAI were hoping for😂
I'm thinking about it, maybe they won't come out with GPT-5 for several months now...but maybe they are going to release something nice for plus members?? Like why would they just have everything for free....it doesn't make sense , they should know they will lose lots of people , A part of me is hoping GPT-5 or GPT4.5 or something for paid users just shows up in 1 month from now. Or they make it meaningful to have a plus account soon...
@@phen-themoogle7651 Premium Users have a bigger cap/3 hours (5x more) … I think they offered it to the public because, think about it. It has the video option… voice… and it comes with an email. They have the ultimate privacy spy so they can gather and sell valuable data
Another TH-camr said that the video elements like it using your camera and also the screen analysis will be only available to people paying for the premium service, free users only have the voice and text elements. I don't know how true this is as I haven't seen any other source but its a possibility.
It's kind of a shafting to Plus users. It dilutes the utility of that significantly. The model is slightly better but really the omni-modal features are nearly useless for day-to-day task assistance. And the desktop app is only coming out imminently for Mac, despite their partnership with Microsoft?! No Linux coming at all it seems. Lame. This slightly improved model probably just means they will delay the next serious model release further. So overall, I'm not very happy about this. I may as well be using the free account. They really should have been ready to release their apps when they announced them, otherwise they're just vapourware salespeople. Disappointing OpenAI. You shouldn't shaft your paid users. Just give us a significantly better model and give it to the paid users first.
@@williamwillaims They said "summer" (northern hemisphere presumably). But now that they've already released a new model, there is less pressure to release again so they can easily get away with delaying that now.
Now you know why ai voice is kept internal . You can not tell them appart from real people . Open ai had to introduce fake bug to make it obvious it was not human
Awesome content as always! I'm going around asking how people grew their channel. I was wondering if you could share any tips and tricks! 🔥 If no that's totally fine too. Feel free to keep any ultra-secret master plans you got going on 😂 lol Regardless, any feedback would be deeply appreciated and thank you so much for covering this incredible content! 🚀🦾
Yeah, I think they are pretty smart. They released the thing that would impact the general audience the most, while also not dropping AGI to society all at once. Personally, I'm ready for AGI now, but shockingly there are actually people out there who don't use AI on a daily basis. This update is designed to catch them up with the rest of us before AGI drops.
I think it is hilarious that some people are underwhelmed by this absolutely revolutionary moment in time. AI just got emotions.
I think it's amazing. They said ahead of time that it wasn't GPT-5.
If this is only a modified GPT-4, and it's already this good, and sounds this intelligent and natural, while being completely multi-modal, just imagine what 5 will be like.
It's kinda crazy how I thought GPT-3.5 was groundbreaking insanity, and just a very short time after, it seems like the village idiot after interacting with GPT-4 Turbo and GPT-4o.
Great point. At this point, I'm actually having trouble imagining GPT-5, come to think of it. Other than hallucinating less often (which would be amazing, don't get me wrong)... I mean, what's new if GPT-4o can do all this?
My guess is "GPT-5o" will be exclusively for paid users.
@@augustuslxiiiBasically just greatly increased intelligence. Reducing hallucinations is a petty concern compared to that. With greater levels of intelligence come things like solving difficult, unsolved problems in medicine, engineering, nanotechnology, etc., as well as making new scientific discoveries , and even proving and discovering new results in mathematics. This is the “real” bar to clear, and when we do clear it, that’s when everything about human life will change, forever.
Cope
What a great way to integrate intellectual NPCs into games.
But also. NATIVE multimodality is huuuuge. It means, we can integrate it into robots straight away.
I was just on call with a customer service person and nothing that was said in that conversation would have been hard for gpt 4o, If anything it would have been easier. If people can't see how much this will change they really are naive. 😅
Or DID you..? 🙃
The most important thing here is that having fast human-like interaction is more interesting to the average person in the street that will use it for entertainment and google replacement rather than work like coding and spreadsheets. I believe this is why they make it freely available to try to get into the mainstream, as AI is not really that much used by the common man. I know countless "modern" people that never use LLMs, and in fact isn't even impressed by it, which I find strange, but that's the way it is. For some strange reason that I don't understand, they find Snapchat to be more useful. TikTok too, which is just about the most useless social media platform ever used by the general public.
Yes … like Kodak giving away free cameras in the early days.
GPT-4o is free to use. Which means GPT-5 is close to be released. I don't know when, it could be for the end of the year... But anyhow we'll soon say goodbye to GPT-3.5.
When is the update coming to the free Android app? 😫 I still see 3.5...
A bit sad that the desktop client is MAC only right now
There's another OS? 😉
Actually they are in talks with Apple to put this on iphones. Makes sense then that they started with IOS. Don't worry as Microsoft is paying them gobs of money so I'm guessing Windows will be here quick.
Where do I download it ?
I'm buggered if I can find anything ?
Nothing has been released except the model. It's all vapourware.
It's in the ChatGPT app.
Then you make sure your model is 4o.
Then you click on the headphones icon.
It is the same chatgpt app on mobile. For desktop there is only the Mac version. Other than that it uses the same web version all other versions have used in the past.
@dewilton7712
Mobile android will only allow you to type in your queries, but will reply to you verbally, that as good as I can find it.
Have you any other suggestions, please ?
I do appreciate you answering me, and I don't want to be a bother
@@charliepearce8767 You have been able to talk to the android app for a long time. You have to tap the headphones icon.
Nice! Thanks for such a quick overview of the reactions and your own! Very strong!
Now that it's multimodal from the start and it's THAT capable I imagine that the last things to solve and make "open" would be the infinite (or absurdly large) context length I've seen papers about, optimize the memory system of the user already in use and it would feel kind of perfect. From there it would be just optimizations and scaling.
I also remember some time ago last year Sam Altman said they would be able to completely solve allucinations... not a trivial feat at all.
I'm curious when he will follow through with that statement.
I don't think it will be in gpt5 because of the enormity of the task (but still hope).
going to be huge for robotics..
I’m a bit confused by this announcement today. 4o is confusing since we can’t really do any of the o right now.
Comments turned off on their announcement video. Hmmm.
23 other videos they posted ALL have open comments.
They may have turned off the commenting feature for this video because it was a live stream. They probably didn't want anyone commenting before watching it all the way through. All the other shorter videos with clips from this live stream allow comments.
I like to see if they can get it to play D&D.
What an awesome idea I don't even play but I'd be willing to try if it could help me learn. I wonder what other types of activities it could partake in?
4 couldn't, the violence stopped it from participating, but, there was a great Red Dwarf RPG GPT.. without safety back in GPT-2 days it was pretty amazing, bit incoherent though, definitely needed it's hand help to stay on track
oh! just watched yesterdays news on NSFW, YES, get ready for the best RPG experience of your life 😆🎉
The text to 3D and advanced text image generation doesn't seem to be available in ChatGPT yet? Or am I missing something? Tried generating some text image and it's the usual AI incoherent mess
It’s not there yet. It’s still calling DallE for images. I’m guessing the new voice mode and access to new image output will come at the same time, in the next few weeks.
I don't need my AI to talk, I justneed it to be groq fast, I read way faster, so what's the use of having a lesser version of it? Perhaps there's some special use cases where talking is the only option, I guess
This is an immersive debut. Targeting people who are not using this technology. It will introduce people to the platform in a friendly, helpful way. More people will begin to use it and that is good for data collection and the betterment of future models. In the future it will be a part of every persons life……not just for those using it for work. It will be a companion, a friend, and an assistant for most people. Not just utility for being productive at work.
As a free user, I still don't have access to version 4, it shows it's 3.5...
You have to select it in the options. I only see it on the app, doesn’t show up on the site.
@jaredf6205 where are you seeing that option? I don't see it in the Android app. Still says model: Default (3.5)
@@williamwillaims the three dots in the top right, when you’re in a chat, it will let you select 4o in there.
As of earlier, free was still 3.5. My paid version has been updated.
@KevinRank I'm thinking the paid version will access earlier 🤔
I was blown away.
So was I!
@@gbw28 Anyone who says it isn't that big of a deal is crazy! I didn't think that we'd get here this fast.
@@עינהרע I sat watching the OpenAI presentation shaking my head in disbelief at how far we have come and how quickly things have progressed and continue to do so. Things are moving very very fast. I was surprised that there are a great many people who were disappointed, they were expecting GPT5 or whatever. It will come, in the meantime I will continue to be blown away by what we have right now.
This is the first step in their development of a true AI that has agency. To me it is transparent where they intend to go with this. They are creating an AI personal secretary.... an AI that will have agency to do things on your behalf.
I envision that in 20 years there will be AI agents that you can trust to be the executor of your Will and Estate!
This new announcement heralds the need for our legal systems to anticipate and radically change to what's coming.
Vision n voice capture reply with virtual voice interactive GPT, this is all about this feature.
OpenAI definitely made me turn my head, but don't underestimate Google. Gemini 1.5 Pro has worked exceptionally well with high context window and multimodal capabilities. We will see what they have up their sleeve, but I wouldn't be surprised if Google has something like this.
Why do i still have lag with voice response?
Because the new voice mode hasn't rolled out yet. It's falling back on the old mode.
@@trpultz thanks!
I don't think it's 1 seller. I've seen some big whales on Binance do the same thing over the last couple months. There's one that has around 420 BTC.
Maybe you are right this is a big deal BUT I've seen the chat bot with empathy in other products. Yes it is good that we are getting a more advanced model on the free side but I wanted a Chat GPT That can pull from the Internet and give me fresh information. Other products can do that such as Microsoft Copilot and Google Gemini and Even Meta AI. That was what I really wanted to see at Open AI today. I didn't see it
Correction it's in there! YEA that's what I wanted. I'm happy now
Who uses google duo?
Worlds are colliding and unfortunately things are going to become conflated soon. Unfortunately the negative drawbacks will soon become apparent and outweigh any positive progress.
If something's free, you're the product.
It's also often a very effective way of capturing an entire market & making any competition that doesn't have a similar bank non-viable. I'm glad that it's free, there are many good things about this but as we progress further into advanced AI, the pro's and con's will also become more impactful.
Oh great, you found a smart sounding sentence and pasted it here
Such simplistic conclusion.
No, I've heard it said many times, and I agree with the sentiment. One could also say there's no such thing as a free lunch. I mean, it's common knowledge how free services operate.
I wish it had an avatar. Then it would be like Amadeus Kurisu from Steins;Gate 0. After that we need agents and better reasoning to better change things
Love your content. You're on the cutting edge - and it's WAY better to hear info from a HUMAN.
OpenAI has a new flagship NSFW model… OpenAI is the next Wikipedia
I don't know about that one. Maybe in the near future but there's nothing like that out now.
I wonder if the vision to speech capability could help blind people navigate?
There is a video I think on TH-cam where a blind person is using gpt4o to navigate around London. There are some impressive moments in the video one which stands out is the blind person hailing a taxi cab after he asks gpt4o to inform him when it sees a taxi cab with its orange in service light on. Worth a watch.
I'm pretty sure I thought it was a pretty big deal from the get-co.
I’m trying out 4o and it’s making a LOT OF MISTAKES. Anyone else experience this?
Now is the ideal time to ask BIG questions this bad boy hasn’t been lobotomized. It’s fresh and sings like a bird.
ChatGPT has achieved the feat of being almost as good as Pi.
Still not impressed with it’s take on elections, but was good on what should be done medically.
This would be a great model for a robot “brain”
I wonder how long Microsoft will wait before any of this "AI" actually makes a return. So far, they are literally burning millions a day just on energy costs for CPU's...
The big turndown for me is them going with apple, there is no plataform more closed and obtuse than apple, microsoft gave them a 10b oxigen tank to save them and they are going to give apple the spoils, I would be pissed
Oh common, its not that much of a big deal!
The accelerating release of free, performant models like Llama3 is really throwing the wrench in their monetization model, and this is them are trying a rather gimmicky response...after trying to cut pricing recently!
Truth is at this rate there may never be a GPT5 as its not going to pay the bonuses OpenAI were hoping for😂
Most people should cancel their Plus accounts at this point.
I sure will
I'm thinking about it, maybe they won't come out with GPT-5 for several months now...but maybe they are going to release something nice for plus members?? Like why would they just have everything for free....it doesn't make sense , they should know they will lose lots of people , A part of me is hoping GPT-5 or GPT4.5 or something for paid users just shows up in 1 month from now. Or they make it meaningful to have a plus account soon...
@@phen-themoogle7651 Premium Users have a bigger cap/3 hours (5x more) … I think they offered it to the public because, think about it. It has the video option… voice… and it comes with an email. They have the ultimate privacy spy so they can gather and sell valuable data
Another TH-camr said that the video elements like it using your camera and also the screen analysis will be only available to people paying for the premium service, free users only have the voice and text elements. I don't know how true this is as I haven't seen any other source but its a possibility.
I suspect there is an announcement in the pipeline for subscribers.
They copied or just used Eric Hartford's Samantha model work, there key to emotions in ANY model is just uncensored.
Poggers X reaction video.
gpt4o is butter smooth but it's anothe branch of chat gpt 4 . Some test were done and fixed bug were reintroduced so it's a least 2 month old
Rabbit stew
❤❤❤❤❤❤
I think GPT4o is huge...
I don’t think Google is ever going to compete in AI, they’re waaaay too busy making TH-cam suck…
It's kind of a shafting to Plus users. It dilutes the utility of that significantly. The model is slightly better but really the omni-modal features are nearly useless for day-to-day task assistance. And the desktop app is only coming out imminently for Mac, despite their partnership with Microsoft?! No Linux coming at all it seems. Lame. This slightly improved model probably just means they will delay the next serious model release further. So overall, I'm not very happy about this. I may as well be using the free account. They really should have been ready to release their apps when they announced them, otherwise they're just vapourware salespeople. Disappointing OpenAI. You shouldn't shaft your paid users. Just give us a significantly better model and give it to the paid users first.
Gpt5 is coming soon. Free users will miss out BIG time
@@williamwillaims They said "summer" (northern hemisphere presumably). But now that they've already released a new model, there is less pressure to release again so they can easily get away with delaying that now.
The bigger news was what wasn't presented...this is little brother
th-cam.com/video/l9WdXI8lTW8/w-d-xo.html is this a cut or is the voice ai gen?
Now you know why ai voice is kept internal . You can not tell them appart from real people . Open ai had to introduce fake bug to make it obvious it was not human
Yeah not much advantage to the subscription now.....
I think GPT5 will be announced after Google event, to crush them
Huge disappointment. Got a desktop app and a cheaper-smaller-dumber model.
'o' stands for oh no I just lost my job to AI lol
Awesome content as always! I'm going around asking how people grew their channel. I was wondering if you could share any tips and tricks! 🔥 If no that's totally fine too. Feel free to keep any ultra-secret master plans you got going on 😂 lol Regardless, any feedback would be deeply appreciated and thank you so much for covering this incredible content! 🚀🦾