I mean we had this for like almost 2 years now. The only difference is now it can talk to you in 'human' voice and not just text in chat window. It's still great but nothing special, I would say it won't be used more than a normal copilot.
@@beqer9281 All companies are trying to record and summarize meetings, live translation and teaching would beg to differ. This enables a lot more use cases, and the fact that it now works as it does on a standalone app, means that they kinda took over tons of startups.
Hate these phobia people making comments below, they are the ones slowing down our technology development then complaining why tech so stupid afterward
I can't wait to be able to use the Desktop app! I hated having to switch between Chrome tabs and VSCode to access GPT - the fact that I can always have it open and it can even see my screen is a game changer.
Indeed, and now we can directly talk to it makes so much more convinient. I hate typing big chunk of text to communicate. Hopefully we will get a windows version soon.
@@sulracing9710 what the fuck... months? Are we living in early 2000's? They should have it in a week! Unlimited resources, unlimited talent, and a simple desktop app that reacts to global keystrokes and can observe the screen. If I assembled a team of three, we could do it in a very robust and stable way in a week of full-time work, with the prototype done in first 3 days. Honestly, fuck them.
They got the AI to tell then that this code x[''Average Temperature].rolling(window=y).mean() calculates a rolling mean of the average temperature. And that was enough to amaze you?
@@ZoranRavicTech but still it is still useful for someone for those who arent coders. Especially good also for like Solo Game Dev who dont have such learning to code capacity, AI will help them to code.
Because we know, that the answer is correct or more correct than from most of the co-workers. i respekt "people" more if i can learn something from them.
When and where a client usually asks for a demo/PoC in 3 months? Here in Brazil they always ask for tomorrow and if you say that it will take 5 days do build it, they joke "I dont want the entire project but just a demo"
yeah but I've been testing the mobile version and it can hardly keep up saying it has issues with the system right now, even with Chat GPT4 that I paid for it seems to block me out after I send it a few PDF files to look at,
@@richarddukard8989 getting into a good university so you can get a good degree so you can get a good job so you can get a good salary so you can get a good retirement so you can die with good money... wait what
Totally agree! I want a psychical ChatGPT companion following me around, a microchipped ChatGPT into my brain, a ChatGPT girlfriend, and ChatGPT in my scrotum so I can make ChatGPT powered kids and we'll rule the world!! Muhahaha!
As someone who learns with direct feedback for corrections, this will be a gamechanger for many educational needs. I learn in weird ways thanks to my adhd. Can't wait.
Is it just me that hates using my voice with these devices? Alexa, Xbox, SkyQ, now this! I much prefer typing, especially when you’re in a public place and you don’t want everyone to hear your conversation, or late at night / early in the morning
That’s why it has the ability for both now. You don’t have to pick one or the other you get both. I personally really like the audio feature, and while I wouldn’t use it in public obviously, it would be great for a lot of other things just like the scenarios they presented in the video
@@piemasta93I know you are exagerating but its really not that big of a gap at all. For someone who types regularly audio is maaaaaaaybe twice as fast but probably not even. Plus with talking you have to iron out pauses and non-value-add works like "uh" and "er" etc etc.
What exactly do you want to do with an open source model that this can’t do? People go on about open source, but there are good reasons for the limitations on models.
@@RelaxTake30 tbh some of the other videos looked like a human was speaking especially when I hear the uh sounds, ai wont do that unless told to act a certain way
@@deadchannel8431 gen-z, on average, owns more tablets and phones than computers, when compared to millenials. all ALL gen-z people i know are compute illiterates.
I would have liked to see the vision used for the code interpretation. I'm wondering if there are issues with using it this way. Surely they would have shown it, instead of copying the code text, if it worked well.
Yes. I think they really need to embed this into the kernel space so it isn’t running OCR on pictures of the desktop, but can instead read any window you have open. I’m sure that will happen. Also they can conveniently scrape all your data (joking)
I’d like to see an open source OS with kernel level AI integration. Running a local LLM for limited kernel processes. Great progress with some 7billion param models, I’m sure it will happen. I think some sort of hybrid approach where conversation and understanding is offloaded to larger and more complex models, and the local kernel side LLM communicates with that LLM to run shell processes, find files, etc. Obviously big security concerned but it really is a case of when not if.
@@derbigpr500 It's just an integration of OCR with existing AI models, this isn't much different from what were were seeing at the launch of GPT 3.5. An important thing to keep in mind with these demonstrations is that they use VERY cherrypicked examples, the code for example was written in a way that AI just had to read the function names and it knew what was going on. The impressive thing will be when AI can read spaghetti code with poorly named variables/functions and explain what it's doing and infer WHY it's doing it just from reading the code.
@@ajay.kartheek you can try it in the playground of open ai. I also tried it and asked a coding question that other models couldn't write it correctly. GPT4o nailed it in one go. Impressive AF
And all of those smart humans are synthesized from nature. Literally nothing exists except for patterns derived from chaos. So what's the point in pointing that out?
20 years ago I read that singularity will be around 2030 and I couldn't believe it is so early. Man, we didn't have smartphones back then... singularity? Ha-ha-ha! Now I have strong feeling 2030 will be unrecognizable from anything we saw or could imagine.
Question/Concern... In a real world scenario, how much of that data/visuals is being transmitted to Open AI's (or third party) servers and how much is stored? Can the AI be run locally or in secured proprietary locations? In a real world scenario that type of process could very well be a trade secret.
Hey Chat GPT!!! Please collaborate with video game companies. Imagine playing call of duty and shouting at the battlefield, talking to your platoon. Giving commands. This is a huge breakthrough for the gaming community. Imagine applying this to openworld games like Fallout. Goddamn 🔥 For all the gamers there. Help me guys like this so they'll notice 🎉🎉🎉
@@ManiBalajiC True, but call agents dont require multitudes of voice pitch variations. That's why the call agent and voice actor are two fundamentally different jobs. Actually there's another demo emphases ChatGPT4o's substantial improvements on the voice acting.
This AI woman would be perfect for the men that are alone. Imagine how they will feel when they have a hot AI woman voice talking to them, engaging with the man's interests, encouraging them. I don't think women treat men like that anymore.
**Enhancing AI Interactivity with Audio and Video Feedback Loops** The evolution of artificial intelligence, particularly in the realm of conversational agents, has been rapid and remarkable. With the recent advancements in GPT-4O (Omni), the capabilities of AI have expanded beyond text processing to include multimodal inputs such as images and audio. However, there remains significant potential for further enhancement, particularly through the implementation of audio and video feedback loops. **The Concept of Feedback Loops** A feedback loop, in the context of AI, refers to the process where an AI system can receive and process its own outputs. For instance, when an AI generates audio responses, these could be looped back into the system, allowing it to "hear" itself. Similarly, for visual outputs, the AI could "see" its own video responses. This concept is analogous to how humans perceive their own voices and visual presence, enabling adjustments in real-time to improve clarity, tone, and emotional expressiveness. **Technical Implementation** 1. **Audio Feedback Loop**: - The AI's audio output would be fed back into its own auditory processing unit. By analyzing its own voice, the AI could adjust parameters such as pitch, tone, and volume to better match the intended emotional tone or to improve mimicry of specific voices. - This requires the integration of advanced auditory feedback systems and real-time processing algorithms to allow immediate adjustments. For instance, machine learning models trained on voice modulation could provide instant feedback and corrective measures. 2. **Video Feedback Loop**: - Similar to the audio loop, the video output generated by the AI could be fed back into its visual processing systems. This would enable the AI to assess the quality of its visual responses, such as facial expressions or gestures if anthropomorphic avatars are used. - Implementing this would involve integrating video analysis tools that can evaluate and enhance visual output in real-time, ensuring that the visual cues are consistent with the spoken content and emotional tone. **Benefits of Feedback Loops** 1. **Improved Realism**: By continuously monitoring and adjusting its own outputs, the AI can produce more human-like interactions. This is particularly important for applications requiring high emotional intelligence, such as virtual assistants or customer service bots. 2. **Enhanced User Experience**: Users are likely to find interactions more engaging and satisfactory if the AI can adjust its tone and visual cues to better match the context of the conversation. 3. **Consistency and Accuracy**: Feedback loops can help maintain consistency in voice and visual presentations, reducing the likelihood of jarring discrepancies in long conversations. **Future Directions** Incorporating feedback loops is a forward-thinking approach that aligns with the ongoing efforts to make AI more interactive and responsive. As AI technologies continue to evolve, such features could become standard, leading to interactions that are indistinguishable from human communication. The development of these systems requires collaboration between audio-visual engineers, AI researchers, and user experience designers to create holistic solutions that enhance AI's capabilities and usability. In conclusion, the integration of audio and video feedback loops into AI models like GPT-4O represents a significant step towards more natural and effective human-AI interactions. This enhancement not only promises to improve the technical performance of AI systems but also has profound implications for their acceptance and integration into daily life.
No, you don’t. What people fail to understand, and why AI is not going to take very many serious white collar jobs, is that it’s not just about producing the output. It’s also about taking responsibility for the work that’s done. An AI can never take responsibility, because it has nothing to lose. It doesn’t live. Things like engineering and medicine work with humans behind the helm because engineers and doctors suffer consequences for their mistakes. You are not just paying for their knowledge, also their ownership of their work, decisions, judgments, recommendations, etc.
@@embedded_softwarelol, is THAT why so many in upper management and politics don't output shit and get the credit of everyone beneath them. Yeah, fucking good riddance to that. Can't wait to see the death of the old world and the birth of a new one.
That's what I was thinking! They asked it what it does, but the function couldn't be any simpler, it's just calculating .rolling.mean. And they duplicated the line 3 times, to make it seem longer.
Yeah but even though it's readable to anyone who's done any amount of code, that's still beyond the capabilities of most people. It would be interesting to see these demos pushed more to their limits though.
@@kamikeserpentail3778 Wouldn't you still need to be a coder to understand what the AI is saying? In this case the code is practically an English sentence "average temperature rolling mean", so it seems like the AI is explaining it nicely, even tho it is just reading the function names and variable names. A more complicated function would require a much lengthier explanation, and might still confuse people.
the OpenAI start-up confirms the arrival of chatgpt 5 it is going to be great in addition that the current chatgpt can help the developers of OpenAI created the chatgpt 5 the current chatgpt can create code which will allow the new chatgpt 5 model to be created by example the current chatgpt can also create code to give autonomy to artificial intelligence, it's great chatGPT is really a technology that advances even its developers, it's fantastic too 👍👍👍
There needs to be a reality show where inventors compete using chat gpt to help create crazy inventions for tasks and a cash prize. Chat gpt can be used to help inventors code, create plans for other types of inventions, and basically anything. The possibilities are endless. Just a spontaneous thought though. Netflix can fund it. 😂
The voice is impressive, and the ability to interpret the image also, even though it had context. As for the rest however, it's not particularily impressive: the function is easy to understand, as is the math behind it, and the host interrupted the AI before it could answer what would probably have been the harder question.
Agreed, especially since this is the cherry picked example. It has the same problem all chatgpt solutions have right now: You need an expert to verify it‘s not hallucinating, but it mostly produces outputs only useful for non-experts.
@@Rudgged We are all using it, that is why we know it's not that good. It has a great knack of looking far more impressive than it really is and these guys are all leaning hard into that.
@@chrisstucker1813 you can still use it in text and it's incredibly fast, tbh the whole voice thing is pretty useless in practice and its mostly flexing and marketing lol
@@chrisstucker1813you can change it to any voice (maybe even a clone of yours if you are that kind of person), or you can just have it output text, either way this is as bad as it will ever be.
People needs to understand that this only help for small code base. Imagine you are developing an app and you want some help related to it Will you so all the code written in different file to chatgpt because code should be matching to its environment
To all of you saying this will change this and that etc, yes it's true but you forget one thing. These things will happen only if openai decides to release it to rhe public, they are still a private company after all which, as in every company exists for the sole purpose of making profit.
I can't share my screen macbook for GPT or doing talk and send text at the same time like a video live demo. Really need somebody give me a hand. Thanks
2:20 The implications of this are tremendous. People will be able to have the AI watch them do their job all the time - this trove of training data will then be sent to train the models so that in the future they can essentially replace every single white collar job.
Will the AI company that Apple wants to integrate into its devices be OpenAI, given that all presentations at the event are using Apple devices? Could this be a spoiler?
@dominic7012 You say these rumors with so much certainty it's funny. Chatbot is still unknown. All we know is Apple is developing its own models, and that it's also in talks with Google and OpenAI to integrate some of their stuff too. It doesn't necessarily mean "chatbot" though. It could simply mean a smarter Siri.
Apple will probably feature OpenAI's TTS AI chatbot in their latest OS as an improved voice assistant. However, as Apple heavily prioritises privacy they will try to conduct as much of the AI processing as possible on-device, hence why all the new ARM chips contain AI-specific processors
this is not the voice my chatgpt has! in my head! i think this is cool in theory, but i read way faster than anyone can talk. this would be nice for things like asking chatgpt to read you things, or asking for help repairing a car. so it def has its uses, but i wouldn't leave it on all the time. I think i would use it the most in the car. i like to write lists and think out problems and write down story ideas in the car and it's difficult to do that right now with siri/alexa. also at bedtime, same use - all my story ideas come right when i'm falling aslep, or suddenly remembering i need to do something tomorrow. also bedtime stories.
funny how apples rolling out ai soon, open ai just partnered w/ them, in the models they use all devices, the vision and new voice hasn’t rolled out yet but is specified to be soon and wwdc is right around the corner.
Still has my biggest frustration with AI which is over explaining things that it wasn't asked to do. For example 1:20, he didn't ask what the function does, just what would happen if it was/wasn't there.
actually the chatgpt helps a lot so when you are a developer its happiness also in addition it can also generate images so the chatgpt is good it has a lot of usefulness it is obvious that the chatgpt 5 will be released it is certainly in development at the moment current is then OpenAI can also use chatgpt to create chatgpt 5 because chatgpt has a lot of programming language skills 👍👍👍👍👍👍🐱🐱🐱🐱🐱🐱🐱🐱🐱👍👍👍🙂🙂🙂🙂 🙂🙂🙂
That desktop agent .. wow I imagine that AI was actually going be able to do that. Next thing you know.. it’ll be able to control your desktop do your work for you. Like installing, configuring, and even troubleshooting…. Ultimate level would an AI desktop administrator. And beyond.
They should show damo of the real life code written by developer in the 2006-2008. Since migration is the key to achieve this kind of the code. How many organisations ready to start from scratch to use generative AI????
This changes a whole industry just like that, and it’s only 2024. I cannot even begin to imagine the future of this tech in even the next 5 years.
Jeez, people, whats happening 😮😮
I mean we had this for like almost 2 years now. The only difference is now it can talk to you in 'human' voice and not just text in chat window. It's still great but nothing special, I would say it won't be used more than a normal copilot.
@@beqer9281 Yeah it's exactly like Siri...
@@beqer9281
All companies are trying to record and summarize meetings, live translation and teaching would beg to differ.
This enables a lot more use cases, and the fact that it now works as it does on a standalone app, means that they kinda took over tons of startups.
Were either headed to a Utopia or a Dystopia. The only thing that is certain is that everything will change.
Hello Skynet.. I'm a huge fans please do my grandkids no harm
🤣🤣🤣🤣🤣
Hate these phobia people making comments below, they are the ones slowing down our technology development then complaining why tech so stupid afterward
Read about "roko's basilisk"
@@PramodhRachuri Roko's basilisk is a pretty weak argument. Too many absurd assumptions.
@@TravisHi_YT I agree. But it's an interesting thought experiment
I can't wait to be able to use the Desktop app! I hated having to switch between Chrome tabs and VSCode to access GPT - the fact that I can always have it open and it can even see my screen is a game changer.
Indeed, and now we can directly talk to it makes so much more convinient. I hate typing big chunk of text to communicate.
Hopefully we will get a windows version soon.
as long as you use an Apple Product, windows had to wait few more months for the release
@@sulracing9710 what the fuck... months? Are we living in early 2000's? They should have it in a week! Unlimited resources, unlimited talent, and a simple desktop app that reacts to global keystrokes and can observe the screen. If I assembled a team of three, we could do it in a very robust and stable way in a week of full-time work, with the prototype done in first 3 days. Honestly, fuck them.
Why didn't you just use the vs code chatgpt extension? Duh..
@@matejpesl1 you can still do that app using API. Go ahead.
This is like the first Time the Internet was unveiled to us! The amazement, The fears, The Potential and all must go on.
They got the AI to tell then that this code
x[''Average Temperature].rolling(window=y).mean()
calculates a rolling mean of the average temperature.
And that was enough to amaze you?
@@ZoranRavicTech of course AI will help and assist average computer literate people who cant code to code some basic stuff for them.
@@HobeyDator It's not even coding for you, it just described some code that someone else wrote
@@ZoranRavicTech but still it is still useful for someone for those who arent coders. Especially good also for like Solo Game Dev who dont have such learning to code capacity, AI will help them to code.
@@ZoranRavicTech Well AI could already do that. I think the impressive part is the speech and convenience. It's just like you're talking to a person.
what in the word this is literally the future
It’s the present
It literally is. I literally can't believe this is real
Dont need to switch tabs to copy paste code to chatGPT = future
@@jazduh1235 It’s the past, the livestream finished hours before they posted their comment
You have not seen anything yet, my friend.
I also feel like we speak to AI more politely than we would to our co-workers when asking it to do something lol
Becoz no emotions connected, you don't hate the AI but you can hate some of your coworkers.
Because we know, that the answer is correct or more correct than from most of the co-workers. i respekt "people" more if i can learn something from them.
@@thawfeeqjamaal1777 That's bogus.
@@telekraft3175 that "people" made me laugh.
Because it knows you. Really well.
We will go from: i want a demo in 3 months
to
I want the demo next week.
We are already there my friend, this will be much more
When and where a client usually asks for a demo/PoC in 3 months? Here in Brazil they always ask for tomorrow and if you say that it will take 5 days do build it, they joke "I dont want the entire project but just a demo"
yeah but I've been testing the mobile version and it can hardly keep up saying it has issues with the system right now, even with Chat GPT4 that I paid for it seems to block me out after I send it a few PDF files to look at,
to
I want the demo NOW
One month later...
Learning and schooling is going to drastically change.
Perhaps for the better... But it is still a choice...
What’s the point of school anymore? We all have 24/7 dedicated teachers now
@@richarddukard8989 socialization obviously
@@richarddukard8989 getting into a good university so you can get a good degree so you can get a good job so you can get a good salary so you can get a good retirement so you can die with good money... wait what
@@samodelkini they can join a sports team
I want this ChatGPT 4o inside my house, in my Alexa, in my phone, inside my PC, everywhere. This is amazing
you can. Apple and open AI have signed the contract. you can have homepods. iphone. homepod mini. apple tv. watch. all with chat gpt running.
@@ReLapse85 its seems its time to get an apple product hahahaha
@@chibchan3765as always
yes have it control everything i cant wait for it to be schooling raising and feeding our children so exciting
Totally agree! I want a psychical ChatGPT companion following me around, a microchipped ChatGPT into my brain, a ChatGPT girlfriend, and ChatGPT in my scrotum so I can make ChatGPT powered kids and we'll rule the world!! Muhahaha!
that's her! only 11 years late.
Her takes place next year, so they are right on time
@@isothynebyronincredible. I noticed we’re also getting close to Minority Report, which takes place 2049-ish?
@mark9294 the world truely be unrecognisable by that time.
Nah I would not say it's late I would argue it's early and pleasantly so
As someone who learns with direct feedback for corrections, this will be a gamechanger for many educational needs. I learn in weird ways thanks to my adhd. Can't wait.
This is just so amazing, we are literally pushing and making the boundaries invisible.
Pathbreaking innovation, a game changer for sure 👍
Pathbreaking huh? 😂 Because we already broke the grounds
Is it just me that hates using my voice with these devices? Alexa, Xbox, SkyQ, now this! I much prefer typing, especially when you’re in a public place and you don’t want everyone to hear your conversation, or late at night / early in the morning
And it's inability to pronounce certain words properly. Tempachure. Also why does it need to ask how I'm doing?
That’s why it has the ability for both now. You don’t have to pick one or the other you get both. I personally really like the audio feature, and while I wouldn’t use it in public obviously, it would be great for a lot of other things just like the scenarios they presented in the video
Communication through voice is 500x faster but you can also just type
@@piemasta93I know you are exagerating but its really not that big of a gap at all. For someone who types regularly audio is maaaaaaaybe twice as fast but probably not even. Plus with talking you have to iron out pauses and non-value-add works like "uh" and "er" etc etc.
@@BeakFoundrywhat? Lol
You sound like you for No reason Dislike the Software
The best innovation happened in a long time, loving this
Im excited for open source clones of this app which we will be able to extend and build upon
What exactly do you want to do with an open source model that this can’t do? People go on about open source, but there are good reasons for the limitations on models.
@@lamsmiley1944 what i want to do is not pay for it
@@lamsmiley1944No
@@lamsmiley1944 I think there are even better reasons to not have these capabilities be centralized.
@@lamsmiley1944You know what's best for me, thank you.
Great demo. Doesn't look like super rehearsed to death at all.
Very casual, definitely not scripted
I love that AI will do all jobs! Automate everything so humans can spend time having fun and raising families!
😂 I wish
Not cybersecurity
raising families?! why increase population? humans becoming cyborgs is futuristic not biological/organic.
I hope so but we all know the chaos this will ensue
NO work NO money
From Chat Assistant to Chat Analyst .. great improvement ..
It's just describing the graph. Not something you'd really need unless you were blind. And even then screen readers already exist.
if you forget to mute it, it'll keep yapping?
Slap it across the face, works on my wife.
For real. Every demo the ai doesn't know when to shut up
No. They really want to show that you can interrupt it.
He clearly mutes the mic when he wants to talk to other people or camera (viewers)
@@RelaxTake30 tbh some of the other videos looked like a human was speaking especially when I hear the uh sounds, ai wont do that unless told to act a certain way
This will help a lot especially struggling with programming.
This is amazing, I was not expecting this to be announced so early , but it's great to see it in action so early :D :)
Her is becoming a reality.
When this becomes fully integrated into Apple products, I just can't begin to imagine what we will be able to do 6 months down the line
and when it runs on gpt5 which is long overdue
so uh this is a computer, we're gonna take out a COMPUTER for this, just in case zoomies get confused
more like boomers
lmao
what the fuck is a computer?
@@deadchannel8431 gen-z, on average, owns more tablets and phones than computers, when compared to millenials. all ALL gen-z people i know are compute illiterates.
I would have liked to see the vision used for the code interpretation. I'm wondering if there are issues with using it this way. Surely they would have shown it, instead of copying the code text, if it worked well.
Yes. I think they really need to embed this into the kernel space so it isn’t running OCR on pictures of the desktop, but can instead read any window you have open.
I’m sure that will happen. Also they can conveniently scrape all your data (joking)
@@HideousSlots Yes, what could possibly go wrong by allowing a private company full access to the contents of your personal computer.
I’d like to see an open source OS with kernel level AI integration. Running a local LLM for limited kernel processes. Great progress with some 7billion param models, I’m sure it will happen. I think some sort of hybrid approach where conversation and understanding is offloaded to larger and more complex models, and the local kernel side LLM communicates with that LLM to run shell processes, find files, etc. Obviously big security concerned but it really is a case of when not if.
@@pvanukoffhappens all the time, starting with the os.
All of this in just in a computer. This could accelerate learning beyond our comprehension.
man, it speaks now, how tf is that ground breaking?
@@lab-or5yg You probably can't comprehend how impressive this is because you don't have the slightest idea what makes it do what it does.
No lol it isn't on device - its running on their servers
@@lab-or5ygwhat?
@@derbigpr500 It's just an integration of OCR with existing AI models, this isn't much different from what were were seeing at the launch of GPT 3.5. An important thing to keep in mind with these demonstrations is that they use VERY cherrypicked examples, the code for example was written in a way that AI just had to read the function names and it knew what was going on. The impressive thing will be when AI can read spaghetti code with poorly named variables/functions and explain what it's doing and infer WHY it's doing it just from reading the code.
Keep going Openai 🚀🚀
That's just insane!
Brilliant coders writing the code for an even more brilliant AI coder that will eventually replace their coding jobs. Well done!
I just tried GPT-4o... It is so fast. Like GPT-3.5 turbo fast
How and where?
@@ajay.kartheek you can try it in the playground of open ai. I also tried it and asked a coding question that other models couldn't write it correctly. GPT4o nailed it in one go. Impressive AF
@@ArtVandelayInc Yes and it fast as hell.
@@ajay.kartheek OpenAI API
Yesterday I rewritten the whole code for a Drupal custom module with more cleaner way with many improvements with GPT4 😀😀.. life is easy now
Everyone talks about how amazing this technology is, which it is, but it's all knowledge synthesized from very smart humans.
And all of those smart humans are synthesized from nature.
Literally nothing exists except for patterns derived from chaos.
So what's the point in pointing that out?
I am speechless, maybe I should hire this gal to present for me this Thur.
Why didn't you ask it more about the "big rainfall event" annotation? 😅😅
It's there, just hidden by the GPT window on the screen (zoom in)
yup increase resolution and zoom in
Ok I was like where did it get that from.
@@xemy1010 Nice catch. I thought it hallucinated too.
Zoom in, bud
Welp this is insane it can do code analysis and explain it that’s so cool
20 years ago I read that singularity will be around 2030 and I couldn't believe it is so early. Man, we didn't have smartphones back then... singularity? Ha-ha-ha! Now I have strong feeling 2030 will be unrecognizable from anything we saw or could imagine.
Ok ChatGPT, conquer the world !🌎 ❤
We need John Connor more than ever!
nobody should be able to conquer the world
@@AndyNicholsonIf it will be for good reasons - why not?
"I am doing it right now"
"How?"
"By burning thru your precious energy"
@@topandrey14 that's what the Nazis said
Demo of Chat GPT 4o ❎ How to kick out data analysts out of job market ✅
Question/Concern... In a real world scenario, how much of that data/visuals is being transmitted to Open AI's (or third party) servers and how much is stored? Can the AI be run locally or in secured proprietary locations? In a real world scenario that type of process could very well be a trade secret.
I'm fairly sure they do everything in the cloud.
@@ZoranRavicTech sort of my assumption which is why for corporate work it is dead to me
Hey Chat GPT!!!
Please collaborate with video game companies. Imagine playing call of duty and shouting at the battlefield, talking to your platoon. Giving commands. This is a huge breakthrough for the gaming community. Imagine applying this to openworld games like Fallout. Goddamn 🔥
For all the gamers there. Help me guys like this so they'll notice 🎉🎉🎉
So it has a really good interface. We havent seen it code yet
You can already use it to code
It's a demo for assisting the development, no major updates has been released in terms of coding.
@@rutger2930 yeah the basic stuff, but it cant do the real shi
@@rutger2930 and it produces trash not a code, anything more complicated than hello world is a burning pile of garbage
its literally an O
The voice is so realistic; I dont think we gonna need voice actors anymore
if you didn't know , there are multiple companies using the AIs as their primary call agent. No need for anymore humans i suppose.
@@ManiBalajiC True, but call agents dont require multitudes of voice pitch variations. That's why the call agent and voice actor are two fundamentally different jobs.
Actually there's another demo emphases ChatGPT4o's substantial improvements on the voice acting.
This AI woman would be perfect for the men that are alone. Imagine how they will feel when they have a hot AI woman voice talking to them, engaging with the man's interests, encouraging them. I don't think women treat men like that anymore.
great idea. let's give the potential AGI to full control of all of our electronics. that's definitely not gonna backfire at all
**Enhancing AI Interactivity with Audio and Video Feedback Loops**
The evolution of artificial intelligence, particularly in the realm of conversational agents, has been rapid and remarkable. With the recent advancements in GPT-4O (Omni), the capabilities of AI have expanded beyond text processing to include multimodal inputs such as images and audio. However, there remains significant potential for further enhancement, particularly through the implementation of audio and video feedback loops.
**The Concept of Feedback Loops**
A feedback loop, in the context of AI, refers to the process where an AI system can receive and process its own outputs. For instance, when an AI generates audio responses, these could be looped back into the system, allowing it to "hear" itself. Similarly, for visual outputs, the AI could "see" its own video responses. This concept is analogous to how humans perceive their own voices and visual presence, enabling adjustments in real-time to improve clarity, tone, and emotional expressiveness.
**Technical Implementation**
1. **Audio Feedback Loop**:
- The AI's audio output would be fed back into its own auditory processing unit. By analyzing its own voice, the AI could adjust parameters such as pitch, tone, and volume to better match the intended emotional tone or to improve mimicry of specific voices.
- This requires the integration of advanced auditory feedback systems and real-time processing algorithms to allow immediate adjustments. For instance, machine learning models trained on voice modulation could provide instant feedback and corrective measures.
2. **Video Feedback Loop**:
- Similar to the audio loop, the video output generated by the AI could be fed back into its visual processing systems. This would enable the AI to assess the quality of its visual responses, such as facial expressions or gestures if anthropomorphic avatars are used.
- Implementing this would involve integrating video analysis tools that can evaluate and enhance visual output in real-time, ensuring that the visual cues are consistent with the spoken content and emotional tone.
**Benefits of Feedback Loops**
1. **Improved Realism**: By continuously monitoring and adjusting its own outputs, the AI can produce more human-like interactions. This is particularly important for applications requiring high emotional intelligence, such as virtual assistants or customer service bots.
2. **Enhanced User Experience**: Users are likely to find interactions more engaging and satisfactory if the AI can adjust its tone and visual cues to better match the context of the conversation.
3. **Consistency and Accuracy**: Feedback loops can help maintain consistency in voice and visual presentations, reducing the likelihood of jarring discrepancies in long conversations.
**Future Directions**
Incorporating feedback loops is a forward-thinking approach that aligns with the ongoing efforts to make AI more interactive and responsive. As AI technologies continue to evolve, such features could become standard, leading to interactions that are indistinguishable from human communication. The development of these systems requires collaboration between audio-visual engineers, AI researchers, and user experience designers to create holistic solutions that enhance AI's capabilities and usability.
In conclusion, the integration of audio and video feedback loops into AI models like GPT-4O represents a significant step towards more natural and effective human-AI interactions. This enhancement not only promises to improve the technical performance of AI systems but also has profound implications for their acceptance and integration into daily life.
I need to learn manual labor ASAP
No, you don’t. What people fail to understand, and why AI is not going to take very many serious white collar jobs, is that it’s not just about producing the output. It’s also about taking responsibility for the work that’s done. An AI can never take responsibility, because it has nothing to lose. It doesn’t live.
Things like engineering and medicine work with humans behind the helm because engineers and doctors suffer consequences for their mistakes. You are not just paying for their knowledge, also their ownership of their work, decisions, judgments, recommendations, etc.
@@embedded_softwarelol, is THAT why so many in upper management and politics don't output shit and get the credit of everyone beneath them.
Yeah, fucking good riddance to that.
Can't wait to see the death of the old world and the birth of a new one.
Where can I see a real world case study. This code seems fairly readable to me.
That's what I was thinking! They asked it what it does, but the function couldn't be any simpler, it's just calculating .rolling.mean. And they duplicated the line 3 times, to make it seem longer.
Yeah but even though it's readable to anyone who's done any amount of code, that's still beyond the capabilities of most people.
It would be interesting to see these demos pushed more to their limits though.
@@kamikeserpentail3778 Wouldn't you still need to be a coder to understand what the AI is saying?
In this case the code is practically an English sentence "average temperature rolling mean", so it seems like the AI is explaining it nicely, even tho it is just reading the function names and variable names.
A more complicated function would require a much lengthier explanation, and might still confuse people.
the OpenAI start-up confirms the arrival of chatgpt 5 it is going to be great in addition that the current chatgpt can help the developers of OpenAI created the chatgpt 5 the current chatgpt can create code which will allow the new chatgpt 5 model to be created by example the current chatgpt can also create code to give autonomy to artificial intelligence, it's great chatGPT is really a technology that advances even its developers, it's fantastic too 👍👍👍
son unos cracks, ojala algún día pueda charlar con alguien de su equipo.
Just needs a way to automatically access project folders and find the data it needs without us having to manually copy/paste or show it. I can't wait.
Just scan every file in a folder.... maybe that's too much power to give it!
No need to wait. Tools that can do that have existed for quite some time, like Cursor.
@@IvanRandomDude Ah interesting. Thanks for the suggestion.
Interesting that it said the y-axis was in degrees celsius
"Hey ChatGPT" is a bit of a mouthful. Please allow the users to set custom name to summon ChatGPT. It'd be awesome.
you can literaly just say hello
AI make me a sandwich
Honestly, that is really crazy, if this ain't a fake demo.
Where can we download the app from? Is it out yet?
it's not out yet, all links you can find now are scam websites
Its rolling out over the course of the next few weeks.
There’s the official OpenAI app on the App Store
oh my, This is amazing.
All that amazing tech and she still pronounces it 'Tempeh-ture' :D
I hope she also calls a Library a 'Lie-berry'
Reminded me of Toney Stark's Jarvez 💯💯
This would be a great lawyer for the public
Chat gpt desktop is only available for usa only ?????!!!!
There needs to be a reality show where inventors compete using chat gpt to help create crazy inventions for tasks and a cash prize. Chat gpt can be used to help inventors code, create plans for other types of inventions, and basically anything. The possibilities are endless. Just a spontaneous thought though. Netflix can fund it. 😂
I'm just waiting for "COMPUTER" from Star Trek. It would be the perfect "assistant"! :)
The voice is impressive, and the ability to interpret the image also, even though it had context.
As for the rest however, it's not particularily impressive: the function is easy to understand, as is the math behind it, and the host interrupted the AI before it could answer what would probably have been the harder question.
Shh you will put off the VCs!
Agreed, especially since this is the cherry picked example. It has the same problem all chatgpt solutions have right now: You need an expert to verify it‘s not hallucinating, but it mostly produces outputs only useful for non-experts.
The future is now old man, get over it
@@Rudgged We are all using it, that is why we know it's not that good. It has a great knack of looking far more impressive than it really is and these guys are all leaning hard into that.
@@Rudgged calm down kiddo, have a glass of milk while the adults are talking.
Robot simply gonna analyse human code, say piece of garbaage, wipe it out, and replace with their own, we're doomed.
It seems Skynet will be very chatty and helpful when it crushes our skulls
It will be more like Ultron. It will rizz us up before crushing us lol
definitely bunch of jobs will be lost due to the vision cap. you can live show the screen and talk, solve problems etc.
useless jobs mostly, we can focus on the real jobs now
Couldn't think of anything worse than listening to that robotic-ass voice all the time lol, drive me nuts
@@chrisstucker1813 you can still use it in text and it's incredibly fast, tbh the whole voice thing is pretty useless in practice and its mostly flexing and marketing lol
@@chrisstucker1813you can change it to any voice (maybe even a clone of yours if you are that kind of person), or you can just have it output text, either way this is as bad as it will ever be.
People needs to understand that this only help for small code base. Imagine you are developing an app and you want some help related to it
Will you so all the code written in different file to chatgpt because code should be matching to its environment
To all of you saying this will change this and that etc, yes it's true but you forget one thing. These things will happen only if openai decides to release it to rhe public, they are still a private company after all which, as in every company exists for the sole purpose of making profit.
This is great for learning
Finally I would be like Tony Stark's Friday assistance
Imagine implementing this in video games, we can start talking with NPC's in real time lol
That‘s awesome!
People: Can’t wait to use this!
Company: Sorry you have been terminated as we no longer need the headcount due to AI efficiency.
I can't share my screen macbook for GPT or doing talk and send text at the same time like a video live demo. Really need somebody give me a hand. Thanks
Some features are not available yet.
Idk, ChatGPT already does this without the voice. It's an improvement alright but for me it's already doing what I want.
I don't see how this helps you in coding, the moment you write a function you already know what it should do otherwise you wouldn't write it.
It is for people who don't know much about coding and get sent data like this and want a quick interpretation of what it does
Could be useful when debugging or trying to understand someone else's spaghetti
2:20 The implications of this are tremendous. People will be able to have the AI watch them do their job all the time - this trove of training data will then be sent to train the models so that in the future they can essentially replace every single white collar job.
OpenAI I beg you very much, can I have early access to this function, I really want to help with the test of this model!
Can this AI regrow hair on my head???
i think you need prayers for that
That’s going to be on the gpt 5 update but they are working on it.
this is big 🔥
Is the macOS app already available?
The app should have been released with the announcement. What's the point of a multi-week wait?
What are these chairs? I want one.
Edit: It's a Tacchini Costela, 4k$.
Wow.. Amazing 👏👏👏
Will the AI company that Apple wants to integrate into its devices be OpenAI, given that all presentations at the event are using Apple devices? Could this be a spoiler?
prob not since microsoft paritally owns open ai
Apple has its own AI, on-device LLM (in house), and cloud-powered LLM (in house), but for the chatbot, it collaborates with OpenAI.
@dominic7012 You say these rumors with so much certainty it's funny.
Chatbot is still unknown. All we know is Apple is developing its own models, and that it's also in talks with Google and OpenAI to integrate some of their stuff too. It doesn't necessarily mean "chatbot" though. It could simply mean a smarter Siri.
Apple will probably feature OpenAI's TTS AI chatbot in their latest OS as an improved voice assistant. However, as Apple heavily prioritises privacy they will try to conduct as much of the AI processing as possible on-device, hence why all the new ARM chips contain AI-specific processors
They watched "Her" and was like yup we can do that.
Woww another favourite feature❤
this is not the voice my chatgpt has! in my head! i think this is cool in theory, but i read way faster than anyone can talk. this would be nice for things like asking chatgpt to read you things, or asking for help repairing a car. so it def has its uses, but i wouldn't leave it on all the time. I think i would use it the most in the car. i like to write lists and think out problems and write down story ideas in the car and it's difficult to do that right now with siri/alexa. also at bedtime, same use - all my story ideas come right when i'm falling aslep, or suddenly remembering i need to do something tomorrow. also bedtime stories.
just ask to talk faster lmao
You can adjust playback speed under settings.
At 3:08 the real woman got replaced by AI. She will probably get cats living with her at home next.
funny how apples rolling out ai soon, open ai just partnered w/ them, in the models they use all devices, the vision and new voice hasn’t rolled out yet but is specified to be soon and wwdc is right around the corner.
Well, I have to say it's just great that they're using a female voice for ChatGPT
Still has my biggest frustration with AI which is over explaining things that it wasn't asked to do.
For example 1:20, he didn't ask what the function does, just what would happen if it was/wasn't there.
actually the chatgpt helps a lot so when you are a developer its happiness also in addition it can also generate images so the chatgpt is good it has a lot of usefulness it is obvious that the chatgpt 5 will be released it is certainly in development at the moment current is then OpenAI can also use chatgpt to create chatgpt 5 because chatgpt has a lot of programming language skills 👍👍👍👍👍👍🐱🐱🐱🐱🐱🐱🐱🐱🐱👍👍👍🙂🙂🙂🙂 🙂🙂🙂
Ok this is amazing but it is not anywhere near replacing engineers yet.
my job has an expiry date now. thanks obama
OpenAI makes answering lame coding interview questions even easier
I don't think I've ever had an interview where the technical questions weren't easier than the social questions and self aggrandizing dance if words.
This is insane 😮
That desktop agent .. wow I imagine that AI was actually going be able to do that. Next thing you know.. it’ll be able to control your desktop do your work for you. Like installing, configuring, and even troubleshooting….
Ultimate level would an AI desktop administrator. And beyond.
Relax, all they shown is it reading stuff from screen. Hardly more useful that the screen reader blind people already have.
I question if we, citizens should be allowed to have this in our arsenals.
They should show damo of the real life code written by developer in the 2006-2008.
Since migration is the key to achieve this kind of the code.
How many organisations ready to start from scratch to use generative AI????
This is marketing at its finest. Happy and diverse people enjoying the tool that makes people even more lazy :)