People whining in the comments section as always. All these updates reflect a great mindset at Openai, ie they still see the need to keep developing the product, even if in incremental steps. I wouldn't even mind if the next update is a punctuation corrector. I respect that. Well done guys. One line of code at a time.
@@smokeyfish7435 By the people that invest Billions into them.This is serious business, not a charity. You either do groundbreaking work or you become irrelevant. Google and Anthropic will eat them if they don't show up with something real soon.
Everybody is entitled I would love to see them take leadership at a company like this just to find out they can't do any better no matter how much they tried.
The reason today and yesterday don't appeal to people is simple: Only 15% of the users have a mac or iphone. These products are not reaching the other 85% of the customers. I think those designers al have a MAC pc and forget that the rest of the world is on Windows and Android. From a commercial perspective this is a disaster because they neglect most of their potential customers by putting so much effort in something that only reach 15% of potential customers. I hope they realize this at some point... They should prioritize where the manjority of their customer base is located, that would be the most cost efficient. Now those 85% were dissapointed two days in a row (first IPhone and now MAC). I hope tommorow is not about another Macintosh product...
Mac for whatever reason is the priority I believe. I'm just using Google's AI Studio for the screen sharing tool. Hopefully Windows gets that upgrade soon.
Yes it is odd, they started with MAC os only. Only 15% of pc's are MAC, what do the other 85% have today with this announcement? Yesterday same problem, they start with an Iphone announcement, while 80 off all the phones in the world are ANDROID ...
@@dbab-liftingjourney7351 They are official apple partners, they are eventually going to be mac exclusive. Apple wants to get people to rely on it and then strip it away and make it a mac selling point
People don’t understand that all these product features will be absolutely essential as prerequisites to more advanced models. Once the feature set is more complete all of those features automatically get better with each new model. This is a great strategy
The voice with the notes is great for us fiction writers who would like ChatGPT to act out parts of a script; one of the things I was wanting. I hope the last day will be agents, ChatGPT 5 or both.
No, because the process of implementing Sky's voice was marketing, i.e. a circus to draw attention to the movie and the development of the Advanced Voice, we have already implemented more voices that are much cheaper. Say goodbye to Sky, she's not coming back unless the actress doesn't charge us so much money.
I have the Plus subscription and I do not have access to Canvas on my desktop. No idea why. I am located in Trinidad and Tobago, so maybe it is a region thing?
Linux? why would they even bother, if it is GTK all the KDE people will hate it. If it is electron every one will despise and let negative reviews if it uses Wine they will receive funny emails. No sane developer will develop a desktop app for linux.
@@afterglow5285 Oh, of course, they market heavily to programmers and then exclude the OS that’s almost entirely populated by them. I think I get why-no beginner who depends on AI would use Linux.
Justin and John "sleigh" in those suits. I also love the hidden Eastern egg of the rotary phone in the upper right hand corner. Nice throwback to yesterday's 1800 CHAT GPT video. Great demo. And, killer saxophone solo. Most useful part for me will be the integration with writing tools.
I would like to one day just open up gpt on my pc and say “ hey, I need help organizing my pc. Place all the photos in one folder and organize them by year. Also find any items I use for Blender throughly my pc and bring them into one folder and organize it by projects” something like this cause I save a bunch of stuff but not very organized about it lol
Unfortunately, that is generally almost impossible. The current feature relies on the MacOS accessibility API, which means the target app must implement this API to access its UI elements and contents. A very resource-intensive workaround involves parsing the screen to determine coordinates for interacting with the apps. It is possible that in the future many apps will have the incentive to implement the accessibility API, but it is still a bit intrusive because the ChatGPT app can read a lot and right now, they can use that data for training their models.
You can make your own agent with CustomGPT and API calling and it will use chatGPT subscription instead of API calls so you have large context and you are paying only the subscription not the api usage.
there are some features ahead on the windows app right now... i believe the canvas app is available in the windows app but not the mac app. first time I've seen windows come out ahead with openai
This video is incredible and highlights the fascinating advances in AI. I’ve been exploring the idea that intelligence might be a universal property of complexity, not just a human-centric concept. By applying principles like the Renormalization Group (RG) and holography, we could better understand how AI systems organize complexity and potentially unlock their full potential in a safer and more interpretable way. It’s a speculative direction, but I believe it could lead to breakthroughs in AI safety and design. I’d love to see OpenAI dive into these interdisciplinary ideas!
Yesterday was a bit disappointing for developers. I got it, which is great for common people. But today you are back ! Hopefully tomorrow is a great day !
Well done to all at OpenAI! These live streams have been so much fun to watch along with and the new features are just awesome! Can’t wait to see what you have planned for tomorrow. 🎄
“Available today on our Mac desktop apps, coming _very soon_ to Windows.” Don’t be a brat. Or listen to the whole thing. Or both. Oh, wait, I forgot what community I’m talking to.
It’s not even just one child, it’s one half of one child. I have an intel MacBook and can’t use it because it only works on Apple Silicon Macs and MacBooks.
One thing to acknowledge: The voice mode is just one commit away from seriously making physically disabled people work with their voice in any document / code editor
I appreciate the direction you are going in with these with text based app integrations. I wonder if there are other users like me that would really like to see it work with Obsidian?
Most of world population use windows not OS. If you really want to capture the world, you MUST GO with Windows. Period! Get out of your own bubble please :)
This is the most frustrating thing. I am not going to replace my Intel Mac any time soon, and in Xcode building a universal binary is the DEFAULT. They had to choose deliberately not to support us.
This doesn't seem to work. It says "Working with Notes" but doesn't see anything in my notes app. I tried using advanced voice with the connected Notes app, but it can't read anything
@@travisgoesthere All the ones available aren’t official and work by using your API key. The ones that don’t use an API key are basically just web apps. I can guarantee there’s no official ChatGPT app, and you can’t even use Wine because the Mac version uses a different architecture, and the Windows version can only be downloaded directly from the Microsoft Store.
This was even more disappointing than yesterday! I hope tomorrow they don't announce "chatgpt app is available in blackberry phones now" or "advanced voice mode is now available in the vatican"
GPT has already been integrated with unique via the Dialogue System and .NET has HTTP features to make API calls if you want to integrate AI into your games. If you mean 3D generation, Unity is working on this with their own AI called Sentis I think.
не совсем понятно - canvas по-определению должен максимально возможное количество параметров объединять из максимально возможного числа источников, по-максимуму разнородных. почему такое искусственное разделение?
It looks like the Mac-app has whisper (speech to text). This is not available in the windows app? (Only voice mode) Please add it for the web and Windows app!
My educated guess on Day 12 will be a new larger model, GPT-4.5 or GPT-5, trained from the ground up. The model is being trained since January and if it took 6 months to finish training and 6 months of safety testing, it should be ready by now. It would also make sense if only the $200 tier gets access to it, because even with all the features showcased from day 1 to day 11, $200 still seems like it's not worth it.
This doesn't automate anything if you still have to copy and paste.
At least they did say that they're working on it.
yeah, it must be the next step. but voice help in the background is great tho
next step will be chatgpt takes control of mouse and keyboard .....get ready
@@ankhenaten2 Cursor and text input.
exactly my thought, specially when I use cursor
that saxophone performance is pretty lit ngl
People whining in the comments section as always. All these updates reflect a great mindset at Openai, ie they still see the need to keep developing the product, even if in incremental steps. I wouldn't even mind if the next update is a punctuation corrector. I respect that. Well done guys. One line of code at a time.
Half of these stuff was promised 6 months ago
Bro they're expected to make large steps in research, not release minor features for ChatGPT.
@@Wegnerrobert2expected? By who? You sound entitled
@@smokeyfish7435 By the people that invest Billions into them.This is serious business, not a charity. You either do groundbreaking work or you become irrelevant. Google and Anthropic will eat them if they don't show up with something real soon.
Everybody is entitled I would love to see them take leadership at a company like this just to find out they can't do any better no matter how much they tried.
Day 12: your agi is in another castle.
you were always the agi you were looking for
day 12: agi is now on gemini!
The reason today and yesterday don't appeal to people is simple: Only 15% of the users have a mac or iphone. These products are not reaching the other 85% of the customers. I think those designers al have a MAC pc and forget that the rest of the world is on Windows and Android. From a commercial perspective this is a disaster because they neglect most of their potential customers by putting so much effort in something that only reach 15% of potential customers. I hope they realize this at some point... They should prioritize where the manjority of their customer base is located, that would be the most cost efficient. Now those 85% were dissapointed two days in a row (first IPhone and now MAC). I hope tommorow is not about another Macintosh product...
I agree that for Iphones the percentages are higher, especially in the united states. But outside, the world is ALL android.
@@henry_room the G in AGI stands for Gemini afterall
Hope the Windows app gets updated soon, this looks coo, thanks openAI
Mac for whatever reason is the priority I believe. I'm just using Google's AI Studio for the screen sharing tool. Hopefully Windows gets that upgrade soon.
Yes it is odd, they started with MAC os only. Only 15% of pc's are MAC, what do the other 85% have today with this announcement? Yesterday same problem, they start with an Iphone announcement, while 80 off all the phones in the world are ANDROID ...
@@dbab-liftingjourney7351 They are official apple partners, they are eventually going to be mac exclusive. Apple wants to get people to rely on it and then strip it away and make it a mac selling point
@@sebkeccu4546 Take a look at the machines that the OpenAI devs use: almost uniformly Apple. Very much like many cutting-edge development studios.
@@sebkeccu4546 apple users have more money) in average)
People don’t understand that all these product features will be absolutely essential as prerequisites to more advanced models. Once the feature set is more complete all of those features automatically get better with each new model. This is a great strategy
Not on apple. Way to feed power to the wealth and apples marketing
welcome to how economy works for the last 2000 years. the rich get riches, as you say.
@@TheAuroraSynapse биология работает еще дольше - в конечном счете у всех одинаковые пустые карманы
People are so impatient
I can say, TOO SLOW. I will ship on the Google ship now
What about windows😢
18:55
@@The_Collective_I Got excited. Then I realized.
People are tired of windows, things constantly break and it has bugs
Yes we're tired of windows, but some apps exist only in windows@@Wanderer2035
And most, self included hate apple. Wonder how much they paid to rush these apple features @Wanderer2035
Still waiting for the advanced voice from the demo, which could sing and imitate voices...
advanced voice mode from the demo couldn't imitate any voice
Pretty sure this was something they specifically didn't want it to do as it could make it much easier to scam using the generated voice.
@@PedroPenhaVerani-ll1wc yes it can
ask putin to make an AI
It's really good at language and accents tho!
The voice with the notes is great for us fiction writers who would like ChatGPT to act out parts of a script; one of the things I was wanting. I hope the last day will be agents, ChatGPT 5 or both.
More like GPT 4.5 not GPT 5
I came to say the say thing 4.5 is my guess as well
Except that if you try to get it to do character voices, the guardrails jump in and block it :(
Can we a Linux desktop app please
Linux 😂😂😂
Also available on Palm Pilot 😂
The OS for people who enjoy their privacy @Jules-hn6un
hu-hu-hu-hooo. Maybe the next year christmas hoo-hooo
They announced you can call via rotary phone yesterday.
@Jules-hn6unI wonder if they realize that half of the humanity they’re doing it for isn’t worth it.
Are we gonna get Sky back?! Also please fix the interruption bug.
No, because the process of implementing Sky's voice was marketing, i.e. a circus to draw attention to the movie and the development of the Advanced Voice, we have already implemented more voices that are much cheaper. Say goodbye to Sky, she's not coming back unless the actress doesn't charge us so much money.
@@KARA-IA Scarlett Johansson is gonna fk you all up.
@@KARA-IA NOOOOOOOOOOO!! REEEEEEEEEEEEEEEEEEEEEEEEEEE!!!!!!
Tomorrow better be really good. The best part of this was the sax solo at the end.
my thumbs up was just for the saxophone
Here's to hoping it's going to be something like Gemini 2's native image output. I'm itching to get my hands on something like that!
Google’s now the king of AI. I’m expecting them to close the gap with o1 in due course.
I have the Plus subscription and I do not have access to Canvas on my desktop. No idea why. I am located in Trinidad and Tobago, so maybe it is a region thing?
Day 11: Windows users... Nothing :')
2 days now. Let's give this new powerful tech to the wealthy first, instead of showing the little guy they can finally have movement.
Day 143: Linux users...💀
I forgive that there was not a windows version ready that supports this, just because the sax in the end was so great 🎶🤩
ChatGPT cannot top that saxophone performance 🙌
Windows users...
they said coming soon to windows
We don't need this ! At least I don't need it.
Sax solo sleighed sharply. Bravo!
can you use this for windows? arent yall owned by microsoft lol what tf is going on
damn. guessing im switching to gemini. never thought this would happen
ChatGPT (nor OpenAI) isn't owned by Microsoft. Microsoft is a major investor though. ChatGPT desktop is also available for Windows.
@@felicitatislibertatisque yeah but i dont see the option to do what they are doing
Girl, I'm saying
@@PremoSaysThings hey does Gemini have something that can interact with your apps that I’m not aware of??
Where is the Linux desktop app?
Linux? why would they even bother, if it is GTK all the KDE people will hate it. If it is electron every one will despise and let negative reviews if it uses Wine they will receive funny emails. No sane developer will develop a desktop app for linux.
The AI will make one eventually
@@afterglow5285 Oh, of course, they market heavily to programmers and then exclude the OS that’s almost entirely populated by them. I think I get why-no beginner who depends on AI would use Linux.
Justin and John "sleigh" in those suits. I also love the hidden Eastern egg of the rotary phone in the upper right hand corner. Nice throwback to yesterday's 1800 CHAT GPT video. Great demo. And, killer saxophone solo. Most useful part for me will be the integration with writing tools.
I would like to one day just open up gpt on my pc and say “ hey, I need help organizing my pc. Place all the photos in one folder and organize them by year. Also find any items I use for Blender throughly my pc and bring them into one folder and organize it by projects” something like this cause I save a bunch of stuff but not very organized about it lol
@@DominequeTaylor OMG if only! That would be the best.
Unfortunately, that is generally almost impossible. The current feature relies on the MacOS accessibility API, which means the target app must implement this API to access its UI elements and contents. A very resource-intensive workaround involves parsing the screen to determine coordinates for interacting with the apps. It is possible that in the future many apps will have the incentive to implement the accessibility API, but it is still a bit intrusive because the ChatGPT app can read a lot and right now, they can use that data for training their models.
I'm like this with my browser tabs
Us non-MacOS users got coal from OpenAI today lol
XD
Hahaha
😂
Only ~84% of the world doesn't use macOS, so it makes sense to make them wait a bit.
Which is most of the population
Not even a word for windows users who are 80% of computers users in the world.
Tomorrow better be agents.
Or what
Probably best for last l hope
nope. they just said that it will be next year
This is the prelude to agents, have to assume tomorrow is 4.5 else they've been completely outclassed by Google
You can make your own agent with CustomGPT and API calling and it will use chatGPT subscription instead of API calls so you have large context and you are paying only the subscription not the api usage.
Hello OpenAI,
When will the "Find message" option return to Windows (app and website)?
Very useful tools introduced today! Plus, it was really fun the way you guys presented it.
Intel Mac and Windows support when?
*It's my birthday. Happy birthday, me!* 🥳
Happy birthday my dude
happy birthday
🎂
And I have no idea why I announced it here. Guess I'm just excited about the 12 Days of OpenAI :)
happy birthday bro
was not expecting a sax performance!! great job man
…. Is this only available on Mac? What we gotta wait a year for it to be available on windows?
Or forever, they have a contract with Apple, and Apple is just so much better anyways everyone should switch
there are some features ahead on the windows app right now... i believe the canvas app is available in the windows app but not the mac app. first time I've seen windows come out ahead with openai
We are working on Windows support!
@@aaronsander8333 I can confirm that mac version still do not have canvas support.
Exactly. Open ai. Ai for all, just the wealth first
This video is incredible and highlights the fascinating advances in AI. I’ve been exploring the idea that intelligence might be a universal property of complexity, not just a human-centric concept. By applying principles like the Renormalization Group (RG) and holography, we could better understand how AI systems organize complexity and potentially unlock their full potential in a safer and more interpretable way. It’s a speculative direction, but I believe it could lead to breakthroughs in AI safety and design. I’d love to see OpenAI dive into these interdisciplinary ideas!
Not gonna lie, that boy can play! (and nice work on the desktop app)
Mac OS only. Should’ve known 😫
Please fix the site interface for phones. The new function is not displayed in the vertical screen.
Can we please get a desktop app for deb and rpm linux? or a flatpak?
Great surprise at the end. That sax is awesome.
Thank you for all your hardwork and innovation ❤
These two guys in suites are amazing - they should do more demos!
Yesterday was a bit disappointing for developers. I got it, which is great for common people. But today you are back ! Hopefully tomorrow is a great day !
Nice ending with the saxophone solo!
🔥🔥🔥 The end of the demo is fire :)! The rest of it is great too..😅
Well done to all at OpenAI! These live streams have been so much fun to watch along with and the new features are just awesome! Can’t wait to see what you have planned for tomorrow. 🎄
The greatest saxophon player ever....👍
This only works on apple? If so, it is useless.
This demo shows exactly why code generated by LLM needs to be checked in order to keep quality high
What quality?just because it is checked doesn’t mean quality. Did you not see any code before 2023?
so now we are giving christmas gifts to one child and not the other.... ugh
macOS is just the better OS
@StarChaser1879Better in what ? Beeing locked in a Walled garden
“Available today on our Mac desktop apps, coming _very soon_ to Windows.” Don’t be a brat. Or listen to the whole thing. Or both.
Oh, wait, I forgot what community I’m talking to.
It’s not even just one child, it’s one half of one child. I have an intel MacBook and can’t use it because it only works on Apple Silicon Macs and MacBooks.
@@carinaschoppe better for most technological and artistic needs, windows is better for the average consumer though.
Tomorrow better be AGI lol
No chance
Probably Orion
i am down with v5
I really didn't understand what they wanted to disclose today. Very confusing.
me too
p
Something that is very useless to someone that is not a writer nor a coder
You're definitely not the only one 😂
you guys are killing it❤ love the progress.
that sax was a pretty cool way to close it out
Best live to date 🎉 Ty for the sax
what about windows users? 😞
One thing to acknowledge: The voice mode is just one commit away from seriously making physically disabled people work with their voice in any document / code editor
Computer use by Anthrophic is superior if you ask me
Great sax solo!
Came for gpt, stayed for the sax
Does this mean OpenAI can "listen" to all copy buffers even if not intended for OpenAI?
The Windows version is not ready. **Microsoft is holding the $11 billion it gave OpenAI and looking enviously at the Mac**
Vous êtes des tueurs❤. C'est vraiment excellent. Dans la présentation comme dans les fonctionnalités. Bravo à vous en attendant le 12eme jour.🎉
Considering I use both Linux and Windows in my projects this is useless for me, but thanks.
Does the 200$ plan support canvas in combination with the o1 model?
i hoped for gpt 4.5 tomorrow and new image generation tools today. It looks like google is taking lead..
These 3 were great presenters.
It would be so cool to eliminate the command + C, command + V
I appreciate the direction you are going in with these with text based app integrations. I wonder if there are other users like me that would really like to see it work with Obsidian?
Most of world population use windows not OS. If you really want to capture the world, you MUST GO with Windows. Period! Get out of your own bubble please :)
Please release desktop for Linux
Hey that saxophone player! Come back tomorrow for another show!!🎉
Sax solo for day 12?
Pleeease allow it access to calendar!
Cool Sax Solo!
Great present! I loved the last saxophone performance.
When are you going to release a mac (intel-based) app?
This is the most frustrating thing. I am not going to replace my Intel Mac any time soon, and in Xcode building a universal binary is the DEFAULT. They had to choose deliberately not to support us.
🦕
@@JustSuds Yea it's really frustrating.
Really cool sax solo! Also: please ship for linux too.
This doesn't seem to work. It says "Working with Notes" but doesn't see anything in my notes app. I tried using advanced voice with the connected Notes app, but it can't read anything
They gonna need more than funny hats to catch up with Google these days.
Is there a way to see the Projects folders in the native chatgpt app? I can see the folders on the browser site but not on the native app.
Great saxophone, please do more :)
A+ for live music
Will there be a release of the desktop app for older macOS versions (e.g., 12.x) and Intel hardware? What about Linux?
Integrate with a Task List, Maps Service and Calendar, and you have yourself a new Executive Assistant also known as Donna.
Loved the musical ending! Thanks so much ❤
You are wonderful, beautiful people. Beautiful people build good things and things that make the world beautiful. Thank you for developing Chatgpt.
sax was epic
Nice presentation Justin and John! The features presented today are looking awesome. Windows App please!
Please, we need Linux app
There already is. Dona bit of research
@@travisgoesthere All the ones available aren’t official and work by using your API key. The ones that don’t use an API key are basically just web apps. I can guarantee there’s no official ChatGPT app, and you can’t even use Wine because the Mac version uses a different architecture, and the Windows version can only be downloaded directly from the Microsoft Store.
I’m excited for tomorrow!
Come on, you're so close to letting us talk to our IDE and code hands-free. hurry up!
This was even more disappointing than yesterday! I hope tomorrow they don't announce "chatgpt app is available in blackberry phones now" or "advanced voice mode is now available in the vatican"
"Live demos work 100 percent of times. The rule of the universe"😅 I got that one
I hope the next step is being able to work with gaming engines with multiple scripts working together.
GPT has already been integrated with unique via the Dialogue System and .NET has HTTP features to make API calls if you want to integrate AI into your games.
If you mean 3D generation, Unity is working on this with their own AI called Sentis I think.
не совсем понятно - canvas по-определению должен максимально возможное количество параметров объединять из максимально возможного числа источников, по-максимуму разнородных. почему такое искусственное разделение?
Does anyone know if there are any plans to create a linux desktop app?
So can it simply open a next episode of the series I watched yesterday?
When can I finally use the realtime voice to control my computer completely?
Is there any hope that we get an update on the windows app this year?
It looks like the Mac-app has whisper (speech to text). This is not available in the windows app? (Only voice mode)
Please add it for the web and Windows app!
My educated guess on Day 12 will be a new larger model, GPT-4.5 or GPT-5, trained from the ground up. The model is being trained since January and if it took 6 months to finish training and 6 months of safety testing, it should be ready by now. It would also make sense if only the $200 tier gets access to it, because even with all the features showcased from day 1 to day 11, $200 still seems like it's not worth it.
Just out of curiosity should we wait the next economical crisis before the windows release or after it?