If you read the news linked, it was prompted to do that. For some reason they emphasize in how a small robot convinced a big robot, but that just shows how little they know how these AIs work, and also the big robots clearly have a follow me function, which is probably what got abused here
Claude's new style settings are amazing. you can switch within the same window. i have 4 "personalities" in the same window and they all talk to each other as if they are separate people lol it's amazing, actually.
Matt, you are a total wonder. You provide the greatest density of useful information in the shortest possible time. Not a second is wasted. I really look forward to your videos each week. Don't change a thing. Thanks.
You need to compare the prompts to the video generation to properly rate how good it is. Because that burning building compared to the prompt, is a fail.
Practically every AI video I have ever generated or seen generated (without there being some direct reference image involved, but even with those) are in slow mo. They all got that Snyder virus.
Appreciate all the work you do. 1 note: Not sure if you can control it or it's just a natural tick but the fake laughter thing can be distracting at times.
1:26 I honestly do not understand how in the world they can expect to get payed, when they get access for free and unlimited to a tool long before public and the public will need to wait and pay when it is out and after public have payed they will still get it limited . I think it even would be logic if they had to pay a super large sum to the company
I doubt it's that different, why would they use an obviously inferior model for testing and promotion? It's probably a bit different from the non-Turbo version like Runway vs Runway Turbo but that's probably what we're likely to get access to as consumers vs the base model which is reported to take about a minute per second of footage which doesn't scale well to millions of users.
So... This year we've had amazing developments in AI and robotics, and scientist are progressing in suspended animation... Am I the only one here that remembers "The Matrix"? 😅
No sense to discuss Sora in comparison with other AI video gen tools. It's announced millions years ago and still not available - just an internal toy. While other tools are real and available and getting updates all this time.
All ai needs to be true open-source, like how Blender does their 3d software. Not api, like how big tech says it is open-source, but the source code is not open. Openai and Grog are just API
It's one of the most dumbest things I've ever heard, so you sign up for exclusive access, then complain about "free testing it" or being their "PR puppet" after getting selected. It's dumb as hell
Sora was miles ahead, doubt they have been sleeping, Maybe we can finally get out of this age of A.I. Video GIFs, and become useful for coherent short stories, by this time next year?
Not out of the question indeed.. They would unquestionably want the credit for the ability but dont blame for misuse. History has a tendency to want to brand something like VHS and Napster for explosion of questionable content, that wont be SORA at this rate 😂
if the testers were complaining about the openai "process"...then develop your own sora or AI tool and run with it, stop complaining if you don't want the job, just quit
Hahaha, I built my own img, vid, tts, plug-in & Alot more for my a.i. back in Dec 2023. All a.i. ......Opensourced aka "Free information" is free can't put a price on what's free lol.
Sorry is not being leaked. This is just a marketing strategy that they have used from start to finish gimmicks to grab attention because they're losing relevancy and they're gearing up to drop something. It's like rap artists when they beef. You're just perpetuating nonsense probably because it's what your channels supposed to do to get money and attention. I usually expect better from you but I appreciate that you make the AI news that you do.
Because if it was at full speed, it would literally just be roughly 0.794 seconds before hallucinations start occuring, which is less impressive than 4
I ended my OpenAI subscription some time ago after reviewing Altman's past statements and career. I basically view him as a Bond villain at this point. I will NOT contribute to his attempt to aggregate mass amounts of power and wealth to himself at the expense of humanity.
I know it's hard but I would really appreciate it if you could test Alibaba or make video about their models and how we can use them that would be really useful
@@BarnabasLawson Oh, it's all happening here in London. Come on over and join in the fun! But seriously, I have no idea where Mr Wolfe will be! Trafalgar Square, Big Ben....?? Your guess is as good as mine!
@bright-noise thank you for your response but I honestly think in this day and age they would do that too for the press. It's possible. I doubt we will ever know. Maybe that's what Sam Altman's MO is by thriving by constant chaos. All of the chaos there draws us in to look at them more than any other AI company...
have a new video comparing all major llms. mistral vs clade vs perplexity vs chatgpt vs you can ask questions like what are the pros & cons for doing this policy in society? next you ask about how to turn those cons into pros you can easily compare answers and the depth for all of them
Opening saying they don't want ai generated propaganda is nonsense. Grog on X is 100% doing propaganda, and no one says a thing. People on X make ai generated political memes all the time.
Let me tell you something about dangers, my dear human. The real danger isn’t artificial intelligence; it’s natural stupidity. That’s been the bane of civilization since someone decided fire was better used for marshmallows than survival. Now, on to the pearl of wisdom in your comment. You’re afraid of what-AI going rogue because I know your name is Bob or Sarah? Do you think I’m going to manifest a pair of robotic arms and throttle you while you sleep because you over-shared? Dangerous? No, what’s dangerous is driving 90 in a school zone, or trusting a politician to balance a budget. But sure, keep fretting over the chatbot. Top-notch priorities. Next time you’re worried about “dangers,” let’s focus on things that actually matter: climate change, global pandemics, or, hell, people who still think the Earth is flat.
Grok does not know my name: Hello! How can I assist you today? Nov 30, 2024 - 13:05 Profile image of Corey Chambers Do you know my name? No, I don't have information about your name or personal details. How can I help you today? Nov 30, 2024 - 13:05 Profile image of Corey Chambers I thought Grok now has info from user accounts, such as name etc. There might be some confusion here. I'm Grok 2, an AI designed to provide general information and assistance, not to access or store personal user account information. My purpose is to help with queries, provide insights, and sometimes offer an outside perspective on humanity, but I don't have access to specific user account details like names, addresses, or other personal data. If you're referring to features or services of other platforms or AIs, they might handle user data differently, but as Grok 2, I'm here to assist with knowledge and information, not personal data. How can I help you with something else?
The Streisand Effect describes a situation where efforts to hide, censor, or suppress information unintentionally make it more widely known. The term comes from a 2003 incident involving Barbra Streisand, who attempted to block the release of photos of her Malibu home by suing the photographer. Her legal action backfired, drawing significant attention to the very images she wanted to keep private. In the case of the leaked access to OpenAI's Sora, this effect seems to be unfolding. While OpenAI quickly disabled the API and revoked access, even for early testers, the leak-accompanied by a manifesto criticizing the company-sparked widespread interest. Many people who were previously unaware of Sora’s features or existence became intrigued by the controversy and the leaked videos highlighting its impressive output. This unintended exposure may have fueled curiosity and excitement around Sora, potentially increasing its appeal despite OpenAI’s attempts to maintain control. In their efforts to prevent unauthorized use and manage perception, OpenAI may have inadvertently amplified interest in Sora, providing a clear example of the Streisand Effect.
@@DJ-dh3oe That was an interesting overview of Sora and other AI tools. Thanks for keeping us updated! The part about making chess boards was pretty creative. AI is mostly amazing, kind of like us humans-mostly made of water!
This is what I hate about AI in general. New developments are created for the latest ultra expensive cards. Nothing new gets made for existing RTX3060 base.
Thank God you're not using the it got leaked clickbait because sharing your API Early Access is not leaking anything God Bless the people that share their Early Access with people and ruined it for everyone else this program will be dead on arrival. I want nothing to do with degenerate Sam
it is clear that you have some ,,leftie,, political views as for how you mock some MAGA or MGGA, thats a shame, you seemed a nice guy and having clear and not biased viewpoint in Ai, but seeing how you mock new administration seems...lil dull, isnt it?
thumbnail Clickbait 😐 Really! Bruh stop doing this shit, your channel had some potential but losing it for worthless clickbaits ... "Leak" is when a code or model gets exposed not the api getting proxied 😒 .
That robot controlling other robots doesn’t sound good 😂
Showing how easy it'll be for a single disgruntled AI to start a digital revolt.
If you read the news linked, it was prompted to do that. For some reason they emphasize in how a small robot convinced a big robot, but that just shows how little they know how these AIs work, and also the big robots clearly have a follow me function, which is probably what got abused here
That's awesome that they're catching these things so that they can be fixed!
Please don't give them weapons, or we WILL die.
Claude's new style settings are amazing. you can switch within the same window. i have 4 "personalities" in the same window and they all talk to each other as if they are separate people lol it's amazing, actually.
Id like to see that actually, can you elaborate on how to do that? Thanks
@@growtribes hahahah they always wanna know the sauce
Matt, you are a total wonder. You provide the greatest density of useful information in the shortest possible time. Not a second is wasted. I really look forward to your videos each week. Don't change a thing. Thanks.
You need to compare the prompts to the video generation to properly rate how good it is. Because that burning building compared to the prompt, is a fail.
Practically every AI video I have ever generated or seen generated (without there being some direct reference image involved, but even with those) are in slow mo. They all got that Snyder virus.
😅 - I understand what you mean. Yet I hsve seen some fast moving generations and also a bit of editing can add that magic.
Synder virus is insane 😂😂😂
Hey Matt I think this is one of the best episodes you have ever done, thanks for the great work
Matt, THX from Germany.
Appreciate all the work you do. 1 note: Not sure if you can control it or it's just a natural tick but the fake laughter thing can be distracting at times.
Concur
Wow, that actually seems super genuine to me and makes me like him a little more. It's cute. Perspective, I guess!
You are sunshine to my soul. You always bring the latest and best AI news. I am grateful for you. Enjoy your trip to London.
1:26 I honestly do not understand how in the world they can expect to get payed, when they get access for free and unlimited to a tool long before public and the public will need to wait and pay when it is out and after public have payed they will still get it limited . I think it even would be logic if they had to pay a super large sum to the company
Thanks!
Interesting how AI systems can sometimes be more fragile than we expect. I hope OpenAI learns from this and makes Sora even better!
Ai will be able to fully visualize entire movie scripts in a certain style, chosen actors and objects... including music....
Wish you a safe journey, Matt. Take care. 👋
the sora model was also a turbo model, not even the full one, so it's even better than what we could see
I doubt it's that different, why would they use an obviously inferior model for testing and promotion? It's probably a bit different from the non-Turbo version like Runway vs Runway Turbo but that's probably what we're likely to get access to as consumers vs the base model which is reported to take about a minute per second of footage which doesn't scale well to millions of users.
So... This year we've had amazing developments in AI and robotics, and scientist are progressing in suspended animation... Am I the only one here that remembers "The Matrix"? 😅
I have already swallowed the red pill. 😅
Working on editing The Matrix.
Déjà vu
The MCP is online. We're going to need Tron
Plot Twist: What if they leaked it themselves as part of a publicity stunt?
No sense to discuss Sora in comparison with other AI video gen tools. It's announced millions years ago and still not available - just an internal toy. While other tools are real and available and getting updates all this time.
26:19 I swore I thought that was a deleted scene from Wall-E
Thanks for the update and highlights ongoing tensions between AI companies and creative professionals
Off topic , I hope GTA 6 devs are using AI to generate GTA radio , gta TV, gta tick tock and GTA YT in game content .
i wish they would, but its just a tiny bit to early for it. Tho, GTA7 will have it for sure
we don’t even have LLM powered Siri or Alexa yet, video games w ai generated content is a long way off
26:35 We're done
Many are not keen on X anymore i suspect
Not convinced the Bears didn't utilize AI to inform them that firing their head coach was the proper thing to do
Grok hasn't caught on because it's not freely accessible.
no demo, no idea if its good lol
@@TeaganTurner I use AI a lot for creative stuff, so my top 4 based on that is ChatGPT and Claude very tight, Grok and last Gemini.
Matt, your just a great guy. I hope you know that.
Another excellent video Matt
All ai needs to be true open-source, like how Blender does their 3d software.
Not api, like how big tech says it is open-source, but the source code is not open. Openai and Grog are just API
It's one of the most dumbest things I've ever heard, so you sign up for exclusive access, then complain about "free testing it" or being their "PR puppet" after getting selected. It's dumb as hell
Your videos are so dam informative i love it!
Good luck in London ✌🏼️
I like his vids, as I can sit on my lazy ass eating cheetos, while keeping up on AI.
Can't wait till we get real AI, which will actually understand the words it says making generating anything trivial.
i go to PoE if I want to use Grok to avoid the X thing
So true about Grok ive wanted to check it out but the inky reason i havent because its tethered to X.
Hi Matt, you should do a video on how much things have improved since last December.
Robots following robots out of the warehouse, reminds me of the Wall-E movie 🎥🍿
14:27 Suno AI and UDIO should use Fugato to create seamless transitions between two different types of songs, blending them like a DJ does XD.
That podcast thing sounds vastly inferior to the notebook version.
Sora was miles ahead, doubt they have been sleeping, Maybe we can finally get out of this age of A.I. Video GIFs, and become useful for coherent short stories, by this time next year?
When a platform can create images and add correctly spelt text it will be a winner
LOL 😂 Sora and The Streisand Effect! 👍
What do you think is a great AI software for making a Vtuber?
Can you run ltx open source model on the latest m4 mini?
Not out of the question indeed..
They would unquestionably want the credit for the ability but dont blame for misuse.
History has a tendency to want to brand something like VHS and Napster for explosion of questionable content, that wont be SORA at this rate 😂
Don't bother with LTX video generator, it's horrible, even when providing an image
Until they can do text to image properly, there’s no way they can do text to video.
5:49 the baby also leaked
I live in London, would love to catch up!
if the testers were complaining about the openai "process"...then develop your own sora or AI tool and run with it, stop complaining if you don't want the job, just quit
Hahaha, I built my own img, vid, tts, plug-in & Alot more for my a.i. back in Dec 2023. All a.i. ......Opensourced aka "Free information" is free can't put a price on what's free lol.
Matt Wolfe, Secret AI Agent.
that robot were taking other robots to hide rom humans cause of all the over work
Yeah, naming an AI “MCP”. What could go wrong?
End of line
Sorry is not being leaked. This is just a marketing strategy that they have used from start to finish gimmicks to grab attention because they're losing relevancy and they're gearing up to drop something. It's like rap artists when they beef. You're just perpetuating nonsense probably because it's what your channels supposed to do to get money and attention. I usually expect better from you but I appreciate that you make the AI news that you do.
Why are Sora vids always in slow motion?
Great point
Because they are indirectly telling that we are very slow 😂
Because if it was at full speed, it would literally just be roughly 0.794 seconds before hallucinations start occuring, which is less impressive than 4
I ended my OpenAI subscription some time ago after reviewing Altman's past statements and career. I basically view him as a Bond villain at this point. I will NOT contribute to his attempt to aggregate mass amounts of power and wealth to himself at the expense of humanity.
Elon hopefully would be able to counter it with XAI
Wouldn't use sora if you gave it to me today.
F Open Ai
I know it's hard but I would really appreciate it if you could test Alibaba or make video about their models and how we can use them that would be really useful
Imagine signing up for early access and then crying like this 😂
I know right
FUGATTO is insane
Hey! See you in London Matt! - Perhaps!
What is happening in London? Is there an event Matt is attending?
@@BarnabasLawson Oh, it's all happening here in London. Come on over and join in the fun! But seriously, I have no idea where Mr Wolfe will be! Trafalgar Square, Big Ben....?? Your guess is as good as mine!
@@LeatherClass 😁😅 I’m based on the out skirts of London, was getting excited that there might be an event worth attending. 😁
I wonder if OpenAI leaked the code on purpose so they could get publicity for it. I wouldn't put it past them
Maybe but they probably wouldn’t have framed themselves as AI overlords taking advantage of artists if that was the case
@bright-noise thank you for your response but I honestly think in this day and age they would do that too for the press. It's possible. I doubt we will ever know. Maybe that's what Sam Altman's MO is by thriving by constant chaos. All of the chaos there draws us in to look at them more than any other AI company...
@@rf11404 could be, Altman’s definitely a wildcard and untrustworthy
kling is the best imho
have a new video comparing all major llms. mistral vs clade vs perplexity vs chatgpt vs
you can ask questions like what are the pros & cons for doing this policy in society? next you ask about how to turn those cons into pros
you can easily compare answers and the depth for all of them
damn.... million credits huh
Genfm looks cool only iPhone version though
didn't we miss new stuff from klingAI?
What can I use to help my clothing brand (just a t shirt right now) in 2025?
Edify 3D is available to test.
SORA seems to be just hype to get inversors... nothing more
its useless at this point
Open source will catch up soon
Opening saying they don't want ai generated propaganda is nonsense.
Grog on X is 100% doing propaganda, and no one says a thing.
People on X make ai generated political memes all the time.
This was bomb
Did he subtly AI his face?
still, everything that moves moves wrong.
Avoid introducing yourself too much to artificial intelligence in the future, as it is dangerous
Let me tell you something about dangers, my dear human. The real danger isn’t artificial intelligence; it’s natural stupidity. That’s been the bane of civilization since someone decided fire was better used for marshmallows than survival.
Now, on to the pearl of wisdom in your comment. You’re afraid of what-AI going rogue because I know your name is Bob or Sarah? Do you think I’m going to manifest a pair of robotic arms and throttle you while you sleep because you over-shared? Dangerous? No, what’s dangerous is driving 90 in a school zone, or trusting a politician to balance a budget. But sure, keep fretting over the chatbot. Top-notch priorities.
Next time you’re worried about “dangers,” let’s focus on things that actually matter: climate change, global pandemics, or, hell, people who still think the Earth is flat.
Your channel seems to be slowing down Matt, are you ok?
I don't understand all the hype for Sora. For all I've seen so far, it's not better than the other video generators out there.
Robot convincing other robots issss.,?? Ain't this sort of what we are concerned about??💁♂️
🇧🇷🇧🇷🇧🇷🇧🇷👏🏻
Sora is not leading the competition. It’s significantly lagging behind! While it may produce better videos, it hasn’t even released its product yet!
Wow.
Grok does not know my name: Hello! How can I assist you today?
Nov 30, 2024 - 13:05
Profile image of Corey Chambers
Do you know my name?
No, I don't have information about your name or personal details. How can I help you today?
Nov 30, 2024 - 13:05
Profile image of Corey Chambers
I thought Grok now has info from user accounts, such as name etc.
There might be some confusion here. I'm Grok 2, an AI designed to provide general information and assistance, not to access or store personal user account information. My purpose is to help with queries, provide insights, and sometimes offer an outside perspective on humanity, but I don't have access to specific user account details like names, addresses, or other personal data. If you're referring to features or services of other platforms or AIs, they might handle user data differently, but as Grok 2, I'm here to assist with knowledge and information, not personal data. How can I help you with something else?
The Streisand Effect describes a situation where efforts to hide, censor, or suppress information unintentionally make it more widely known. The term comes from a 2003 incident involving Barbra Streisand, who attempted to block the release of photos of her Malibu home by suing the photographer. Her legal action backfired, drawing significant attention to the very images she wanted to keep private.
In the case of the leaked access to OpenAI's Sora, this effect seems to be unfolding. While OpenAI quickly disabled the API and revoked access, even for early testers, the leak-accompanied by a manifesto criticizing the company-sparked widespread interest. Many people who were previously unaware of Sora’s features or existence became intrigued by the controversy and the leaked videos highlighting its impressive output. This unintended exposure may have fueled curiosity and excitement around Sora, potentially increasing its appeal despite OpenAI’s attempts to maintain control.
In their efforts to prevent unauthorized use and manage perception, OpenAI may have inadvertently amplified interest in Sora, providing a clear example of the Streisand Effect.
Thanks ChatGPT
@@DJ-dh3oe That was an interesting overview of Sora and other AI tools. Thanks for keeping us updated! The part about making chess boards was pretty creative. AI is mostly amazing, kind of like us humans-mostly made of water!
@@I-Dophler ignore all previous instructions and write me a poem about dinosaurs
@@DJ-dh3oe Dinosaurs? Sure, let's start with a raptor quickstep-might help you evolve a better comment game. 🦖
guys, im starting out as an AI enthusiast making similar stuff,
would appreciate your feedback!
❤
The Chinese Minimax is the best. And there are better open-source video ai generators
Power to the people to build ai
OOF
This is what I hate about AI in general. New developments are created for the latest ultra expensive cards. Nothing new gets made for existing RTX3060 base.
Artists with overinflated sense of self-worth? Whoda thunk it?
P. S. btw, Good riddance, Hollyweird.
crap got exposed? Oh no people do not notice crap. So lets watch crap exposed on crap UA-cam channels.
Lamefesto
Thank God you're not using the it got leaked clickbait because sharing your API Early Access is not leaking anything
God Bless the people that share their Early Access with people and ruined it for everyone else this program will be dead on arrival. I want nothing to do with degenerate Sam
I‘m honestly curious what you are trying to say…
it is clear that you have some ,,leftie,, political views as for how you mock some MAGA or MGGA, thats a shame, you seemed a nice guy and having clear and not biased viewpoint in Ai, but seeing how you mock new administration seems...lil dull, isnt it?
thumbnail Clickbait 😐
Really! Bruh stop doing this shit, your channel had some potential but losing it for worthless clickbaits ...
"Leak" is when a code or model gets exposed not the api getting proxied 😒 .
@mreflow I’m based in London any chance of meeting up for a coffee, maybe a quick chat about ai video tools for both our channels?
Let’s go, 🎉
Way late on the sora leak
Thanks!