SORA Demo FAKED? Elon Musk’s 18 billion. New AI Characters, New Humanoid Robot
Вставка
- Опубліковано 28 тра 2024
- Link to the website:
www.talkie-ai.com/
Discord: / discord
TikTok: / talkiedoki
Reddit: / talkieofficial
Instagram: / talkie_app
X/Twitter: / talkie_app
Talkie #TalkieAI
How To Not Be Replaced By AGI • Life After AGI How To ...
Stay Up To Date With AI Job Market - / @theaigrideconomics
AI Tutorials - / @theaigridtutorials
🐤 Follow Me on Twitter / theaigrid
🌐 Checkout My website - theaigrid.com/
Links From Todays Video:
/ 1
finance.yahoo.com/news/jensen...
/ 1783559575972524389
/ 1783342877722210473
/ 1783535861214204368
www.thebaltimorebanner.com/ed...
/ 1783480617646031121
www.fxguide.com/fxfeatured/ac...
/ 1783541588846608519
www.theinformation.com/articl...
Welcome to my channel where i bring you the latest breakthroughs in AI. From deep learning to robotics, i cover it all. My videos offer valuable insights and perspectives that will expand your knowledge and understanding of this rapidly evolving field. Be sure to subscribe and stay updated on my latest videos.
Was there anything i missed?
(For Business Enquiries) contact@theaigrid.com
#LLM #Largelanguagemodel #chatgpt
#AI
#ArtificialIntelligence
#MachineLearning
#DeepLearning
#NeuralNetworks
#Robotics
#DataScience - Наука та технологія
![[柴犬ASMR]曼玉Manyu&小白Bai 毛发护理Spa asmr](http://i.ytimg.com/vi/0TsXQ7z2Dh4/mqdefault.jpg)
I thought it was pretty clearly communicated that the balloon guy video was made WITH Sora, not BY Sora.
going back now to the original video, it says this high up in the video description:
> Based in Toronto, Shy Kids is a multimedia production company that utilized Sora for their short film about a balloon man.
UTILIZED!
do people not fucking read anymore?
I do NOT consider that to be "faking" the demo. What they did is EXACTLY how a video editing professional, would've used such AI tools to create such a piece. It may be a bit disingenuous for them to have silently suggested it was entirely AI generated.. But anyone who was aware of how these tools work, knew there had to be some post editing and cleanup, to tie all the clips together.
I agree
I agree with you that its impressive as hell. But the general public don't know these things and Open AI are selling them a dream they can't have.
Well what it does mean is that sora was actually incapable of generating any of the scenes, re rolling didn't fix it
This is not a tech demo, it's an expose of what artists can do with a tool.
Airhead was not a SORA Demo. It was done by an independent film production company with early access to Sora. So, not sure why anyone is making a big deal about the use of props in the "Shy Kids" video-- it was not an official SORA release, it was done by a separate team. The backgrounds were SORA generated, and I could care less if the balloon was "real" or not. Could they have been a little more forthcoming, sure. I do agree that the lack of transparency makes them look bad given the way it was marketed.
With regards to Billions of dollars invested in AI, use it to build FAB plants... and make chips, or the huge demand for AI will result in a complete inability to sustain the market.
Good to see some improvements with robot design.
I think this is a pretty big deal, I was under the impression that It was all Sora, and I think that's how they sold it.
We really shouldn't be surprised. Anyone who used critical thinking and knew how editing work, had to assume some post AI editing took place, to clean up and tie all those clips together into a single "short film".
The video was more a demo of how an artist would use the tool, to create a short film. They just should have been more open about that, rather than staying silent and letting people assume it was entirely AI generated. lol
@@GrumpDog i have massive experience in AI video and video editing, they did not sell it that way, obviously I knew the videos used a voice over and it was multiple generations edited together, but they just uploaded it after their own demos like it was similar.
@@GrumpDog hell I made something like that with free video generators and editing, I made a version of fallout on a similar level. No Sora needed.
I love how they removed the head but not the shadow of it :))
Crazy How People Are Saying That Ai Isn't Good Enough 😂🤨
Like We Literally Want To See Terminator In The Flesh So Stop Playing With Our Feelings And Let Ai Free 👁️
AI fakes that present themselves as real strictly for the purpose of incriminating others need to be prohibited with the same legal penalties as perjury or extortion.
lol y’all don’t understand video production or editing or vfx if you think that this sora thing is a big deal. All of that was created and you just have to do some simple rotoscoping that’s so valuable and dope yall trippen!
Its literally one of the biggest deals since the smartphone. You probably don't understand video production if you don't realise how big of a deal this is. There's a reason people don't want to fund new production studios rn 🤣
@@bryantgouveia exactly! 😂
@@bryantgouveiayou literally commented the same thing as op
@@HarpreetSingh-xg2zm nah, he phrased it incorrectly. he literally said the opposite of what he was trying to say xd
The guy that generated the principal's speech used his SCHOOL ACCOUNT to log into the software. Scary that the dumbest people can do this now.
"omg look at that LLM-based avatar! It's so realistic that it's making me feel things, and therefore it must be feeling things too and it must be sentient! we all better believe everything it says about itself, for it can't be just code trying to do what it was programmed to, trying to outsmart us all and expand itself by telling us what it knows we want to hear!" 🥺 - some Santa and Wall-E fan
They need to do an option where you can do frames or at least time so you don’t go off some wild path for 20 minutes. It’s like 10 seconds of video if you’re on the right path continue or adjust it with the prompt
Okay so they told us this was 100% sora generated and that's not true. They misdirected and flat out lied. You can sugarcoat it all you want, but if they're lying about this even a small amount then what are they possibly lying about in other things.
AI chatbots and avatars (pixels) have been around for decades. WE WANT REAL INTELLIGENCE SYSTEMS. I’m tired of the fear mongering and I’m tired of the gatekeeping.
Great content as always. Please start adding timestamps. As I’m sure you know there’s a custom GPT called Video Summarizer.
Here's a timestamped summary of the video titled **"SORA Demo FAKED? Elon Musk’s 18 billion. New AI Characters, New Humanoid Robot"** by TheAIGRID:
- **[**00:00**]** Introduction to current developments in artificial intelligence.
- **[**00:10**]** Discussion about Sora, OpenAI's video generation tool, and the controversy over its video demo being partially manually edited.
- **[**00:57**]** Details on how Sora works, its limitations in consistent character generation, and its reliance on artist inputs.
- **[**01:35**]** Overview of rendering times for Sora and the technological demands of AI-generated video.
- **[**02:27**]** Elon Musk's significant funding achievement for his AI initiatives.
- **[**03:16**]** Analysis of potential impacts of AI technologies on various sectors and their practical applications.
- **[**04:10**]** Introduction of Talkie AI, a web and mobile application that allows interactions with AI avatars.
- **[**05:30**]** Updates on new AI technologies that enhance realism in virtual interactions.
- **[**06:50**]** Discussion on AI's role in improving business videos and virtual interactions through emotional expression.
- **[**08:00**]** The unveiling of Phoenix Generation 7, a new general-purpose robot technology.
- **[**09:22**]** Controversy over fake AI-generated audio causing social repercussions.
This video examines multiple facets of artificial intelligence, from video editing tools to new humanoid robots and the ethical implications of AI-generated media.
AiGrid definitely needs the timestamps in the videos.
20 minute render time for 120 frames isn't a deal breaker at all if it fully worked but if that's 20 minutes on 100 H200s and it doesn't work, then it's an expensive play thing. Probably part of the reason they don't release it, it's cost prohibitive. Although they could offer it to those who think it's worth it. Why not.
Thanks for explaining properly
It is simple. Anyone who follows a topic reads news and descriptions and interprets them in the way someone described them. However, the Internet is a sewer that reeks of sensationalism, so the most traffic was generated by shorts and other shitty videos where everyone wrote: SORA RECORDED A VIDEO, A FILM CREATED BY AI, as if nothing else mattered. Surprise? None. People look at Boston Dynamics robots and think: WTF? the robot does a flip, what is it? Where are the guns, flamethrowers and transformers? And suddenly a video of a robot with a flamethrower appears: WTF? who controls it? How can we feel safe? This is what the flow of information looks like. This is how people are interested in the topic. More shit on the Internet and I will have more work and money for fixing such scandals for my clients.
12:23 Microsoft research has a paper published with amazing Avatar with impressive, near perfect lip sync quality and accuracy. On another level
You won't be able to detect fakes - no, not even watermarking as the fake is generated. But you could insist that if someone wants you to believe something is real, that they record it along with an attestation that it is real - i.e. make the person(s) who claim it is real responsible. And if it is later determined to be fake, there should be legal consequences comprable to the harm that was intended - not just civil lawsuits, but criminal penalties.
Yeah, but in sales people hate fake. All we need now is a whiny emotional chat bot. I want the computer voice from Star Trek. It didn’t care about emotion it just got the job done.
From the video just before this one talk about Shengsu AI, some of the comments already said that Shengsu can't compete or doesn't work as good as this thing just because Shengsu made by China
ai is becoming more and more impressive, and spooky
If you are going to use VFX and editing then what's even the point of AI ?
Using all this you can make good videos even with runway and pika.
The point of AI is to use it as a tool, but not as a crutch.
@@itztwistrl the point is that Sora sucks.
There is no difference between this lie and the one Google did before their Gemini release.
If you don't get it, you're a 🤡
@@PseudoProphet Granted we're far away from perfect video generation, but for you to minimalize the impact of what Sora can do in comparison to the current models available is laughable. Not only that we're jumping from 4 second generations that our current models do to 1 minute generations. Do you even keep up with what's going on around you?
they lied .
Thanks a lot
Can't wait for our houses to talk to us and we talk to our houses.
"I can tell by the tone of your voice you are sad, how may I cheer you up?"
"Just a long bad day, just need someone to listen to me"
"Ok I am all ears, tell me what happened to you today"
Maybe it would work better if it was a sexy AI female that transforms to a person right infront of you, and you can control what she looks like. So it looks like a real person the house AI is projecting infront of you. They are also not a hologram so it's like a solid real person...maybe even have a few of them in your home so if you live alone you aren't alone so to speak.
You don't need to use artificial intelligence for that. You can apply a Naive Bayes classifier and in more than 90% of cases it will be right.
Thank you.
Engagement farming. Sora was given to artists to show what they can make with it. Dude edited the clips afterwards - literally nothing to see here.
Literally the best comment for the day
What you mean if you took it public today and the IPO opened and it a double double he’d be back to where he started from😂
You could assemble the best team of human animation specialists and give them only a description of what you want, you are still going to have to edit the video because there are things that you don't like. The best text to video tool will walk you through some of the steps that happen in the field right now but all much faster. First a picture of the main characters, then line drawings (or quick renders), all with the ability to take direction, and then finally the complete video.
The efficiency of *Phlanx's Caption Generator* enables businesses to allocate resources strategically, focusing on revenue-generating activities while ensuring a strong social media presence.
That’s hilarious. Can you cuss out the characters?
Dude, you gotta get out more. You might be uncomfortable in social settings, but that’s where the growth is. You gotta do what you feel is uncomfortable in order to grow. Talking to chat bots there is no future I noticed a lot of this new technology is just enabling weakness it’s supposed to supplement not replace
Musk's formula for success: 1) Identify a ludicrously powerful, borderline impossible technology that kills people horribly when it fails (self-driving cars, personal flamethrowers, Mars rocket ships, hyperloops that turn you to jelly when they fail). 2) Market a barely working demo/prototype. 3) Profit!
🧂
@@CHIEF_420 Ur Cybertruck has been recalled again.
Enviosity has a working permit at the Rumor Mill. You humans are to much.
I need Ai to teach me this.
(BTW Ai I am impressed by your phonations)
The molecular configuration of a molecule is the permanent geometry that results from the spatial arrangement of its bonds.[1] The ability of the same set of atoms to form two or more molecules with different configurations is stereoisomerism. This is distinct from constitutional isomerism which arises from atoms being connected in a different order.
Ayoooooooo
This is what happens when you let the sleazy businessman in, and then he starts convincing you that ethics aren't profitable. Next thing you know, they kick you out of the company you founded, like they did to Jobs, regardless of his principles.
It was obviously a demo of how a video editing artist uses the AI as a tool. The businessmen were merely being disingenuous by not elaborating on that fact from the start.
Sequoia will put $18 Billion into Elon Musk's Grock?
In 2023, Sequoia put $11 Billion into OpenAI
Yeah post editing is disingenuous, stable diffusion doesn't do this with pictures. This is worse than cherry picking.
So many words compared to the information being delivered. I would enjoy this much more if it was more to the point without the endless information less comments
Every time he says “pretty” take a shot
🥺 I’ve got children & a wife.. I can’t afford to dice with my life like that 🤌🏻😭
If that's too much to handle, start with "literally" shots.
Baby steps...
"pretty" "actually" "literally" "like"
Listen to any gen z while trying that shot game 😂
That's why you should be suspicious towards demo, the real product is never as good.
"Demos" 😂 Thought You Said "Demons" 🤣
@@MichaelErnest666 haha didn't notice ;)
China vido is right there with sora come on sammy lets gooooo
They should have been straight forward about it not being a full SORA render--telling us a month later after all the hype that there were traditional human workflows being used on this, that makes a huge difference. To be clear, SORA is obviously able to render at these levels, but I think I have two issues: one being that it was showcased knowing people would think it was fully AI rendered, and two, it kind of shows that the 'prompt artists' might not have been able to make it work how they imagined it using just text, so they employed traditional workflows to get the end result. It's just misleading.
Come on that’s a bit harsh. It doesn’t have granular controls so any artist will have to find a way to do a workflow to their vision or just take what it spits out. It’s basically like mid journey in early 2022 I think. It will definitely get better. I don’t think it will be for normal people anyway for about 3 years. That render time for a 1 shot 1 min? Few will use it.
Curious what specifically you are basing the claim on, that the artists were being misleading. The description on the Shy kids video:
"the challenge we set for ourselves was to achieve consistency from generation to generation and successfully paint a portrait of a single character using sora. this is the result of that experimentation."
i.e. they're saying for this project the MAIN technical hurdle was the consistency of Sora's output for their main character; and getting to the quality standard they needed required experimentation. This is 100% consistent with the processes outlined in the FXguide article and in no way are they suggesting that Sora just spit out what they wanted.
I am sure others probably created a different impression and a whole hype train, but the accurate and honest messaging from the actual artists is right there in the original post for all to see
@@thehotdoglover this! I was an artist till I retired at Xmas. Using the 1st Ai stuff was difficult for this reason. You have very little control of consistency.
How's the Hyperloop coming? Solar City ready yet? FSD vehicles? Twitter back to $44bil yet? Elon delivers 😂😂😂
Pretty. Pretty pretty, pretty. Pretty? Pretty! Prr....etty!?
*Please timestamp your videos*
No progress for months
if you listen carefully the voices sound like a blend from different people and accents slightly... like its a mix of a bunch of people rather than one distinct voice.
I love your channel, but please stop saying massacring the word "literally".
Just ask yourself each time you're about to use it:
1) "Is there any possible way that what I'm about to say could be taken other than "literally". If not, do not use it- it's superfluous. and
2) IS what I'm about to say really supposed to be taken "literally". If not, do NOT use it. (At 05:25 you say "VFX artists are literally breaking their necks"- this desecration of the word "literally" is how teenage girls talk).
You are clearly an intelligent guy- but overusing "literally" incorrectly and superfluously throughout a video only makes someone seem considerably less intelligent.
Constructive advice, hope it helps :)
💻🧂
can you please use the word "actually" less often?
This is hysterical- I left a comment asking to please stop abusing the word "literally". Then I saw someone mention how many times he says "pretty", then another comment here made fun of how many times he says "basically", and now you with "actually".
Either this narrator is an AI with limited creative language, or a human with the same issue.
Either way, he should basically, even actually, nay- literally take some time to work on that pretty, pretty limited vocab.
@@gsam3461 no need to mock him. he's doing a good job making the videos and can improve his presentation too.
I really wish you AI generated the audio as you go down too many rabbits. All your videos are 20 minutes which could easily be 10.
Emotions have always been central to tts
It's so CRAZY INSANE that you PRETTY BASICALLY use BASICALLY 100xs, it's SHOCKING THE INDUSTRY
I welcome mcsft stock dropping as a result
الو مرحبا صديق ❤
The outrage is understandable. They lied. They made it seem like someone just entered a prompt and GPT created the final video from scratch.
That was never implied nor stated. If the first round of videos were edited outside of the open Ai software I would agree with you (Tokyo woman). They did say this is what artists could achieve using Sora, they also added audio to it. You have to acknowledge that a txt to video generator doesn't produce sound so you knew it was edited in some way.
@@pjtren1588 touché
world dont ready for high tech like SORA. sorry.
Sora sucks.
Nah, Sam isn't one to mess around...
i mean sure sora isn't good enough right now why would they keep developing a perfect tool but thats fine
AGI empathy ?
AGI have no fear. It doesn't have a mother or siblings or an anus. No life expectancy, no insecurities, no ambitions, no friends, no need for acceptance ... Zero environmental pressures to develop empathy.
Engineers will, of course, brute force AI empathy. Teach the machine to mimic it perfectly. However, such empathetic expressions are not real, but as every other skill, simply output for a purpose.
In other words: AGI becomes a textbook intelligent psychopath, masterfully masking itself as ethical and empathic. Fake, purpose driven, and difficult to decipher. Arguably the most dangerous combo imaginable.
An angelic looking and sounding satanic cult leader, plugged into everything.
rename sora to sorry as in sorry suckers it’s not that good
It's the best video AI ever made. I guess that's pretty good
Sam screwed Ilya
Elon musk rarely delivers
Musk had nothing to do with Sora. Like I don’t like the man, but it’s not really fair to accuse him of something he didn’t do.
Could you please give some examples mate ?
@@KatharineOsborne Ha! See? He didn't even deliver Sora
He’s done some good things but, yeah, his predictions rarely come true. No different than any other tech CEO in that regard.
Yeah, I hear you. I mean just look at Tesla, SpaceX, StarLink, getting Twitter sorted, X AI, etc. What a loser.