EMO by AliBaba is scary-good at taking any supply picture plus audio and making the picture seem like they’re saying/singing the …
source
Meet the Unitree G1 AI Humanoid Robotic! This superior know-how is altering the sport in robotics. Uncover the capabilities of ... source
Read more
Should Alibaba open-source this project? Or is it too dangerous?
What's funny with AI…humans can imagine this in their minds, but can't put it into media… its like AI can output the thoughts onto media we can observe….very quickly.
The twist: this entire video is in EMO.
You can tell it's AI Though. 🤷🏼♀️
can totally tell they are fake, what are you even on about?
It just is noticeable…For now…
Its seems AI has similar trouble with natural eye movement as it does this frequently with hands
This is nothing new, Trump and Biden both died several years ago but most people think they see them every day.
Install this technology in the positronic neural network of an advanced generation biologically indistinguishable-from-human android, and viola, instant West World/Bladerunner!!
Audrey Hepburn + vocal fry = No thank you
The big question is WHY are we creating technology that WILL be used for mostly illegal and nefarious purposes.
The most realistic thing about it is the fact that the neck and throat are moving accurately. There actually looks like vocalization is actually happening naturally. This is the thing that makes it more believable for me. Thankfully just using the photos (the Weak part of that long title of the project) it isn’t as accurate on shading or depth, yet.
Roma Holiday movie is easy to forget
What will happen is the same as visual effects in movies, it will all look "realistic", but we will become accustomed to the little inconsistencies and recognise the tells… So we will know its AI except in the right circumstances.
GF!=AI (wait 2 years you understand )
Beginning of Running Man
Mindblowing.
Holy shit! This an eye-opening!
I mean TBH- the fakes are fairly obvious
This is the same old bullshit from all of the artificial intelligence companies. Oh, look, honey, a new AI application that we don’t get to use! But we can watch videos of other people using it. The future is so futuristic.
We're Doomed
Like, I think I'm looking at the most painfully sunken eyes I've ever seen – like I can literally see your skull. Surely that's CGI.
The detail of the face of the girl wasn't there but may get there and the animation of the face to the audio was awesome.
the radical liberal deepstate has done everything and thrown everything at trump as election interference only making him more popular , it's just a matter of time before they deepfake him as a racist fake video etc. if that doesn't work they will j.f.k. him , like they did Kennedy. rather then to let trump get re elected and save America from radical communism.
Anti-Trump College Student Can't Back Up Claims
https://www.youtube.com/watch?v=5rn3Ct5mACY
The Inconvenient Truth About the Democratic Party:
https://www.youtube.com/watch?v=g_a7dQXilCo&t=2s
Hawley Exposes Biden HHS Officer For Facilitating Largest Child Trafficking Ring In U.S. History
https://www.youtube.com/watch?v=HZDebmlu6Nk
"FIRST THEY GO AFTER TRUMP, NEXT WILL BE YOU, IF THE RADIAL LIBERAL DEMOCRATS CAN JAIL OR GO AFTER THEIR ELECTION OPPONENTS WITH LIES,MISINFORMATION, CROOKED LIBERAL D.O.J., F.B.I. AND C.I.A.,JUDGES AND PROSECUTOR'S ,LIKE OTHER COMMUNIST COUNTRIES DO AND EVEN GO AFTER A FORMER PRESIDENT OF THE UNITED STATES IN ELECTION INTERFERRENCE TO KEEP TRUMP FROM BEING RE-ELECTED TO SAVE AMERICA, WHAT DO YOU THINK THEY WILL DO WITH ALL OTHER AMERICANS LATER ON THAT CAN'T FIGHT THEM IF THEY WIN?"
TRUMP 2024 LAW AND ORDER AND TO SAVE AMERICA FROM RADICAL LIBERAL DEMOCRATS THAT ARE PUSHING COMMUNISM AND SOCIALISM AND ARE A THREAT TO DEMOCRACY TO AMERICA AND THE WORLD!! . LET'S ALL STAND UP FOR THE "AMERICAN" FLAG AND RING THE LIBERTY BELL!!
COMMUNISM CHINA LIKE CENSORSHIP OF AMERICANS FREEDOM OF SPEECH THAT IS GUARANTEED BY THE CONSTITUTION IS NOW BEING CENSORED BY LIBERAL COMMUNITY GUIDLINES CLAIMING HATE SPEECH AS AN EXCUSE TO CENSOR AMERICANS !
PEOPLE THAT CALL TRUMP "HITLER", WANT HATE SPEECH LAWS!
https://www.youtube.com/shorts/qlauedHyfyA
hate speech :
Here are six videos of Democrats calling for violence or physical confrontations that are still active on Twitter and other liberal social media sites!!
https://www.washingtonexaminer.com/news/democrats-physically-confont-twitter
" NOT MUCH HAS CHANGED HAS IT, EXCEPT THEY NOW HAVE MORE CONTROL OVER MOST OF SOCIAL MEDIA AND NEWS OUTLETS TO SPREAD THEIR HATE, MISINFORMATION AND LIES!!"
That is ABSOLUTELY AND TOTALLY NUTS!!!
I know how to program and I've written a huge bunch of books about it – and I've worked with REALLY VERY COMPLEXE SOFTWARE SYSTEMS!
Let me tell you this: EVERY PROGRAMMED SOFTWARE WILL DEFINITELY NOT FUNCTION THE WAY YOU WANT IT TO!!!
There will be mistakes and failures in GARGANTUAN MASSES!!!
G U A R A N T E E D !!
If you are satisfied with what AI delivers in the first step or version then okay! Go with that!!
That will lead to a situation where nearly every software will look kind of similar in a certain way!! Because the main challenge is that people in general are BY FAAAAR not as creative as they "HAS TO BE"!!
So, professionals will use some total other skills, features and dimensions of a software tool than "usual" people will do!!!!
The conclusion will be, that EVERYBODY who wants to correct an AI-software, has to be able to control that AI-Software line by line and will have to programm small patches and extensions just right INTO the AI software till it fits the demands of the producer!!!
And the demands of customers can be ENDLESS AND CRUEL!!! BELIEVE IT!
And the good paying customer will never be satisfied with AI-"quickshots"!
N E V E R E V E R !!!!!
So the opposite is true: learn programming as good as you ever can! You will surely be the one who will have to controll AI! AND TO CORRECT IT!!! INEVITABLY!!!
See the DIY markets for hobbyists!
The more the people will build and tinker with hands the more little und unpredictable tools they will need for their endless masses of corrections!
The same will come with AI-software!
People will desperately search for software specialists who will give them these small tools, repairs and individual patches!
Wait … and you will undoubtedly see it….
Imagine those giant time, knowledge and efforts off all the others who then will want to or shall form a proper prompt for those complex corrections and extensions!!
Nothing will be more complicated than that!! And the very most of those who has to try it, will hopelessly give up!
The average IQ of mankind is at 100!! AND THAT MEANS THAT HALF OF ALL PEOPLE LIES LOWER!!
Be well prepared… ;O)
PS: who has questions about that subject, ask under "jamboozy@gmx.de"
HEUREKA
Given how rapidly AI/ML technology is advancing and the resulting options for DeepFake manipulation, the question inevitably arises: how can we ensure the authenticity of video and audio recordings 100%?
One feasible solution is to always require encoding of any image, streamed video, or streamed audio to take MORE time than the fastest possible live generation of a medium-specific key would need!
This means that an ad hoc key generation, using a combination of biometrics, steganography, VPN (and possibly PGP) via quantum computing, MUST always be faster than the fastest possible encryption analysis and calculation of a corresponding stream decoding!
VPN sets the example by making the current live transmission key increasingly complex, making it impossible for a computer to calculate and transmit it instantly using state-of-the-art methods!
In the era of quantum technology, a "cheap" arithmetic key unfortunately no longer suffices. Even if it's a gigabyte long! Those naive and "cozy" times are definitively over with AI!
The sender of a graphical or phonetic message MUST simultaneously transmit a live-stitched "watermark" based on biometric live data alongside the stream transmission, which is absolutely unique for every trillionth of a second of the current direct transmission!
During live communication, such as video chat, a reliable camera must continuously track the sender's eye movements and "living" iris features. Using steganographic calculations, this watermark is then incorporated (stitched) into the respective streaming medium. Steganography is used because without the specific key and knowledge that the medium has been encoded, it is impossible to detect an encoded medium as encoded!
The recipient will only see or hear media that are legally transmitted along with such a publicly verified key!
Alternatively, all unverified graphic media could be marked with a blinking red "F" (for "Fake") in the upper left corner of the image or with an unpleasant sound (whistling) in the audio stream!
Verification could be carried out by a cloud-based, publicly and ABSOLUTELY trustworthy UN institute, for instance!
The time required for the detection, analysis, calculation, and generation of such a stitch key MUST IN ANY CASE be longer than the key verification performed by the UN Trust Institute!
For already previously stored media, a new and unique live key MUST be generated by the verification institute at the time of retransmission and live integrated into the current stream!
PS: If you have questions about that issue, then ask under "jamboozy@gmx.de"
———————————-
HEUREKA
Bei der gegenwärtigen atemberaubenden Geschwindigkeit der AI/KI-Entwicklung und den damit einhergehenden DeepFake-Optionen stellt sich unweigerlich die Frage, wie wir die Echtheit von Video- und Soundaufnahmen 100% gewährleisten können!?
Eine machbare Lösung ist, dass die Encodierung eines beliebigen Bildes, Streamvideos oder Streamsounds IMMER mehr Zeit in Anspruch nehmen muss, als die schnellstmögliche Live-Generierung eines mediumabhängigen Schlüssels benötigen würde!
Das heißt, dass eine AdHoc-Schlüsselerzeugung durch eine Kombination von Biometrie, Steganotechnik, VPN (und evtl. auch PGP) per Quantumcomputing IMMER schneller sein muss, als die schnellstmögliche Encryptinganalyse und Berechnung einer entsprechenden Streamdecodierung!
VPN macht es vor, indem der aktuelle Live-Schlüssel der Übertragung immer komplexer ist, als ein Computer ihn instantan mit modernsten Mitteln berechnen und übertragen könnte!
In Zeiten der Quantumtechnik reicht ein "billiger" arithmetischer Schlüssel leider nicht mehr aus. Und sei er ein Gigabyte lang! Diese naiven und "gemütlichen" Zeiten sind mit der AI ein- und für allemal vorbei!
Der Sender einer grafischen oder phonetischen Nachricht MUSS parallel zur Streamübertragung ein live-gestitshtes (eingenähtes) "Wasserzeichen" auf Basis biometrischer Live-Daten übertragen, dass mit JEDER Nachricht mitgesendet wird und welches für den trilliardstel Sekundenmoment der aktuellen Direktübertragung absolut einmalig ist!
Dazu MUSS bei einer Live-Kommunikation (Videochat) eine vertrauenswürdige Kamera z.B. ständig die Augenbewegungen und die "lebendigen" Iris-Merkmale des Senders feststellen und dieses Wasserzeichen anhand von steganotechnischen Berechnungen in das jeweilige Streamingmedium eingearbeitet (gestitcht) werden!
Steganotechnik deshalb, weil es ohne den konkreten Schlüssel und das konkrete Wissen, dass das Medium codiert wurde, unmöglich ist, ein codiertes Medium zu erkennen!
Der Empfänger sieht oder hört dann nur solche Medien, die zusammen mit einem solchen öffentlich verifizierten Schlüssel legal gesendet werden! Bzw. könnten alle nichtverifizierten Grafikmedien mit einem blinkenden roten "F" (für "Fake") in der linken oberen Bildecke oder mit einem unangenehmen Geräusch (Pfeiffen) im Audiostream gekennzeichnet werden!
Die Verifikation müsste z.B. live ein cloudbasiertes öffentliches und ABSOLUT vertrauenswürdiges UN-Institut übernehmen!
Der Zeitaufwand für die Erkennung, Analyse, Berechnung und Erzeugung eines solchen Stitch-Schlüssels MUSS IN JEDEM FALL größer sein, als die Schlüsselverifikation des UN-Vertrauensinstituts!
Bei bereits gespeicherten Medien MUSS zum Zeitpunkt der erneuten Übertragung ein neuer und einmaliger Live-Schlüssel vom Verifizierungsinstitut generiert und live in den aktuellen Stream eingebunden werden!
PS: Fall Sie Fragen dazu haben, dann fragen Sie unter "jamboozy@gmx.de"
not at all believable, lag stutter, skin eye movement ect .. all lacking "reality"
Only face to face would be trusted … for now I guess xD
Where and how to use it? Not sure if links are blocked in some countries as Google is not showing any relevant link.
Bring back loved ones who have departed long ago.
Alibaba rule the world ehehehh